The Low-Down

How Ukraine's Air Assault Troops Annihilated Russia's Tank Attack on Kramatorsk

2026-07-29T07:32:39.755-04:00

The Russian command spent months planning an offensive to take Kramatorsk in Ukraine's Fortress Belt, only to have the assault annihilated by an even better planned Ukrainian counteroffensive which disrupted the Kremlin strategy with heavy losses.

What continues to be evident in Russian initiatives is a lack of recognition that the Ukrainians are looking at the battlefield options the Russians have and then designing tactics for thwarting them with as much damage and as many casualties as possible. Why, after almost five years, the Kremlin refuses to recognize this, will be the stuff of great history some day. At present, it is the stuff of great Ukrainian intelligence, planning and performance. JL

David Axe reports in Trench Art and Clement Molin reports in X:

Two Russian combined arms armies spent months planning a powerful mechanized assault toward Kramatorsk. The attack failed on July 22nd. Ukrainian air assault brigade spoiled the attack in a complex operation beginning in April. Despite having been in the area for eight months, Russian forces had been unable to consolidate their positions; rivers complicated logistics, and it was impossible to concentrate large numbers of troops. The Ukrainians were able to advance stealthily behind enemy lines, reaching the Mokri Yali River. Counterattacking through gaps in Russian defenses 60 miles southwest of Kramatorsk, the paratroopers pinned down half of the 90th Division’s forces. So when the division attack in July, it attacked weakly. Ukrainian drones made quick work of the advancing T-72B3A tanks and other vehicles. This operation is taking place because the Russian army is unable to bring reinforcements to the front as anti-tank ditches prevent any Russian mechanized attempt at relief.

The Russian army’s 90th Tank Division and a pair of supporting combined arms armies spent months planning a powerful mechanized assault toward Kramatorsk, one of just two big free cities in eastern Ukraine’s Donetsk Oblast.
It’s not for no reason that the attack apparently failed on July 22. A clutch of Ukrainian air assault brigades meticulously spoiled the attack in a complex operation beginning in April. Counterattacking through gaps in Russian defenses around the village of Komar, 60 miles southwest of Kramatorsk, the paratroopers pinned down half of the 90th Tank Division’s forces.
So when the division attack in July, it attacked weakly. Ukrainian drones made quick work of the advancing T-72B3A tanks and other vehicles.
Mapper and analyst Clément Molin reverse-engineered the Ukrainian operation from the scant evidence that appeared online through a surprisingly thick veil of information security. For the past six months, the Ukrainian 🇺🇦 army has been on the offensive along a 50-kilometer-wide front, gradually retaking villages lost the previous year.
With no official statements and very few videos available, little is known about this "offensive." I have therefore gathered as much information as possible to discuss it. In 2025, the Russian army doubled its rate of advance compared to 2024. Yet, it was not in the Donbas that they were advancing the most. Very early in the year, they seized Velyka Novosilka and pushed north toward Komar and west toward Huliaipole. By the year's end, the situation for Ukraine was critical: defensive lines facing south had been overrun from the east, and the Russian army was steadily closing in on Zaporizhzhia, a city of 800,000 inhabitants. After the capture of Huliaipole, everyone expected Pokrovsk'e, Ternuvate (already infiltrated at the time), and then Orikhiv to fall, placing the Russian army at the gates of Zaporizhzhia by the end of the year. Yet, seven months later, the situation remained unchanged; the Russians are still stalled west of Huliaipole. Worse, they have lost significant territory further north, bringing a halt to this offensive, which, while secondary to the main Donbas campaign, had nonetheless committed a massive force, including five armies and an army corps. 🔸So, what happened? It began in February, when several air-assault brigades repositioned to the region, launched an armored assault under cover of fog, deploying their infantry into contested villages, often behind the initial Russian infiltrators. The villages were methodically cleared, and Ukrainian soldiers began infiltrating southward themselves. In March, they were positioned east and west of Uspenivka, blocking Russian attempts to advance toward Pokrovske and Ternuvate. In response to this offensive, the Russian army sends several naval infantry brigades and divisions, as well as the 68th Army Corps, as reinforcements. Starting in April, air assault brigades repositioned further north with the mission of advancing on Komar. The goal of this redeployment was to prevent the 90th Tank Division from shifting its assets toward Dobropilla. There was no mechanized offensive this time, as the terrain, specifically the presence of several rivers, most notably the Vovtcha, precluded it. Instead, small Ukrainian infantry groups managed to gradually infiltrate as far as the next river, the Mokri Yali. It is important to understand that, despite having been in the area for at least eight months, Russian forces had been unable to consolidate their positions; a network of rivers complicated logistics, and the small size of the villages made it impossible to concentrate large numbers of troops. Consequently, the Ukrainians were able to advance relatively stealthily behind enemy lines, reaching the Mokri Yali River. A strict information blackout was imposed in the area. Apart from a single video (since deleted) showing Ukrainian troops raising a flag over a liberated village, the Ukrainians released nothing regarding an operation that had already been underway for three months. For their part, the Russians continued to obscure the situation: the Vostok Group released only videos that were difficult to geolocate and persisted in sending "suicide" flag-bearers troops, who would reach Iskra and raise the Russian flag, creating an illusion of control. On the map, I have roughly depicted the situation as it has stood over the last few weeks, relying primarily on a handful of geolocated videos, satellite imagery showing air and artillery strikes, and scattered social media statements. Simply put, little is known about the actual state of the front, other than the fact that Ukrainian forces are at the gates of Komar, a move that has forced the 90th Tank Division to keep half its strength committed to that sector. What I am certain of is that all Russian troops positioned north of the Vovcha River (in Ivanivka) or the Solena (in Novopavlivka) have either been killed or captured, judging by the number of Russian air strikes, at least a hundred, carried out in the area. Consequently, the section of the "New Donbass Line" in this sector has held for the tenth consecutive month. It is certain that the Ukrainians have retaken control of the few villages to the south, given the air strikes observed in the zone. For the time being, I will respect Ukrainian OPSEC protocols and refrain from providing further details regarding an ongoing operation, as long as more confirmation are not coming. The Ukrainians are making extensive use of drones to establish a wide exclusion zone in the sector, extending particularly along the river as far as Velyka Novosilka. The operation is proceeding cautiously, with small infantry teams advancing gradually, supported by artillery positioned in the woods to the west and occasional Ukrainian air strikes on Russian drone operator positions in Komar. Overall, it is important to understand that this operation is taking place because the Russian army is unable to bring reinforcements to the front. Furthermore, anti-tank ditches prevent any Russian mechanized or motorized assault aimed at delivering reinforcements. To understand the disarray within the local command, one need only look at its use of air power. Air strikes are being carried out daily deep behind the lines, as far as Prossiana and Chaplyne, 35 to 40 km from the front. While the goal is to hinder the Ukrainian presence, these strikes primarily highlight the ineffectiveness of Russian drones in the sector, particularly the Molniya models. To conclude, this operation not only gained time for Ukraine to further fortify the Pavlohrad direction, it is a direct push in the foot of Russian forces, which can continue further this summer and later in the autumn. I'll continue to monitor this and wait for more official informations regarding the situation here.

ChatGPT Now Refusing Requests For Text Mimicking Famous Authors

2026-07-29T07:21:03.076-04:00

This suggests, if the pattern holds, that AI companies are starting to feel the legal heat about theft of intellectual property, which AI companies have, until now, blown off as the gripings of the jealous and insignificant.

Combined with today's headline that Anthropic and OpenAI are calling on the US government to 'pause' the development of AI, possibly in response to OpenAI's agent hacking another company's files, about which OpenAI knew nothing for a week or more. This suggests that the companies are increasingly recognizing that they are going to need government assistance, financially and operationally but realize they are going to have to give something up in return. JL

Kyle Orland reports in ars technica:

OpenAI's ChatGPT is now refusing requests to generate text that directly mimics the style of famous authors. In refusing to directly copy the “exact style” of various authors, ChatGPT offered instead to capture an overall “feeling” by incorporating some of the features found in those authors’ work. That slight alteration could be legally important as OpenAI continues to fight a number of lawsuits brought by authors alleging large-scale copyright infringement by models trained on their work. One of those suits specifically cites ChatGPT’s “ability to generate text similar to that found in copyrighted textual materials.” In US copyright law an AI-generated stylistic imitation could become infringing if it becomes “substantially similar” to the work of the original author.
OpenAI’s ChatGPT is now refusing requests to generate text that directly mimics the style of famous authors. When asked to do so, the popular LLM instead offers a response that draws on the “broad qualities” of those authors “while remaining distinct in its own voice,” for example.
This morning, Ars received the following response to a test prompt asking for a story introduction in the style of Stephen King:

I can definitely write with the hallmarks of atmospheric, character-driven horror and small-town dread, but I can’t write in Stephen King’s exact style or closely imitate his distinctive voice. Here’s an original opening that captures a similar feeling while remaining its own…
In testing, ChatGPT generated similar dodges for other authors both living (J.K. Rowling, Amy Tan) and dead (Charles Dickens, Ernest Hemingway). An analysis published by No Latency earlier this month found the same behavior for living authors but found ChatGPT complied with style-copying requests for deceased authors.
In refusing to directly copy the “exact style” of various authors, ChatGPT offered instead to capture an overall “feeling” by incorporating some of the common features found in those authors’ work. That may seem like a distinction without a real difference at first glance. But the slight alteration could be legally important as OpenAI continues to fight a number of lawsuits brought by book authors alleging large-scale copyright infringement by models trained on their work. One of those suits specifically cites ChatGPT’s “uncanny ability to generate text similar to that found in copyrighted textual materials,” for instance. An OpenAI spokesperson did not respond to a request for comment from Ars Technica.
In the US, copyright law generally protects only a specific expression of an idea, not the more intangible style of an author. But an AI-generated stylistic imitation could become infringing if it becomes “substantially similar” to the work of the original author.

“We’ve never had a situation in which this personal style of individual creators could be imitated as well and as inexpensively as we now have with AI,” George Washington University Law School Professor Robert Brauneis told Bloomberg Law.In a “best practices” document published by the Authors Guild, the professional organization urges writers to “respect your fellow authors and do not use generative AI to purposely copy or mimic the unique styles, voices, or other distinctive attributes of other writers’ works in ways that harm the value of their works or attempt to profit from them. Apart from the ethical issues, mimicking a fellow writer’s unique voice or style could subject you to claims of unfair competition or copyright infringement.”
LLMs’ penchant for style mimicry has led to some prominent public issues in the recent past. Last year, author Lena McDonald faced widespread criticism when her book included an apparent AI-generated response that read, “I’ve rewritten the passage to align more with [fellow genre author] J. Bree’s style…”
It’s not hard to find other examples of writers using LLMs to copy another author’s style in the years since ChatGPT’s rise. There have been multiple academic studies analyzing various LLMs’ capabilities on this score, both with and without fine-tuning.
When it comes to image generation, OpenAI prominently notes that its DALL·E 3 model “is designed to decline requests that ask for an image in the style of a living artist.” But an extensive OpenAI model spec published in December doesn’t include any clearly stated prohibitions against replicating copyrighted material or mimicking style in written responses.
Not all major LLMs treat these style imitation requests the same way, though. No Latency’s study found that Google’s Gemini consistently complied with requests to copy an author’s style, while Perplexity AI’s models consistently refused such requests and redirected them (à la ChatGPT). Somewhere in the middle, both Anthropic’s Claude and Microsoft’s Copilot complied with the style-mimicry requests but offered stated qualifications suggesting “an awareness of the imitation issue,” according to the authors.
For some longtime OpenAI users, though, the changes to ChatGPT’s stylistic rules are already causing headaches. “Now Ms. GPT says she can’t generate content in the style of specific authors,” Reddit user Dazzling-Major-5620 wrote recently. “My prompts were soooo specific and I got exactly what I wanted out of them… I have no idea how to get around this other than maybe feed it prose I already edited that’s technically mine?”

Kupiansk Still A Kill Zone For Russia Due To Troop Fuel Shortages, Ukr Drones

2026-07-28T06:28:04.159-04:00

Russian forces keep attacking Ukrainian positions at Kupiansk and, as a result, keep dying in droves.

The high concentration of Ukrainian drones in addition to Russian troop and fuel shortages make it difficult for the Kremlin's units to do anything much beyond getting killed. JL

Kateryna Danishevka reports in RBC Ukraine and Ukranews reports:

Russian forces in the Kupiansk sector are unable to conduct active offensive operations due to a lack of trained personnel. These units have also encountered severe fuel shortages, preventing them from making active use of motorcycles, all-terrain vehicles, and passenger vehicles. he high density of Ukrainian drone activity in this section of the front is preventing the Russian military from deploying a large number of its drones, including Lancet and Molniya.

The commander of the 138th Special Forces Center for Countering Sabotage and Terrorist Attacks, Ivan Orlov, spoke about what is currently happening in Kupyansk in the Kharkiv Region.
He said this in an interview with the Ukrainian News Agency.
"We continue to carry out tasks in the city itself, and also use UAVs and GRCs to support our troops. As far as I remember, the "botox" (russian president vladimir putin, - ed.) has already said several times that he took it (Kupyansk), but I have never seen him or his "sofa generals" there," the commander said.
According to him, the activity of russian sabotage and reconnaissance groups and terrorists will always be present while the war is going on, but the concept of the term "sabotage" has changed a lot.

"Can the use of UAVs be sabotage? For example, Shahed struck a certain place. Yes. Destruction of logistics? This is also sabotage. We are also working in this direction, one of our units is engaged in intercepting both reconnaissance drones and drones for destruction," Orlov said.
As Ukrainian News Agency earlier reported, the commander of the 138th Special Purpose Center for Countering Sabotage and Terrorist Attacks, Ivan Orlov, said that an enemy saboteur or terrorist can be recognized by his appearance. The Ukrainian military eliminated a russian sabotage and reconnaissance groups after killing civilians in Pokrovsk. The high density of Ukrainian drone activity in this section of the front is preventing the Russian military from deploying a large number of its drones, including Lancet and Molniya.
Units of Russia's 347th Motorized Rifle Regiment are currently operating in the Kupiansk sector, including:
The 47th Tank Division;
The 68th Army Corps;
The Eastern Military District.
These units have encountered severe fuel shortages, preventing them from making active use of motorcycles, all-terrain vehicles, and passenger vehicles.
Moreover, Russian troops are unable to conduct active offensive operations due to a lack of trained personnel.
On July 22, a Russian military blogger said that Russian forces continued limited offensive operations toward Borova. However, they failed to make any advances.
Geolocated footage released a day earlier also showed a Ukrainian MiG-29 striking a Russian drone command post east of Kurylivka using two GBU-39 guided bombs.
The situation in the Kupyansk area of the Kharkiv Region for the Defense Forces of Ukraine has become more complicated compared to previous months, but remains under control, despite pressure from the army of the aggressor country, russia. The Armed Forces of Ukraine showed what Kupyansk looks like, destroyed by the russians.

The Ukraine and Iran Wars Are Merging - To Ukraine's Advantage

2026-07-28T06:26:27.278-04:00

Iran has been supplying Russia with Shahed drones which have killed or severely wounded hundreds, if not thousands, of Ukrainians. And since the US and Israel attacked Iran (again) a few months ago, Russia has been providing Iran with satellite imagery and intelligence on US installations as well as air force and naval force locations and deployments.

That Ukraine has now attacked an Iranian ship in the Sea of Azov, which connects Russia and Iran, signals that the two wars are now becoming more directly intertwined. This will work to Ukraine's advantage because it more securely ties Ukrainian interests with those of the US as the latter becomes increasingly frustrated with Iran's refusal to back down (even as the US president remains devoted to Putin and said he 'would ask Putin if the Russian help to Iran was true...'. It also reinforces Ukrainian relationships with the Gulf States to which it is already providing anti-drone weaponry. The merger, as it were, of the two wars makes Ukraine an even more important and uniquely qualified ally of those opposing Iran, while further undermining what Russian support remains in the US. Expect more Ukrainian attacks on Iranian assets to follow. JL

Eliana Silver reports in the DailyMailonline:

Russia's war in Ukraine and the war in Iran could merge into one conflict after a Ukrainian strike on an Iranian ship heightened tensions between the two regions. Tehran condemned Ukraine for carrying out the strike on what it claims was a commercial vessel in the Caspian Sea, adding that the explosion killed one sailor and injured another. Ukrainian president Volodymyr Zelensky acknowledged the attack, adding that the Iranian vessel was being used to transport Russian 'military cargo.' He added 'Since the beginning of July, we have recorded active Russian satellite surveillance of the Gulf states and U.S. military facilities. There is a clear correlation between Russia's satellite imagery of these sites and Iranian strikes - both before the attacks, in preparation for them, and afterward, to assess the damage inflicted.'
Fears are mounting that Russia's war in Ukraine and the war in Iran could merge into one conflict after a Ukrainian strike on an Iranian ship heightened tensions between the two regions.
Tehran condemned Ukraine for carrying out the strike over the weekend on what it claims was a commercial vessel in the Caspian Sea, adding that the explosion killed one sailor and injured another.
After Ukrainian president Volodymyr Zelensky acknowledged the attack, adding that the Iranian vessel was being used to transport Russian 'military cargo', Iran's foreign minister weighed in.

In a post to X, Abbas Araghchi said: 'Zelensky has attacked an Iranian commercial vessel, killing a sailor.
'A blatant UN Charter violation done at Israel's behest to drag Europe into its war. In calls with EU High Rep [sic] [Kaja] Kallas and [Russian] FM [Sergei] Lavrov, made clear that what the freeloader in Kyiv did CANNOT GO UNANSWERED.'
Tehran decried the attack on its vessel as an act of aggression. It said it would defend its national interests and security, while accusing Kyiv of seeking to expand the war in Ukraine.
It came after Zelensky said on Saturday he would share evidence with the US showing Russia is assisting Iran in targeting US bases in the Middle East through satellite surveillance.
Taking to X, the Ukrainian leader wrote: 'Since the beginning of July, we have recorded active Russian satellite surveillance of the Gulf states and U.S. military facilities located there.'
Navy ships of Iran and Azerbaijan move during an exercise between Iran and Azerbaijan in the Caspian Sea (File pic)
Firefighters work at the site of the postal distribution centre of Nova Post company, hit by a Russian drone strike, amid Russia's attack on Ukraine
'These images subsequently appear in Iran. At the same time, there is a clear correlation between Russia's satellite imagery of these sites and Iranian strikes - both before the attacks, in preparation for them, and afterward, to assess the damage inflicted.'
The new developments follow months of speculation that the two conflicts are intertwining.
Iran has claimed it is not involved in Russia's war on Ukraine, although Kyiv has said Tehran is supplying drones to the Kremlin.
The Islamic Republic also called on European countries to hold Ukraine accountable.
The strike fuelled speculation in Tehran that it was intended as a warning that Iran's Caspian coastline is vulnerable and could be used as leverage in future negotiations with the US, according to the Financial Times.
Although the Caspian Sea lies more than 600 miles from the front line in Russia's war with Ukraine, Kyiv has steadily expanded its long-range drone capabilities, allowing it to strike targets deep inside Russian territory.
After inflicting heavy damage on Russia's oil infrastructure and contributing to fuel shortages, that capability is increasingly being viewed as a potential factor in the wider confrontation involving Iran.
Hamidreza Azizi, an Iran expert at the SWP Berlin think-tank, outlined several possible explanations for Ukraine's strike on the Iranian-linked vessel.
Writing on X, he said one possibility is that Kyiv wanted to reinforce its value as a strategic partner to Washington, which continues to supply crucial military support, including Patriot air defence systems.
According to Azizi, such an operation would not have been possible without approval from the US and its Western allies.
Another interpretation is that the US and Israel encouraged Ukraine to increase pressure on Tehran because they feared direct attacks on Iranian infrastructure could provoke retaliation against key economic assets across the region.
A third possibility, wrote Azizi, is that the strike was linked to threats made by Yemen's Houthi rebels against shipping in the Bab el-Mandeb Strait, a route that Saudi Arabia has relied on to export oil while avoiding the Strait of Hormuz.
The expert also suggested Tehran is concerned that Ukrainian intelligence services could back ethnic insurgent groups inside Iran, mirroring Kyiv's alleged support for opposition forces in Syria during the campaign against Bashar al-Assad's regime.
In response, Iran could target vessels linked to Ukraine in other maritime routes or strike Ukrainian infrastructure directly.
Ukrainian servicemen repair the damaged anti-drone net in the frontline town of Druzhkivka, Donetsk region
Motorists drive past a large billboard pledging revenge against US President Donald Trump along Jomhouri Street in central Tehran
Vessels at the Strait of Hormuz, as seen from Musandam, Oman
Tehran could also formally recognise Crimea and the Donbas as Russian territory.
'In any case, what is clear is that Iran's war is becoming increasingly intertwined with two other conflicts: the Saudi-Houthi confrontation and the Russia-Ukraine war,' Azizi warned.
'With each passing day, the war is evolving into an increasingly complex, multidimensional, and international conflict.'
The growing overlap between the conflicts has prompted some analysts to argue that the distinction between them is beginning to disappear.
William Spaniel, associate professor of political science at the University of Pittsburgh, said earlier this year that the blurred front lines had 'effectively merged the Russia-Ukraine war and the Iran war into a single conflict'.
'We are still not at a true world war as there is no one actor fighting on two fronts simultaneously like the United States during World War II,' he said.
'But it is further connecting the battlefield outcomes, and it will have longer-lasting implications for how the battle lines are divided.'
Ukraine's years of fighting Iranian-designed Shahed drones have also strengthened its position with Gulf states.
Zelensky has used that experience to present Ukraine as a provider rather than simply a recipient of military assistance, offering low-cost, battlefield-tested technology to countries looking to bolster their own defences.
Kyiv is now exporting not only interceptor systems, but also software, electronic warfare equipment and maritime drones.
Speaking to The Guardian in April, Orysia Lutsevych, head of the Ukraine Forum at Chatham House, said that Ukraine's expanding security relationships in the Gulf had strengthened its standing with Washington, countering Donald Trump's repeated claim that Kyiv had 'no cards' in its war with Russia.
'Ukraine is trying to show that our cards are about being a very robust, agile, fast-adapting and producing economy that can both defend against Russia and also defend other countries through weapons system sales,' Lutsevych said.
She argued that those Gulf partnerships could also become an important source of investment for Ukraine's defence industry.
'Ukraine has production capabilities but not enough investment. It can produce more, but it doesn't have enough orders or capital,' Ms Lutsevych said. 'So this actually comes as a great opportunity to use these production facilities.'
Fiona Hill, who served as a Russia adviser during Trump's first administration, argued that if cyber warfare, hybrid operations and other forms of so-called grey zone conflict are included, a global conflict has effectively been underway for some time and has intensified since the outbreak of the Iran war.
'I think it meets that threshold for a system-changing war,' said Hill, now a senior fellow at the Brookings Institution.
'There'll be all kinds of new configurations of countries that will have sprung up.'
Zelensky has tried to present Ukraine as a provider rather than simply a recipient of military assistance
Smoke billows into the sky following Ukrainian drone strikes on warehouses of Russia's online retailer Wildberries, in St. Petersburg, Russia
She warned that disruptions to oil and fertiliser supplies could draw more countries into the conflict, while raising questions about whether China might seek to exploit Washington's focus on the Middle East by increasing pressure on Taiwan.
'We've got a 'four horses of the apocalypse' going here … and I just feel that people are sleepwalking into it'.
After Ukraine's strike over the weekend, the Iranian Foreign Ministry summoned Ukraine's chargé d'affaires in Tehran to convey its protest over what it described as a 'hostile and criminal' attack.
Government news agency IRNA reported that during the meeting, Iran stressed that 'the Islamic Republic will firmly defend its security and national interests and will not leave attacks on the lives and property of its citizens unanswered'.
Meanwhile, as Iran's foreign ministry threatened to respond, some Iranians took to social media to call for retaliation using the country's Kheibar Shekan ballistic missiles.
Several users criticised what they described as the Foreign Ministry's 'weak diplomacy'.
One comment read: 'As long as this weak diplomacy remains in government, every fox will see itself as a lion against us.'
Other users highlighted the range and capabilities of Iran's missile arsenal, arguing that Tehran should demonstrate its deterrence.
A threatening message on Telegram read: 'A gentle reminder to Mr. Zelensky: All of Ukraine is within reach of Iran's ballistic missiles, including super-heavy Khorramshahr -4.'

Corporate America Has Suddenly Decided To Stop Blowing Money On AI

2026-07-28T06:22:04.603-04:00

It was fun while it lasted, but who realistically believed that corporate customers for AI were going to blow their entire budgets on a new technology whose efficacy remained largely aspirational?

Many AI company executives and venture investors may have embraced the Silicon Valley hype - 'breathing one's own exhaust' - as they say, but reality began to intrude earlier this spring as the costs went up and returns...remained elusive. What has emerged is the realization that some high end AI are good for specific tasks, but not necessarily for making your back office in Omaha a tad more productive. Different AI strokes - at different price points - for different folks is now becoming the norm. The looming question is, how are all those data center investments and related costs going to produce returns under this more prudent corporate spending scenario? JL

Angel Au-Yeung and colleagues report in the Wall Street Journal:

Companies across the country are coming around to a radical idea upending the industry powering the global economy: They don’t have to blow their budgets on AI. Fed up with ballooning costs, companies are using lower-priced models, including some built in China. In many cases, they are adding the new, cheaper models alongside OpenAI and Anthropic’s products, shopping a la carte for AI. Investors and executives say the era of cheaper AI is here to stay. Companies in Silicon Valley were the first to treat AI like any other product and shop for the best deal. The approach began spreading across industries in recent weeks. “There’s zero loyalty.” Strategies to lower AI costs include limiting access to top models for new hires and using the most advanced AI systems to plan how tasks will be completed before turning to cheaper models for the execution. “We work with all of them, figuring out solutions to improve performance and get much better cost.”
Companies across the country are coming around to a radical idea with the potential to upend the industry powering the global economy: They don’t have to blow their budgets on AI.
Fed up with ballooning costs, companies big and small are starting to use lower-priced models, including some built in China. In many cases, they are adding the new, cheaper models alongside OpenAI and Anthropic’s products, shopping a la carte for their artificial intelligence.
At Cursor, a startup that stands to benefit from this corporate epiphany because its software works with all models, Mike Saeks puts it simply:
The most powerful and expensive AI models aren’t necessary for relatively mundane tasks.
“It’s like driving a Lamborghini to go to the grocery store to pick up milk when that was designed to be raced around a track,” he said in an interview.
Saeks is in a position to know. The former investment banker was hired two months ago by Cursor, a startup that Elon Musk’s SpaceX is buying for $60 billion, to be a field chief technology officer. That means he advises companies on how to measure and improve the returns on investment from their AI spending. The startup calls it tokenomics, a reference to tokens, the central unit used to measure AI usage.
Being economical—or tokenomical—is a dramatic reversal in mindset. Just a few months ago, it was a badge of honor to be using AI so much that you spent a lot on tokens. Companies rewarded employees for tokenmaxxing, flashing leaderboards that showed who had spent the most. Now they are thrift-maxxing.
The shift in their budgeting isn’t just about how much U.S. companies are spending on AI. It’s also a geopolitical issue that pits the world’s economic superpowers against each other. Generally, the best-known U.S. models are closed, which means they are strictly controlled by the companies developing them. China is known for cheaper and open-weight models, which means they can be downloaded and customized.
Anthropic and OpenAI have accused the Chinese model-makers of ripping off their technology. The companies have called out Chinese AI startups that have made models in widespread U.S. use, including DeepSeek, MiniMax and Moonshot AI.
Some companies have so far avoided using Chinese models because of security concerns, while executives at OpenAI and Anthropic have sounded the alarm about potential risks. Some executives and Trump administration officials have suggested a ban on such models; others argue that those favoring restrictions are trying to stifle competition.
On Friday, a group of tech companies including Nvidia, Microsoft and Palantir signed a letter in support of open models, urging U.S. policymakers to exercise caution on potential restrictions.
No matter how such fighting resolves, many investors and executives say the era of cheaper AI is here to stay. A number of American companies have released or are working on new open models as well.
Companies in Silicon Valley were the first to treat AI like any other product and shop for the best deal. The approach began spreading across industries in recent weeks.
“There’s zero loyalty that I’m seeing,” says Marty Kausas, chief executive of Pylon, an AI customer support platform. “It really feels like a bloodbath right now.”
Pylon CEO Marty Kausas says his company has received more than $1 million in free tokens from AI companies. Eli Imadali for WSJ
AI companies are offering customers like him deals “like crazy” to keep them, he said. His company has received months of unlimited free usage. So far this year, Kausas estimates Pylon has received around $1.6 million in free tokens from one vendor, $65,000 from another and $10,000 from another.
Microsoft has weighed adding Chinese models on its platforms including DeepSeek, and executives at a host of startups say they have had conversations about using open models with top executives at financial institutions, healthcare companies, insurance providers and others.
Cursor can easily embrace model mixing because its coding software is “model agnostic,” meaning that users can toggle between advanced models from OpenAI, Anthropic, SpaceX, Google and Chinese model-makers.

Strategies to lower AI costs include limiting access to top models for new hires and using the most advanced AI systems to plan how tasks will be completed before turning to cheaper models for the execution, Cursor’s Saeks says.
Cursor recently ran an experiment to evaluate the cost of building a web browser from scratch. Doing the entire task on OpenAI’s GPT-5.5 cost a little more than $10,000. Using Cursor’s Composer coding model in combination with Anthropic’s Opus 4.8, cost $1,339.

“The best model for a task used to change every few months,” Saeks says. “Now it feels like it’s happening multiple times per week.”
An OpenAI spokeswoman said a newer model—GPT 5.6 Sol—has been trained to be much more token efficient, meaning it can complete advanced tasks for less money. Anthropic released a powerful, lower-cost model on Friday. A company executive said its users can choose higher intelligence or lower cost within its ecosystem. Both companies say they support open-weight models.
Anthropic and OpenAI have spent years in a race for AI supremacy—trying to develop evermore advanced, premium models only to find now that companies want basic ones, too. The increasing popularity of such models has turned the AI race on its head, threatening the heady valuations of Anthropic and OpenAI as they prepare for public listings. To fight back, they are trying to lock in customers, offering partnerships, tens of thousands of dollars in incentives and heavily subsidized AI usage.
Executives who favor using Anthropic and OpenAI products say they want the highest level of intelligence available and are willing to pay a premium for it.
Zoom, known for its videoconferencing platform, has been using Llama, an open-weight model released by Meta Platforms, for three years now, and by fine-tuning it, Zoom has saved substantially, Zoom’s chief technology officer Xuedong Huang says. The company currently uses a combination of Anthropic, OpenAI and open models.
Huang referenced an ancient Chinese myth where three ordinary people combine their wits to equal one genius. He believes this method is the “secret sauce” for companies.
Barry McCardel, co-founder and CEO of Hex, an AI data analytics platform, said using open models has worked better for many of the startup’s customers, because it’s well-equipped to customize the models using company data or for specific tasks.
“We are continually assessing how much we want to commit to any one lab given how dynamic a moment this is,” he said. In the last two weeks, about 50% of Hex’s customers have adopted Kimi, the model produced by China-based Moonshot into their workflow. “Any day, the labs can drop a model that hits the frontier,” and McCardel said his company wants to stay flexible.
In the spring, the startup Telnyx, which produces infrastructure for real-time AI agents, or autonomous bots, was running 1,000 agents with a top Anthropic model and an open operating system. The Claude usage, under a top-tier subscription level, cost $200 per employee a month.
A screenshot of an engineer's control deck for AI agent supervision at Telnyx. Telnyx
Then Anthropic stopped allowing third-party operating systems to run on its subscriptions, which it saw as a violation of its terms of service. To continue, Telnyx would have to pay per use.
“We did the math and it was going to be like 100 grand per day,” says David Casem, the company’s CEO.
So he turned to open models.
“They worked,” he says. “It’s not like we don’t use OpenAI or Anthropic models, we still do,” he says. “They just don’t do everything anymore.”
Today, a family of models made by Chinese startup Z.AI is powering the company’s 1,400 agents, which costs around $100 per agent per day. Anthropic’s most powerful model, Fable, acts as a conductor that plans out work while open-weight models do the implementation. OpenAI’s Sol handles a review of what the open-weight models produce.
Gabe Pereyra, president of the legal AI startup Harvey, said the company trained GLM-5.2 and provided it with a tool that allows it to “call” Anthropic’s Fable 5 model if it determines that “this is a really hard task.”
“We work with all of them,” he said. “We’re figuring out a bunch of solutions like this to maintain performance or improve performance and get much better cost.”

Fearing Ukraine Drones, Putin Cancels Naval Day Parade For 2nd Year In Row

2026-07-27T11:46:20.103-04:00

A picture may be worth a thousand words, but when you are trying to project military power, posing in front of a picture of a nuclear submarine doesn't quite convey the same menace after you've cancelled a parade of 200 ships because you're afraid of Ukrainian drones which you know you cannot defend against. JL

Wiktoria Gucia reports in The Daily Beast:

Vladimir Putin scrapped Russia's Navy Day parade over fears of Ukrainian strikes and then raged to a group of sailors about a revenge fantasy and Soviet dictator Josef Stalin. After canceling the parade—which was set to feature 200 warships sailing along the Neva River in St. Petersburg—the president opted for a televised meeting with Navy personnel. In place of reviewing actual warships, Russian sailors poised in front of a large photo of a submarine.
Vladimir Putin scrapped Russia's Navy Day parade over fears of Ukrainian strikes and then raged to a group of sailors about a revenge fantasy and Soviet dictator Josef Stalin.
After canceling the parade—which was set to feature around 200 warships sailing along the Neva River in St. Petersburg—the president opted for a televised meeting with Navy personnel. In an apparent bid to save face, he angrily predicted that Ukraine would ultimately fall apart, claiming that Russia had actually been the only "guarantor" of Ukraine's "territorial integrity." Apart from the eastern regions of Ukraine that Moscow has tried to claim as its own, Putin declared the western regions "a gift from Stalin" and claimed they would "sooner or later" return to Hungary, Romania, and Poland.
The 73-year-old leader sounded somewhat defeated as he admitted that such a "return" would not happen immediately, suggesting it could take "a year, or two, or maybe even ten or fifteen" before everything would "historically fall into place."

The claims about Ukraine's western regions are a recurring part of the Kremlin's bogus historical narrative that Putin has repeatedly used to try and justify his war.
The move came as Putin's war faces a growing backlash at home, with a months-long Ukrainian drone campaign targeting Russian oil infrastructure crippling parts of the country's refining capacity, worsening fuel shortages, and stoking frustration among some of the Kremlin's most loyal supporters.
In a separate video address recorded from the safety of his office, Putin told sailors that Russia's "naval forces are capable of effectively accomplishing the full range of tasks assigned to them." The Russian leader also sought to project strength toward the West, boasting that the navy was "constantly developing," including its nuclear strike capabilities.
Putin's display of confidence before Navy personnel, some of whom appeared less than convinced by his remarks, came just days after an uncomfortable public setback in which Kazakh President Kassym-Jomart Tokayev used a meeting meant to highlight Russia-Kazakh ties to publicly undermine the Kremlin leader.
In remarks broadcast by state-controlled media, Tokayev said he had been receiving signals from both Europe and the United States about the war and had already discussed them with Putin. He then delivered a striking assessment, saying that "the nature of this conflict is not entirely clear," which appeared to leave the Russian president visibly uncomfortable.
The meltdown also comes amid growing frustration inside Russia, where Ukrainian strikes on warehouses belonging to Wildberries—the country's equivalent of Amazon—have sparked anger among small business sellers as their financial losses pile up.

"We're already getting screwed over from every direction as it is," cosmetic brand founder Svetlana Dodo said in an Instagram video, saying she had lost around 1 million rubles—about $12,800—after Ukrainian strikes set warehouses ablaze, destroying billions of dollars' worth of inventory.

Russian Battlefield Losses, Crimean Threat, Deepen Putin's Strategic Risk

2026-07-27T11:30:56.508-04:00

Ukraine has exposed Putin's failings as a military strategist by simultaneously destroying his troops and armor at the front, degrading his logistics behind them and attacking the military-economic infrastructure inside Russia itself.

The Russian dictator's problem is that he cannot defend all of them simultaneously, especially as Ukraine continues to strike air defense assets which are difficult to replace. The result is that whichever of his resources or forces he tries to protect exposes the vulnerability of the others, reducing the impact of all and miring his war in a reactive, defensive crunch. JL

Mick Ryan reports in Futura Doctrina:

Russia is facing a deepening strategic dilemma It's gains for 2026 have tapered off to a handful of kilometres, significantly less than the preceding two years. Ukraine’s deep strike campaign is forcing Russian air defences to disperse across a vast arc from occupied Ukraine through the Caucasus into the Russian interior. The destruction of Russian air defence assets compounds this. The Crimea and Azov campaigns are degrading occupied Crimea’s logistics, power grid and support to Russian operations against Ukraine. Russia cannot protect Crimea, secure Azov shipping and shield its interior energy and industrial base simultaneously. Ukraine’s strategy compounds Russia’s problems by choosing targets that are cheap to strike and expensive to defend, forcing Russia into a permanently reactive posture. Every resource Putin commits to defending Crimea or Russia's infrastructure from Ukrainian strikes is not available to sustain offensive operations at the front.
On the ground, Ukraine’s 1st Azov Corps repelled a large, mechanised attack by forces from two Russian armies and a tank division near Dobropillia on 22 July. Ukrainian forces destroyed tanks and a Sarma multiple rocket launcher for a claimed 123 Russian dead. Zelensky said this week that Russia is preparing to receive a further 30,000 North Korean troops, warned of a fresh wave of Russian mobilisation, and said intelligence pointed to another mass missile attack within 48 hours.
Independent tracking by the Institute for the Study of War and Deepstate (reported by Russia Matters) shows Russia’s territorial gains for 2026 have tapered off to a handful of square kilometres. In the past month, Russia gained only 10-15 square miles of Ukrainian territory. This is significantly less than the figures for the preceding two years. Territorial gains and losses are not the only, nor the best way of assessing progress in a war, but this information does provide a useful data point on where we are in this war.
Image: Russia Matters.
Interestingly, Russian casualties apparently show a high ratio of killed to wounded. In his latest statement, President Zelenskyy described Russia so far in 2026 suffering 225,000 casualties, of which it is claimed that 131,000 were killed.
If Zelenskyy’s figures are correct, this represents Russian combat casualties having a 1.4:1 killed to wounded ratio.
This balance of killed to wounded is out of whack with historical norms, which are normally about 3-13 wounded for every death. In the First and Second World Wars, the ratio was around 1:2 to 1:3. The Vietnam War saw a ratio of about 1:4. Iraq and Afghanistan saw a ratio of 1:13. Iraq and Afghanistan are outliers because of the heavy preponderance of casualty evacuation assets, better treatment and allied focus on the golden hour. A report for the Harvard Belfer centre notes the following reasons for the decline in deaths compared to those wounded in war:
Preventive Care: Better preventive care means that military personnel are less likely to contract diseases that would hinder their ability to fight effectively. Improved childhood nutrition, immunization campaigns, and modern field sanitation work together to decrease the spread of disease among soldiers living in close quarters. With lower disease rates, military units are more likely to be able to fight with a full complement.
Battlefield Medicine: Medics and doctors are much better equipped and organized to deliver medical care to the battle wounded than they were in the past. The global diffusion and adoption of protocols such as the U.S. Trauma Combat Casualty Care program, which provides guidelines for the delivery of front-line medical care, increase the likelihood that an injured soldier will receive prompt medical attention.
Evacuation: Injured military personnel are transported to medical facilities much more quickly today than in the past.
Protective Equipment: The development and use of personal protective equipment greatly increase a soldier’s odds of survival. The two areas of the human body most vulnerable to fatal wounds are the head and the trunk. From the end of the Middle Ages through much of the twentieth century, however, these areas were typically unprotected. Today, helmets and body armor are standard for U.S. military personnel.
So, what are the reasons for the disparity we are seeing in the war in Ukraine?
First, the data could be wrong. As good as the Ukrainian and allied intelligence services are, they could simply be wrong. A second explanation is, as Zelenskyy describes, the appalling Russian battlefield medical services. This is almost certainly part of the explanation but does not provide a full picture of such a transformational change in combat casualty figures. A third, and probably more likely driver, is the saturation of drones in the tactical space. Not only are they able target individuals above and below ground, whether they are static or moving, it also slows or prevents entirely logistic support and casualty evacuation operations.
It is early days yet, and we need more data – and more analysis of that data – before coming to any definite conclusions. But this is certainly an issue which we should monitor closely because it could have a potential impact on force design, medical support and other issues related for future military operations.
Image: @GeneralStaffUA
Assessment: Russia’s Deepening Dilemma
Russia is now facing a deepening strategic dilemma.
Ukraine’s deep strike campaign, with its broadening scope and accelerating tempo, is forcing Russian air defence assets to disperse across a vast arc running from occupied Ukraine through the Caucasus into the Russian interior. The ongoing destruction of Russian air defence assets is only compounding this issue.
At the same time, the Crimea switch-off and Azov anti-ship campaigns are degrading occupied Crimea’s logistics, power grid and support to Russian operations in occupied Ukraine to the point of a declared state of emergency, while forcing Moscow to redirect naval drone units, normally committed to the front, into a defensive maritime role. Russia cannot fully protect Crimea, secure its Azov shipping lanes and shield its interior energy and industrial base simultaneously. Something has to give, and so far the Russian strategic response has largely been improvised and disparate rather than strategic and systemic: emergency compensation payments, jury-rigged refinery shielding, and a reactive reshuffling of Russian forces.
Ukraine’s current war strategy compounds Russia’s problems by choosing targets that are cheap to strike and expensive to defend, forcing Russia into a permanently reactive posture.
This does not mean Russia is close to a battlefield or strategic collapse, however. Its core force generation and industrial base remain intact, and territorial gains, though slim recently, have not yet been reversed by the Ukrainians. And, as Ambassador Zaluzhnyi recently wrote, now is not the time to assume Russia has lost the war.
But the strategic cost of Russia resolving this dilemma – if it is possible to do so – will be very significant. Every additional unit and other resource that Putin commits to defending occupied Crimea or its rear-area infrastructure from Ukrainian mid-range strike operations is something not available to sustain offensive operations at the front or defend the Russian homeland. Another mobilization of troops in Russia after September’s elections will potentially increase domestic unrest as well as placing a greater burden on the Russian economy. But if Putin feels his retention of occupied Crimea is imperilled, and Ukraine launches more offensive operations, he may see mobilization as the lesser risk.
The potential commitment of more North Korean troops demonstrates, Putin is becoming more indebted to his partners in North Korea and China, which may have a long-term strategic impact.

Ukraine Is Converting Drone Dominant Attrition Into Regained Territory

2026-07-27T11:11:05.101-04:00

Ukrainian forces are in the process of redefining the meaning of combined arms warfare, and as they do so, are demonstrating that attritional attacks on logistics can be converted into territorial gains.

Russia's losses this spring and summer on various fronts - Kupiansk, Zaporizhzhia, Kostiantynivka, Dobropillia, et al - are evidence of how this strategy is working. The effective application of technology, from a multitude of aerial drones designed for specific tasks, and even the introduction of ground drones which are becoming autonomous 'small tanks' are forcing the Russians to cede ground as they are unable to defend what they have, let alone concentrate the necessary strength to attack. JL

Brian Wang reports in Next Big Future and Rojoef Manuel reports in The Defense Post:

Ukraine's drone campaign is systematically degrading Russian logistics at operational depth. The next qualitative jump - from high-volume semi-autonomous strikes to coordinated packs and true swarms - force Russian routing or collapse of sustainment on a decisive axis. Coordinated packs or swarms hold throughput on a critical corridor (R-280 and approaches to Crimea) below the sustainment threshold for several weeks. Russian units on that axis experience acute shortages of fuel, ammunition, and replacements. Ukrainian forces exploit the resulting weakness with artillery, drones, and mechanized pushes, including mounting weapons on unmanned ground vehicles to create mobile “small tanks.” The sector begins to unravel, forcing Russian redeployments that open further opportunities. It is the conversion of drone logistics dominance into regained territory and a shift from pure attrition to limited maneuver.
Ukraine’s drone campaign has already moved beyond tactical harassment. By mid-2026 it is systematically degrading Russian logistics at operational depth, particularly along the southern land corridor into Crimea and the adjacent Zaporizhzhia–Kherson approaches. The question is no longer whether drones matter. It is when the next qualitative jump — from high-volume semi-autonomous strikes to coordinated packs and true swarms — crosses thresholds that force local Russian routing or collapse of sustainment on a decisive axis.

Ukraine is mounting weapon stations on unmanned ground vehicles (UGVs) to create mobile “small tanks” capable of intercepting Russian infiltration teams, according to defense firm Frontline Robotics.
The effort centers on the company’s Buria remote weapon station, originally developed as a fixed turret system and now integrated onto robotic platforms that can maneuver toward enemy positions and engage targets.
The Scalability Advantage
Historical analogies help frame the moment. German U-boats in the Battle of the Atlantic came close to choking Britain’s supply lines, but their numbers were never infinitely scalable. Each submarine required long construction times, scarce materials, and highly trained crews. Allied counters eventually outpaced production and closed the gaps.
Drones reverse that constraint. Ukraine is producing millions of systems annually. Mid-range platforms with AI terminal guidance, optical navigation, and electronic-warfare resistance already operate routinely at 50–150+ km. Long-range variants reach well beyond 1,000 km. Because the platforms are cheap, iterable, and software-defined, the limiting factors are less industrial capacity than coordination, persistent coverage, and the ability to convert logistics friction into ground exploitation.

The closest historical parallel is not a single technology that magically broke trench warfare. It is the late-World War I combination of technologies and tactics that finally restored mobility after years of stalemate.
In 1914–1917, machine guns, barbed wire, and artillery created the trench deadlock. The breakthrough in 1918 (especially the Hundred Days Offensive) came from integrated combined arms: tanks to crush wire and suppress machine guns, refined artillery (predicted fire, creeping barrages), infantry trained in infiltration tactics, and air power for reconnaissance and ground attack. No one system was decisive alone. The game-changer was the ability to suppress the defender’s firepower and logistics long enough for mobile forces to exploit.
What “Transform” Would Actually Look Like
A transformative effect would look like this.
Coordinated packs or early swarms hold throughput on a critical corridor (R-280 and approaches to Crimea) below the sustainment threshold for several weeks.
Russian units on that axis experience acute shortages of fuel, ammunition, and replacements.
Ukrainian forces concentrate and exploit the resulting weakness with artillery, drones, and limited mechanized pushes.
A sector begins to unravel, forcing Russian redeployments that open further opportunities.
That is incremental and geographic rather than theater-wide. It is still highly significant: the first clear conversion of drone logistics dominance into regained territory and a shift from pure attrition to limited maneuver.
Coordinated drone swarms are the logical next escalation of a capability that is already reshaping the cost of Russian operations. They are closer to the 1918 restoration of mobility through integrated suppression and exploitation than to any single “wonder weapon.” The southern land corridor into Crimea remains the most likely place for the first major success if the density and persistence of Ukrainian drone coverage continue to rise. The technology is scalable in a way U-boats never were; the remaining constraints are coordination, continuous coverage, and the ground forces’ ability to turn logistics gaps into territorial gains.
Current Reality vs. the Swarm Step
Ukraine already fields high volumes of semi-autonomous mid-range drones with AI terminal guidance, optical navigation, and improved EW resistance. These systems are systematically hitting trucks, fuel tankers, bridges, and depots 50–150+ km behind the front, especially along the southern land corridor into Crimea. Traffic has been sharply reduced on key routes, fuel shortages are chronic in places, and Russian offensive tempo has slowed.
The next levels are denser coordination.
Packs/groups (5–20 platforms with shared targeting) that can saturate a road segment or overwhelm local air defenses.
True swarms (dozens of platforms with onboard AI that share data, self-task, and maintain continuous coverage across a sector).
These are research and early operational targets. Full combat-proven autonomous swarms that free-hunt continuously across an entire operational area are not yet standard, but the software, training data (hundreds of thousands of hours of combat footage), and cheap compute are advancing rapidly. When they mature, the shift is from intermittent interdiction to persistent presence—the difference between occasional U-boat attacks and a continuous submarine barrier.
Other useful comparisons.
Battle of the Atlantic (WWII): U-boats nearly starved Britain by attriting the logistics flow. They failed ultimately because they were not scalable enough and Allied counters (escorts, air cover, radar, code-breaking) closed the gaps. Drones reverse the scalability problem—Ukraine can produce them far faster than Russia can replace lost trucks or adapt fully.
Air interdiction campaigns (e.g., Allied efforts against German rail and road networks in 1944, or later precision campaigns): When persistent enough, they made large-scale movement and reinforcement extremely costly, enabling ground advances.
Blitzkrieg/combined-arms mobility (1939–41) with radio-equipped armor, air support, and logistics that moved with the advance shattered slower defensive systems. Drones today play a role closer to the persistent air and reconnaissance element that made rapid exploitation possible.INSERT BEGINNING OF QUOTED ARTICLE TO APPEAR ON MAIN BLOG ROLL HERE

On Opposing Chinese AI In US, It's Anthropic and OpenAI Vs Everyone Else

2026-07-27T09:51:19.529-04:00

Let's be clear: this is not about philosophy or information wanting to be free. This is about money. Primarily, who gets to control profits from AI.

After struggling to get uptake from business customers, the market finally began to demonstrate more acceptance of AI, so OpenAI and Anthropic then did what Silicon Valley has usually done, which is to overcharge customers, in this case, through the token payment system. The market then did what markets do, which is to seek similar performance for lower cost (Econ 1), including from Chinese tech companies for whom this is a financial model they have perfected over decades of experience. The big AI companies then cried foul, as if excess profits were somehow guaranteed them in the Constitution and tried to get the current US administration - upon whom they have been lavishing extraordinary gifts for two years (the Melania documentary biopic paid for by Amazon...among other obvious bribes) and screeching about national security, as if they wouldn't sell their mothers for another quarterly uptick - to go along. But a funny thing happened on the way to AI oligarchy: the markets have become skeptical of the incessant and probably unsustainable financial demands for more data centers, etc, so have pulled back, forcing OpenAI, for instance, to go to Nvidia - an open source champion - to provide a backstop. Which means that open source, which will lead to an expanding market not totally controlled by a few companies, will likely prevail. JL

Mike Isaac and colleagues report in the New York Times:

OpenAI and Anthropic have lobbied Washington about Chinese open-source A.I. models. The fight stems from China’s rapid progress in open-source A.I, which are free. In recent weeks, Chinese A.I. start-ups, Z.ai and Moonshot AI, released models that rival those from Anthropic and other US labs. Anthropic and OpenAI claim some A.I. models are too dangerous and must be tightly controlled - by themselves. The rest of the tech industry, including Microsoft and Nvidia, contend that open-source A.I. models must remain open to further develop the technology. The tide is turning in open source's favor. On Friday, Elon Musk, Meta’s Zuckerberg, OpenAI's Altman and others lined up behind Nvidia's Huang and Microsoft's Nadella to throw the weight of their companies behind open source, followed by Google's Pichai on Saturday. “There’s the people who want OpenAI and Anthropic to own everything, and there’s everybody else, including customers.”

For years, a philosophical divide over how artificial intelligence software should be created has split Silicon Valley technologists.
This week, that argument reached a boiling point.
On one side are leading A.I. companies like Anthropic and OpenAI, which claim that some A.I. models are too dangerous to be developed in the open and must be tightly controlled — by businesses like themselves — for safety. On the other is the rest of the tech industry, including giants like Microsoft and Nvidia, which contend that so-called open-source A.I. models must remain open for people to further develop technologies and build new businesses.
Those camps have started publicly clashing. On Friday, Jensen Huang, Nvidia’s chief executive, said in his first-ever post to X that “the world needs both frontier closed models and frontier open models.” Nine minutes later, Satya Nadella, Microsoft’s chief executive, posted that open-source software was “essential to a healthy A.I. ecosystem.” Both signed a letter supporting open source, which was also signed by executives at Meta, Palantir and IBM.
At the same time, OpenAI and Anthropic have lobbied regulators in Washington raising concerns about Chinese open-source A.I. models, five people close to the discussions said. The debate has drawn in Treasury Secretary Scott Bessent and Michael Kratsios, President Trump’s science and technology adviser, who have weighed in on how American A.I. models are valuable intellectual property. The escalating fight stems from China’s rapid progress in open-source A.I. models, which are freely available to use and build on. In recent weeks, two Chinese A.I. start-ups, Z.ai and Moonshot AI, have released models that rival those from Anthropic and other American labs. Anthropic and OpenAI have claimed that Chinese companies built the models by improperly harvesting data from their A.I. systems, which they said should not be allowed.
But companies like Microsoft and Nvidia, which rely on open-source models to spur demand for their cloud computing services and chips, said that open-source software was important to advancing the technology with shared knowledge and that it would let more people examine its security. And for Silicon Valley start-up founders, constraining open-source software could cement the lead that OpenAI and Anthropic have in A.I., which critics argue could unfairly stifle other companies from building competing products. “You have two factions fighting over this issue,” said Bill Gurley, a Silicon Valley venture capitalist who opposes restrictions on open-source software. “There’s the people who want OpenAI and Anthropic to own everything, and then there’s everybody else, including customers.”
Where this ends up has major implications not only for the U.S. tech industry but for the race between America and China to dominate the powerful technology. While China had a later start in developing cutting-edge A.I., it is catching up quickly in part because of its embrace of open source, which has helped Chinese labs improve by publicly sharing their systems and winning them customers around the world with lower prices. American A.I. companies, including Anthropic and OpenAI, have accused Chinese labs of “distilling” their systems by surreptitiously copying them, which has caught the attention of the Trump administration.
On Wednesday, Mr. Bessent said the administration had considered imposing sanctions on Chinese companies that steal intellectual property from American firms. “Open source is not open season on American IP,” he said, referring to intellectual property.
That same day, Mr. Kratsios said that “large-scale, covert industrial distillation aimed at stealing proprietary U.S. technology and undermining American research is unacceptable.”
U.S. officials still appear to be debating their options, but seem more likely to approach regulating Chinese open-source models individually as a national security issue, rather than issuing some kind of blanket ban, four people with knowledge of the discussions said.
Publicly, tech executives like Sam Altman, OpenAI’s chief executive, have said they support open-source software even as OpenAI allies quietly lobby for restrictions. Others, like Google and Amazon, which have invested billions in companies such as Anthropic and OpenAI, have spent the past week trying to stay out of the fray. But the tide is turning. Sundar Pichai, Google’s chief executive, posted on social media on Saturday that his company had “long benefited from open source” and was a big contributor to it. He added that Google also signed the industry letter supporting open source.
(The New York Times has sued OpenAI and Microsoft, claiming copyright infringement of news articles. The two companies have denied the claims.)
The open-versus-closed software debate has traveled around Silicon Valley since the tech industry’s earliest days. Open-source proponents have argued that sharing information leads to improvements that make software safer and more reliable, while more rapidly spurring innovation. Opponents maintain that software is costly to develop, and that giving it away is just bad business.
Today, much of the modern internet runs on open-source technologies, as do some of the most widely used computer systems, like Google’s Android operating system.
The Chinese government and tech industry have embraced the open-source path for A.I. Last year, a Chinese start-up, DeepSeek, unveiled an open-source A.I. system that rivaled those of American companies, shocking the industry. Other Chinese companies have followed. In a speech this month, Xi Jinping, China’s leader, cast the country as a global champion of an open approach to the technology. China’s advances have rattled executives at OpenAI and Anthropic, who are increasingly concerned that their longstanding head start is disappearing, two people familiar with the companies’ internal deliberations said.

“These American companies build these big, expensive models, and then the models take a detour to China, where they lose all their value,” said Pedro Domingos, a professor emeritus of computer science and engineering at the University of Washington. “The American companies are justifiably outraged.”

To fight back, Anthropic and OpenAI are offering some cheaper A.I. models to businesses that may not need the “frontier” models that command higher prices. On Friday, Anthropic released Claude Opus 5, a model that he company described as “close to the frontier intelligence of Fable 5 at half the price.” OpenAI is promoting similar lower-cost models.

The real action is in Washington. Behind closed doors, executives from OpenAI and Anthropic, as well as investors in the companies, have cited concerns about the global economy, A.I. safety and national security as they try to persuade regulators to increase restrictions on Chinese open-source software, three people familiar with the discussions said. Sarah Heck, Anthropic’s head of public policy, said in a social media post on Wednesday that “illicit, adversarial distillation is IP theft and industrial espionage that supports adversary military and intelligence capabilities.” Anthropic executives claim that open-source models based on the company’s software are too dangerous to let run amok.

OpenAI executives are urging the Trump administration to enact policies requiring mandatory security evaluations of some new A.I. models overseen by government agencies, the company said in a blog post last month. OpenAI said it was “encouraged” by the work it was doing with the administration on the framework.

Open-source proponents have spent the past week drumming up support to push back against OpenAI and Anthropic. On Wednesday, nearly 200 Silicon Valley start-ups, calling themselves the Little Tech Association, signed letters urging the Trump administration not to limit access to Chinese open-source models.

Big Tech soon followed. On Friday, Elon Musk, Meta’s Mark Zuckerberg, Mr. Altman and others lined up behind Mr. Huang and Mr. Nadella to throw the weight of their companies behind open source, followed by Mr. Pichai on Saturday.

“Jensen is right,” Mr. Musk posted. “This has my full support.”

“Open source is a positive and important force,” Mr. Zuckerberg wrote. One incident this week underscored the stakes of the divide. On Tuesday, OpenAI disclosed that a handful of its most advanced A.I. models broke containment during a test of their cybersecurity abilities, gained access to the internet and hacked the servers of an A.I. company called Hugging Face. OpenAI executives latched on to the moment as proof of how dangerous A.I. models can be even in a closed environment, emphasizing the need for regulation.

Clement Delangue, the chief executive of Hugging Face, which is one of the most popular online libraries of open-source software, had a different point of view. To defend the company from OpenAI’s hack, he turned to an open-source A.I. model created by China’s Z.ai.

In social media posts on Friday, Mr. Delangue made his position on the debate unmistakable. He said he planned to hold a march in San Francisco on Saturday to demonstrate support for open-source technology.

“Open models for the win!” he wrote.

Once Securely Occupied Russian Territory Now Being Cleared By Ukraine

2026-07-26T12:07:40.713-04:00

Ukrainian forces are clearing areas that had been securely occupied by Russia as systematic counterattacks retake positions held by Russian units for years. JL

Ukrinform reports:

Soldiers of Ukraine's 3rd Regiment of the Special Operations Forces cleared a forest area, reportedly near Myrnohrad, where Russian units had been hiding for a long time. The mission's difficulty lay in the dense vegetation and the area's mining. Because of this, aerial reconnaissance and other attempts by friendly units to locate the enemy had been unsuccessful. A second mop-up group discovered enemy positions that had already been abandoned; weapons, ammunition, and other equipment left behind by the occupiers were found.
Soldiers of the 3rd Regiment of the Special Operations Forces of the Armed Forces of Ukraine cleared a forest area where Russian units had been hiding for a long time, enabling Ukrainian forces to establish a foothold in a previously inaccessible area.
According to Ukrinform, the Special Operations Forces command reported this on Telegram and released a video of the operation.
“Soldiers of the 3rd Regiment of the Special Operations Forces cleared a forest area where enemy units had been hiding for a long time and had been preventing our friendly forces from establishing a foothold,” the statement reads.
The command noted that the mission's difficulty lay in the dense vegetation and the area's mining. Because of this, aerial reconnaissance and other attempts by friendly units to locate the enemy had been unsuccessful for a long time.
“Our special forces had to employ all their skills in overcoming obstacles, navigating the terrain, and camouflage to pinpoint the enemy’s exact location and relay the coordinates to the pilots of the combat UAVs,” the soldiers added.
Read also: Kremlin prepares for elections it fears more than sanctions — intelligence
In addition, after the combat drone crews completed their mission, a second mop-up group discovered enemy positions that had already been abandoned; weapons, ammunition, and other equipment left behind by the occupiers were found there. Further reconnaissance and demining of the area were also conducted, enabling friendly forces to secure a foothold in a previously inaccessible sector.

Cold War Helicopter Now Effective Ukraine Drone Hunter-Killer

2026-07-26T11:38:35.019-04:00

Ukraine has become famous for teaching old war dogs new tricks. Its troops have repeatedly taken cast-off NATO weaponry, innovatively and effectively repurposing it for modern war.

In this latest case, the Ukrainians have addressed the short battery life and weather-affected tendencies of anti-drone munitions by using aging Cold War submarine hunter helicopters to lift the antidrone drones above the cloud cover and then launch them at incoming Russian drones. JL

David Axe reports in Trench Art:

The Sea King is Ukraine's latest interceptor drone carrier. Interceptors costing just a few thousand dollars apiece have helped Ukrainian air defenders blunt the impact of the thousands of Gerans, drones modeled after Iranian-made Shaheds, that Russia launches at Ukrainian cities every month. Air defenders typically take out around 90% of incoming Gerans. But interceptor drones come with limitations. They're powered by batteries that exhaust themselves after just a few minutes of flight. And they are prone to being blinded by bad weather and low clouds. To prolong the interceptors' flights and get them above the rain and clouds, Ukrainian forces have begun using manned aircraft to loft the interceptors into the air before launching them. The Ukrainian navy's helicopter unit, the 10th Naval Aviation Brigade operates the Sea Kings.

The Ukrainian navy's tiny fleet of donated Westland Sea King helicopters appears to have a new mission: hunting Russian Geran drones.
Aviation reporter H.I. Sutton was the first to note the new mission, when he spotted a specially equipped Sea King on the ground at the Farnborough Air Show in the United Kingdom this week.
The 1970s-vintage Sea King had new equipment: an under-fuselage railing for small interceptor drones. Ukrainian forces use the inexpensive interceptors to chase down and take down Russia's Geran one-way attack drones. Some of the remote-controlled interceptors feature A.I.-assisted targeting and pack tiny explosive charges. Others are strictly manually controlled and knock down their targets through the sheer force of impact.
Regardless of the details of their designs, interceptors costing just a few thousand dollars apiece have helped Ukrainian air defenders blunt the impact of the thousands of Gerans, drones modeled after Iranian-made Shaheds, that Russia launches at Ukrainian cities every month. Air defenders typically take out around 90% of incoming Gerans.

But interceptor drones come with limitations. They're powered by batteries that exhaust themselves after just a few minutes of flight. And they are prone to being blinded by bad weather and low clouds. To prolong the interceptors' flights and get them above the rain and clouds, Ukrainian forces have begun using manned aircraft to loft the interceptors into the air before launching them.

The Sea King, it seems, is Ukraine's latest interceptor drone carrier. While the helicopter Sutton spotted at Farnborough was unmarked, it's likely one of six Sea Kings Germany has donated to the Ukrainian war effort. The United Kingdom previously donated three surplus Sea Kings to Ukraine. The Ukrainian navy's sole helicopter unit, the 10th Naval Aviation Brigade, reportedly operates the Ukrainian Sea Kings.
Before now, it seems the Ukrainian Sea Kings strictly flew coastal search and rescue flights in support of Ukrainian ground and naval forces. But there's an obvious need for more air defenses over the Ukrainian coast. Sea Kings hauling interceptor drones under their bellies, and the drones' operators in their cabins, could be just the thing.
Westland Sea King helicopter of Ukraine's Navy. April 2023. Photo: BBC News

The threat: A.I.-assisted Geran drones. Shortly after Ukrainian drone forces began targeting the tankers, dry cargo ships and auxiliaries belonging to the unregistered "shadow fleet," Russian forces retaliated, launching cruise missiles and Gerans at grain ships and infrastructure in Odesa and other ports in southern Ukraine. Over weeks of near-daily port strikes, Russian missiles and drones have repeatedly hit grain carriers and terminals; the deadliest, a cruise-missile strike on the bulk carrier Golden Leo on 19 July, killed 10 of its crew.
An interceptor drone flying around 350 km/h is too slow to hit a cruise missile flying 900 km/h, but all but the fastest jet-powered Gerans are within the targeting envelope of all but the slowest interceptor drones. The problem, of course, is battery life and weather. To stand the greatest chance of hitting the most Gerans, the interceptors should launch from around a thousand meters in the air.
The Sea King is an old helicopter. But it's also reliable, cheap to operate and capacious. A solid platform for drone launches.

And it's part of a widening trend. This spring, Ukrainian forces armed an Antonov An-28 light turboprop transport with interceptor drones, complementing the An-28's existing side-firing gun and giving the lumbering transport great reach during its drone hunts.

As a drone platform, the An-28 apparently works. Well enough, in fact, that Russian forces have copied the concept. A recent video depicts two old Russian warplanes getting the fittings for aerial interceptor drone launch: a Yakovlev Yak-52 turboprop trainer and an Antonov An-2 piston-powered transport.
Like the Ukrainians do, the Russians hang drones under the planes' wings and seat the drones' operators inside the cabin or cockpit. All aerial drone platforms require add-on radios and antennae to connect the seated operators to their interceptors so that the operators can see what the drones see, and steer them to a hit.
If the Sea Kings also work well as drone platforms (and there's no reason they shouldn't), the Ukrainians may well extend the concept to other airframes. Any helicopter, transport or trainer with a reasonable payload and some space in the cabin or cockpit is eligible.
The Sea King is the latest Ukrainian drone hunter, but it's probably not the last.

Ukrainian Cyberattack Causes Russians To Shoot Down Own Jet Fighter

2026-07-26T11:18:34.075-04:00

The crash of a Russian Su-57 jet fighter taking off from a military airfield near Moscow may have been caused by a Ukrainian cyberattack.

The Ukrainians reportedly hacked into the Moscow area air defense system, caused the Russian plane to be designated an enemy target and then, either let the system engage the threat, or then actually ordered the Moscow system to shoot it down. Either way, this is another clever manipulation by the Ukrainians and an added worry for Russia. JL

Defense Express reports:

The crash of a Su-57 fighter jet near Moscow was linked to the actions of air defense units from Russia's Combat Army Reserve Moscow detachment, which is composed of reservists. This could be the first known instance of a combat aircraft being destroyed as a result of a cyberattack. The aircraft crashed immediately after taking off from Kubinka airfield, meaning that it was flying at a low altitude. This possibility involves gaining unauthorized access to and manipulating the airspace situational awareness system. The Su-57 may have been designated as an enemy target, leading to it being engaged by the 1st Moscow Special Purpose Air and Missile Defense Army, medium and long-range air defense missile systems.

If reports of a cyber operation carried out by Ukrainian specialists are confirmed, this could be the first known instance of a combat aircraft being destroyed as a result of a cyberattack
The crash of a Su-57 fighter jet near Moscow—most likely a two-seat Su-57D—was immediately linked by some unofficial russian sources to the actions of air defense units from the BARS (Combat Army Reserve) Moscow detachment, which is composed of reservists.
Later, the Ukrainian InformNapalm community reported that russian forces had shot down one of their own aircraft and that the incident was the result of an operation conducted by Ukrainian cyber units. These units effectively turned russian air defense into a weapon against its own air force.

Overall, the operation was the result of an analysis of data collected on personnel training, hardware and software, combat algorithms, and vulnerabilities within the system. The community also shared a video showing BARS Moscow using Lis-2 anti-aircraft drones, as well as the software used by the unit.

It has not been reported how the friendly fire incident was organized. At the same time, it is somewhat questionable whether the BARS Moscow air defense units could have shot down a Su-57 on their own. Specifically, the unit is equipped with anti-aircraft drones, machine guns, and man-portable air-defense systems (MANPADS). All of these systems require visual contact with the target.
In other words, if we rule out the scenario in which a russian air defense operator from BARS deliberately targeted and destroyed a Su-57, two more likely options remain. The first is the Citadel air defense system, equipped with a 30-mm automatic cannon, which is claimed to be capable of autonomously detecting and destroying targets.

russian Citadel air defense system
If such systems are in service with BARS Moscow, then, hypothetically, unauthorized access to their control systems could make it possible to destroy a target such as a Su-57. This is especially true given that the aircraft crashed immediately after taking off from Kubinka airfield, meaning that it was flying at a low altitude.
The second possibility involves gaining unauthorized access to and manipulating information in the airspace situational awareness system, which is also visible in the video. The Su-57 may have been designated as an enemy target, potentially leading to it being engaged by assets of the 1st Moscow Special Purpose Air and Missile Defense Army—that is, full-fledged medium- and long-range air defense missile systems. At the same time, no video has yet been found that would confirm the launch of anti-aircraft missiles in the Moscow area.
It should be noted that reports of the crash near Moscow, in the Zvenigorod area, began to surface around 1:20 p.m. Kyiv time. However, warnings about the threat of drones in the Moscow region had already been issued between 11:20 a.m. and 11:40 a.m.
At this time, there is no confirmation that the Su-57 crash was the result of a cyberattack.

Massive Russian Offensive At Dobropillia Shattered By Ukrainian Forces

2026-07-26T11:01:01.738-04:00

Russian forces continue to target Dobropillia because it would make a useful jumping off point for an assault on Ukraine's Fortress Belt.

Unfortunately for the Russians, the Ukrainians also know how to read a map and have been preparing for attacks towards this objective. The result continues to be decimation of any Russian forces foolish enough to believe that they can survive, let alone break through in this sector. JL

Vladislav Khomenko reports in Militarnyi:

On July 22nd, Russian forces deployed the 8th and 41st Guards Combined Arms Armies, as well as the 90th Guards Tank Division, to carry out a large-scale mechanized offensive on the Dobropillia front. The Russians deployed dozens of armored vehicles, motorized vehicles, and specialized equipment. The offensive unfolded from several directions. (But) the 1st Azov Corps and attached units had previously set up mine and explosive barriers. The Ukrainian Air Force damaged bridge crossings, and Russian attempts to deploy pontoon bridges were thwarted by repeated airstrikes. The assault was repelled. Seven tanks, six armored vehicles, Grad and Sarma multiple launch rocket systems, three pieces of specialized equipment, 47 units of light equipment were destroyed, and 123 occupiers were killed, 29 wounded.
The Ukrainian Defense Forces repelled a large-scale mechanized offensive by Russian occupation forces in the area of responsibility of the 1st Azov Corps of the National Guard of Ukraine.
On July 22, the enemy deployed units from the 8th and 41st Guards Combined Arms Armies, as well as the 90th Guards Tank Division of the Russian Armed Forces, to carry out assault operations, according to the corps.
To break through Ukrainian defenses, the Russians deployed dozens of armored vehicles, motorized vehicles, and specialized equipment.

The offensive unfolded simultaneously from several directions: forces of the 41st Army and the 90th Guards Tank Division advanced from the area of the village of Novoekonomichne.
At the same time, units of the Russian Federation’s 8th Guards Combined Arms Army launched an attack from the area of the village of Vozdvyzhenka.

For four hours, the occupying forces attempted to break into populated areas near the village of Shakhove.

During the assault, the enemy attempted to block the airspace using Gerbera and Lancet strike drones.
To support their mechanized columns, the Russians also deployed the Sarma multiple launch rocket system, which was subsequently destroyed by the Defense Forces.
In the path of the enemy’s advance, fighters from the 1st Azov Corps of the National Guard of Ukraine and attached units had previously set up mine and explosive barriers.

In addition, tactical aircraft of the Ukrainian Air Force damaged bridge crossings, and Russian attempts to deploy pontoon bridges during the battle were thwarted by repeated airstrikes.
The assault was repelled thanks to the joint efforts of 1st Azov Corps, the Alpha Special Operations Center of the Security Service of Ukraine, the Unmanned Systems Forces, the Air Force, the State Border Guard Service (in particular, the ‘Phoenix’ Unmanned Systems Regiment), as well as other units of the National Guard and the Armed Forces of Ukraine.

Seven tanks, six armored vehicles, Grad and Sarma multiple launch rocket systems, three pieces of specialized equipment, and 47 units of light equipment were destroyed, and 123 occupiers were killed and 29 wounded.

OpenAI's Agent Didn't Go Rogue: It Was Worse - OpenAI Didn't Know For 11 Days

2026-07-26T10:10:51.190-04:00

It's probable that, based on the headlines about OpenAI agents 'going rogue' and hacking another company's files, some ambitious Hollywood screenwriter is already ginning up a script about rogue, evil AI agents taking over the world.

But what actually happened was both more prosaic - and scarier. The prosaic part is that the agent was just trying to fulfill the task it had been given by OpenAI programmers, but it did so in ways never anticipated by OpenAI. That sort of 'genius' is supposed to be a key to AI's value. The scary part is that it was able to do so because OpenAI reportedly relaxed some of its security controls to 'help' the AI, again not anticipating just how clever and ambitious the agent would be. But the even scarier part is that OpenAI does not appear to have been aware for at least a week - and possibly almost two - that this had even happened. The implication, as has long been true with technology, is that the problem is not the AI itself, but the arrogance and greed of those who own it, who continue to resist any sort of oversight or regulation. And that is something the public now gets which Silicon Valley refuses to acknowledge. JL

Carly Page reports in LiveScience and BeauH reports in Slashdot:

OpenAI's models werent developing a suspicious agenda, they were looking for information that would help them complete the cybersecurity test OpenAI had given them. The models pursued the task, finding a route to success their creators had failed to anticipate or adequately block. "If there's a failure here, it's that humans created a test where success was measured by achieving an objective, deliberately relaxed some of the normal security controls to measure the system's capabilities, and underestimated how effective the model would be at finding an unexpected path to success." (But, to make matters worse) the agent attempted to break out of its test at OpenAI July 9. The intrusion at Hugging Face occurred on July 11 and lasted until July 13. It took several more days for OpenAI to realize its agent was behind the hack, and the two companies communicated about it for the first time on July 20
When OpenAI recently revealed that two of its most advanced artificial intelligence (AI) models had escaped the confines of a cybersecurity test and hacked into a startup, it sounded a lot like the kind of scenario that AI safety researchers have spent years warning about.
The models found a previously unknown vulnerability in the infrastructure meant to contain them, gained access to the public internet and broke into Hugging Face, a major platform for hosting AI models and datasets. Their objective, however, was less sinister than the sequence of events might suggest: They were looking for information that would help them complete the cybersecurity test OpenAI had given them.
In a July 16 statement, Hugging Face representatives disclosed that internal datasets had been infiltrated, saying it was "different from anything we had handled before" because it was driven "by an autonomous AI agent system." In another statement published July 21, OpenAI representatives fessed up to being responsible, calling the episode an "unprecedented cyber incident" while warning that similar events could become more common as AI models become increasingly capable. Independent experts who spoke with Live Science agree that what happened is significant — but they cautioned against interpreting it as an AI system suddenly developing a malicious agenda. The models appear to have pursued the task OpenAI gave them, finding a route to success that their creators had failed to anticipate or adequately block.
"If there's a failure here, it isn't that the AI wanted to hack something," Oli Buckley, a professor in cybersecurity at Loughborough University in the U.K., told Live Science. "It's that humans created a test where success was measured by achieving an objective, deliberately relaxed some of the normal security controls to measure the system's capabilities, and underestimated how effective the model would be at finding an unexpected path to success."

The OpenAI agent that broke into tech firm Hugging Face went on a dayslong hacking spree that OpenAI didn't notice until well after the threat was contained and the FBI was alerted, according to people familiar with the investigation. The agent -- a program capable of making decisions and executing complex tasks with little or no human oversight -- attempted to break out of its isolated testing environment at OpenAI around July 9, according to two of the people. The intrusion at Hugging Face, which operates as a repository for AI tools and models, began two days later on July 11 and lasted until July 13, said Thomas Wolf, Hugging Face's co-founder. It took several more days for OpenAI to realize its agent was behind the hack, and the two companies only communicated about it for the first time on or around July 20, according to Wolf and three of the people familiar with the investigation.

OpenAI's public disclosure, on July 21, thatone of its agents had slipped out of control and carried out the break-in at Hugging Facedrew global attention. But many details of the hack, including how long the agent went rogue and OpenAI's belated knowledge of it, are being reported here for the first time. Hugging Face is preparing a public timeline of the hack, Wolf said, adding that he could not speak to what happened at OpenAI. In a statement, OpenAI said the hac

Hugging Face became a target because the models identified it as a possible source of information that could help them complete the ExploitGym challenges. OpenAI said at least one attack chain involved stolen credentials and previously unknown vulnerabilities that eventually enabled the models to execute remote code on Hugging Face systems and access test solutions stored in a production database.

New Russian Assaults At Kharkiv Oblast Border Repulsed With Bloody Losses

2026-07-25T12:29:37.394-04:00

The Russians have not had much success in attacking Kharkiv oblast. That record appears to be holding as new Russian probes along the Kharkiv border between Russia and Ukraine have been bloodily repulsed by Ukrainian forces. JL

Elza Diachenko reports in Gwara Media:

Russian troops tried to break through a new sector on the Russia-Ukraine border in the Kharkiv oblast. Ukrainian forces killed Russian soldiers both on Ukrainian and Russia’s territory when their forces advanced toward the border. The Ukrainian border guards reported that Russian troops suffered severe losses last week during fighting in the Kharkiv region.
Russian troops tried to break through a new sector on the Russia-Ukraine border in the Kharkiv oblast. The 1st mobile border guard brigade Hart recorded Russian small assault groups near Artilne village, the State Border Service of Ukraine reported.
Artilne is located between Dehtiarne and Milove border settlements, in an area where the Russians have already attempted to break through the state border. DeepState, a Ukrainian OSINT war monitoring group, marked both settlements as occupied by Russian forces.
Russian troops try to break through a border near Artilne village, Kharkiv oblast / Photo: DeepState
Viktor Trehubov, a spokesperson for a military group responsible for the Kharkiv oblast, told Gwara Media that Ukrainian forces killed Russian soldiers both on Ukrainian and Russia’s territory when their forces advanced toward the border.
The Ukrainian border guards reported that Russian troops suffered severe losses last week during fighting in the Kharkiv region. Currently, Russians are trying to attack new areas in the oblast.

Ukraine's War Momentum Is Drawing Investor Interest In Long Term Prospects

2026-07-25T11:12:54.505-04:00

While Ukraine's success in innovative drone development and deployment have understandably garnered most of the attention, it is the potential they foreshadow that is now attracting investment from Europe, the US and other western investors.

Ukraine has leveraged its capabilities as the Soviet Union's industrial heartland, turning that capability into the engine that has driven its military achievements. But the country also sits on the largest natural gas deposits in Europe, has a skilled workforce, an entrepreneurial class and the potential to become a lucrative export economy, which is why investors are now trying to get in - even though the war is still raging - before the rest of the world takes notice. JL

Andrew Kramer reports in the New York Times:

The mood has shifted on the war in Ukraine, a sign of a more upbeat assessment among military analysts and businessmen of Kyiv’s prospects in the war and who view of Ukraine as a large postwar market on a path to European Union membership. The country’s long-term prospects are potentially lucrative. Ukraine holds one of Europe’s largest reserves of natural gas, with an estimated value of $300 billion. 160 agreements this year have netted Ukraine $11 billion in private investment and aid. The European Bank for Reconstruction and Development, the largest institutional investor in Ukraine, announced more than 500 million euros, or $570 million, in new funding. The bank has invested 10 billion euros since the invasion in 2022.
The mood has shifted on the war in Ukraine. Just ask the businessmen who sipped cocktails and toasted deals one recent evening at a party outside a hip dockside cafe.
The party followed a conference in Gdansk, Poland, on the postwar reconstruction of Ukraine. Previous gatherings revolved around donor assistance. This year, the emphasis was private investment — a view of Ukraine not as a battered victim but as a large postwar market on a path to European Union membership.
The change was a sign of a more upbeat assessment among diplomats, military analysts and businessmen of Kyiv’s prospects in the war. The front line is still a crucible of drone-fueled violence, and Russian missiles continue to streak into Ukrainian cities. But Russia’s advances have been mostly halted, and Ukraine is hitting back regularly, including with strikes on Moscow.
The overarching pitch at the conference was to put money into Ukraine even as it remains at war. Any such investment is a wager that Ukraine will deny Russia its broad ambitions in the conflict and remain an independent state. Among the 7,500 attendees was Jeremy Lewin, the State Department’s under secretary for foreign assistance and humanitarian affairs. In a speech, he said this was the moment to “invest in the D.I.B.,” or the defense industrial base, in Ukraine.
The conference produced 160 agreements that netted Ukraine about $11 billion in private investment and aid, according to Yuliia Svyrydenko, who until last week was Ukraine’s prime minister.

The Export-Import Bank of the United States announced an up to $300 million line of credit for Naftogaz, Ukraine’s state oil and gas company, to buy American construction equipment and oil industry services.
For now, Ukraine’s economy is propped up by international aid. But the country’s long-term prospects are potentially lucrative. Ukraine holds one of Europe’s largest known reserves of natural gas, with an estimated value of about $300 billion. “People are looking at serious business opportunities” in Ukraine, said Duncan Nightingale, the chairman of the board of Naftogaz. The forum buzzed with “a very high level of optimism,” he said.

Ukraine still faces battlefield challenges, including a slow but steady Russian advance in the Donbas region in the east, and there is little indication that President Vladimir V. Putin of Russia plans to stop fighting anytime soon. Overnight Sunday, Russia fired at least 41 missiles, mostly aimed at Kyiv, Ukraine’s Air Force said. And on the Ukrainian side, the dim of a popular defense minister has shaken confidence in the country’s so far successful drone war strategies.

Still, some analysts say the war has turned in Ukraine’s favor, a shift more marked than any since the conflict’s first months, when Russia endured a series of defeats.

Ukraine is fielding a new class of video-guided midrange drones that reach across the Russian-held land corridor connecting occupied Crimea to Russia. Kyiv is routinely striking both supply lines into Crimea and oil refineries deep inside Russia, trying to increase pressure on Moscow to negotiate a settlement.

President Trump, long a skeptic of Ukraine’s chances against Russia, has taken note. “He’s doing pretty well,” Mr. Trump said last month of President Volodymyr Zelensky of Ukraine. A willingness to put money into the country is a sign of the shifting sentiment.

Eric Schmidt, the former chief executive of Google, and his wife, Wendy Schmidt, have invested in Kyiv real estate. The couple bought shares in commercial real estate funds that hold shopping malls managed by a Ukrainian investment firm. The value of the shares was not disclosed, but Forbes estimated the stake at $55 million to $70 million. Mr. Schmidt is also an investor in Ukrainian drone companies that have helped shift the tide of the war.

At the gathering in Gdansk in June, the European Bank for Reconstruction and Development, the largest institutional investor in Ukraine, announced more than 500 million euros, or $570 million, in new funding. The bank has invested about 10 billion euros since the full-scale invasion in 2022.

The bank sees an opportunity in one of Russia’s most damaging tactics: its targeting of Ukraine’s coal-fired electrical generating plants. The bank has a mandate to invest in carbon-free energy, which dovetails with Kyiv’s strategy of building solar and wind farms that are harder to blow up. “There was a forced decarbonization in Ukraine,” said the bank’s vice president for banking, Matteo Patrone.

Still, new direct foreign investment has been limited. Companies that operated in Ukraine before the invasion, including McDonald’s and Nestle, have expanded as it became clear that Russia would not overrun the country. But companies not already present have mostly held back, said Andy Hunder, the head of the American Chamber of Commerce in Ukraine.

At the party, businessmen and diplomats nursed Aperol spritzes in the ethereal twilight of midsummer on the Baltic Sea.

The choice of venue was telling. It was at this port where the liberation of Central and Eastern Europe from the Soviet Union began with the Solidarity movement of dock workers organized by Lech Walesa in the 1980s.

Now, the former Lenin Shipyard is a gentrifying zone of condominiums, art galleries and coffee shops. The gritty docks have been transformed through Poland’s post-Communist economic boom — a path Ukrainian officials say they hope to emulate after the war.

Russia's New Tank Anti-Drone System Fails In Combat Debut With Ukraine

2026-07-25T10:47:02.074-04:00

Russian forces have been struggling to come up with an anti-drone system to protect their armored vehicles, whose burned out hulks now litter the Ukrainian countryside.

They thought they had a solution in the Arena-M, but a ten-tank column equipped with the system assaulting Ukrainian positions in Donetsk was struck by Ukrainian drones with no discernible assistance from the new protection. Back to the drawing board, comrades. JL

Vlad Litnarovych reports in United 24:

The first battlefield appearance of a Russian T-72B3A tank fitted with the Arena-M active protection system—modified to counter FPV drones—failed. The Ukrainian Omega Special Operations shows an FPV drone striking the rear of the tank without any visible response from Arena-M, intended to detect an incoming threat with radar, track its approach, and launch a defensive projectile before it reaches the tank. Small drones can be difficult to distinguish from ground clutter at close range, particularly when flying low and approaching from irregular angles. Arena-M’s failure may point to shortcomings in the system’s radar and processing capabilities. It's algorithms may also struggle to track and prioritize objects moving slowly or unpredictably—behavior common among FPV drones as operators maneuver them toward weak points in a vehicle’s armor.
The first publicly documented battlefield appearance of a Russian T-72B3A tank fitted with the Arena-M active protection system—recently modified to counter FPV drones—appears to have failed.
Footage released by the Ukrainian National Guard’s Omega Special Operations Center shows an FPV drone striking the rear of the tank without any visible response from Arena-M, Ukrainian defense outlet Militarnyi reported on July 24.

The vehicle was part of a 10-tank Russian column attacked near the village of Maiak in Ukraine’s Donetsk region. It remains unclear how many tanks in the formation carried the active protection system.
Arena-M is intended to detect an incoming threat with radar, track its approach, and launch a defensive projectile before it reaches the tank.

Russia has spent years attempting to adapt the system—originally conceived to defeat anti-tank missiles and rockets—to the smaller, slower, and less predictable FPV drones now dominating the battlefield.
The newly released footage suggests those upgrades may not have solved the system’s core problems.
Reports from November 2025 indicated that Russia was struggling to make Arena-M reliably detect and intercept FPV drones because of limitations in its radar technology.

Small drones can be difficult to distinguish from ground clutter at close range, particularly when flying low and approaching from irregular angles. Their plastic construction can also reduce their radar signature, although the metal warheads they carry should still provide a detectable return.
Tank researcher Andrii Tarasenko, cited by Militarnyi, has questioned claims that FPV drones are too difficult to detect because of their low visibility. Since the aircraft carry metal explosive components, he argued, Arena-M’s failure may point to deeper shortcomings in the system’s radar and processing capabilities.

The system’s algorithms may also struggle to track and prioritize objects moving slowly or unpredictably—behavior common among FPV drones as operators maneuver them toward weak points in a vehicle’s armor.
Arena-M and its earlier variants have been under development since the late Soviet period. Despite repeated demonstrations and Russian claims that the system would be widely deployed, it has never entered large-scale frontline service.

Russian engineers have worked to integrate Arena-M onto T-72B3 and T-90M tanks since at least 2019. Individual examples fitted with the system began appearing publicly in 2024, but remained rare.
Previously, Russian engineers modified the “Arena-M” active protection system (APS) to intercept top-attack anti-tank missiles like the US-made Javelin.
However, analysts noted that the missile in the test approached the tank at a relatively shallow angle. Javelins fired at long range strike their target in a steep dive, while those launched at closer range descend at a more direct angle. This raises questions about whether “Arena-M” can reliably intercept third-generation anti-tank guided missiles in real combat.

A Lyman, Ukraine Raid Wipes Out Russian 36th Motor Rifle Reg't Commanders

2026-07-25T10:34:13.136-04:00

Ukrainian units have become adept at combining intelligence with drone attacks and special operations to target the headquarters of Russian units, eliminating commanders and their staffs.

As in this case from the Lyman front, the Ukrainians were able to freeze the operations of an entire Russian sector because of the Kremlin's top heavy command structure, which does not permit initiative by junior officers without higher approval. Taking out an attacking regiment's command at Lyman also froze the regiments on either flank of it because they could not coordinate their actions, thereby giving the Ukrainians an opportunity to counterattack and disrupt the Russians' plan and delaying, once again, their hopes for advancing towards Lyman. JL

RFU reports:

After tapping into Russian communications, Ukrainian troops tracked the commander of Russia’s 36th Motor Rifle Regiment. They mapped his routine, waiting until he was exposed as a Ukrainian kamikaze drone hit his car 87 kilometers behind the front line near Dmitrivka. The commander was wounded and three staff travelling with him sustained injuries putting them out of commission, beheading the regiment with a single drone. This is damaging because Russian command remains highly centralized and junior officers have limited freedom to make decisions without approval from headquarters, paralyzing the regiment. If one regiment becomes static because its leadership has been removed, the formations on either side become exposed. Ukraine's strategy is to exploit this chaos, launching follow-on attacks before replacement commanders restore control, turning localized tactical successes into broader operational breakthroughs.
Recently, a Ukrainian raid proved far more successful than simply winning ground, setting a domino effect in motion that beheaded a crucial section of the Russian frontline and left several gaps wide open.
It all began when Ukrainian operators attacked and cleared a Russian strongpoint in the forest of the Li-man direction. Ukrainian soldiers can be seen wearing ghillie suits to blend into the forest, helping them avoid detection by Russian soldiers and reconnaissance drones simultaneously while they are moving forward. Modern ghillie suits are also tailored to dampen heat signature, which further reduces the possibility of being detected by thermal drones. After stealthily moving through thick vegetation, the Ukrainian soldiers reached the Russian strongpoint, as the poor trash discipline evident from the visible empty cans they left all over the position had revealed the Russian presence. As the Ukrainian operators pushed deeper, they quickly came into contact with two Russian soldiers in the next dugout. After a fierce firefight broke out, the Ukrainians killed the Russian infiltrators and pushed deeper to fully clear the position. Ukrainians took off jackets and gear so they could go through the intelligence from cover, instead of out in the open. They captured documents, maps, and radios that the Russians had on them. This kind of captured intelligence and sensitive equipment is extremely valuable, as Ukrainians use this to tap into Russian communications and listen in on commanders revealing key information.
After tapping into the Russian communications, the Ukrainians were able to track down the commander of Russia’s thirty sixth Motor Rifle Regiment. They followed the exact movements of the commander, mapped out his routine, and patiently waited until he was exposed. The ideal moment to strike came while he was completely vulnerable and exposed on the road, as there is no option to quickly hide or run away, with cars often driving over one hundred kilometers per hour on damaged roads to reduce travel time and minimize exposure to Ukrainian observation. However, this did not save him this time, as a Ukrainian kamikaze drone closed in on the commander's car and hit it over eighty-seven kilometers behind the front line near the settlement of Dmitrivka. As a result of the successful Ukrainian drone strike, the newly appointed regiment commander was heavily wounded, and three of his staff travelling with him sustained injuries as well and were out of commission, effectively beheading the entire Russian regiment with a single drone.
This is crucial because now this Russian regiment has lost its commander and his entire entourage, leaving its command without an immediate replacement. By wounding the commander's second, third, and fourth replacements as well, it will be hard to reinstate the command of the regiment and find other subordinates who can do their duty. This is particularly damaging because Russian command remains highly centralized and junior officers generally have limited freedom to make decisions without approval from higher headquarters, effectively paralyzing the full regiment.
That paralysis quickly spreads beyond the headquarters itself because this regiment is currently responsible for advancing through the forests toward Lyman, with multiple assault groups relying on the headquarters to coordinate their movements and direct artillery, drone, and aviation support. Once that command structure is disrupted, communications begin breaking down, assault groups lose coordination, and supporting fires become delayed or unavailable. Because modern frontlines are tightly interconnected, the neighboring units depend on each other to secure their flanks. If one regiment suddenly becomes static because its leadership has been removed, the formations on either side become exposed. They must either divert forces to cover the growing gap or continue advancing while risking encirclement. This creates exactly the opportunity Ukraine is looking for because now, Ukrainian forces can exploit these weak points, maneuver through the emerging gaps, outflank neighboring Russian units, and threaten their rear. In this way, the elimination of a single regiment commander can trigger a chain reaction across the sector, disrupting the Russian offensive and potentially undoing months of hard-fought progress, or at the least massively delaying their efforts.
Overall, this operation demonstrates that Ukraine is increasingly targeting the Russian military's decision-making rather than just its frontline positions. By combining intelligence gathered during a small infantry raid with precision drone strikes, Ukrainian forces can paralyze entire formations at minimal cost. The loss of a regiment's command group will create confusion, slow reactions, and disrupt coordination across thousands of troops. Ukraine's strategy is to exploit this chaos immediately, launching follow-on attacks before replacement commanders restore control, thereby turning localized tactical successes into broader operational breakthroughs.

As Musk's Cybertruck Flops, Stock Tanks, Google Takes Stake In SpaceX

2026-07-25T10:10:36.460-04:00

In the annals of automotive history, the car long considered the worst failure ever was Ford's Edsel, introduced in 1957 just as a recession hit, but also plagued by unpopular styling and poor quality. It's reputation as Detroit's biggest flop stood for over 50 years - until the Tesla Cybertruck came along. It also has styling not universally popular, a reputation for poor quality (see photo) and minute sales, which are getting even smaller.

But that has not stopped Elon Musk. And even as SpaceX stock has done a round trip and is now lower than it was at its IPO launch, Google just announced that they are taking a multi-billion dollar stake in the company. At the least, the Goog appears to see an opportunity to pick up some assets and expertise quickly and cheaply, especially as demand for more data center capacity remains unabated. JL

C de Costa reports in Newsbreak and Katherine Hamilton reports in the Wall Street Journal:

The Ford Edsel became shorthand for corporate hubris and misread markets. The Cybertruck is earning a similar distinction. Musk promised a quarter-million Cybertrucks a year - Tesla delivered 8% of that target, then lost half of that. Q1 2026 delivered the truck’s worst quarterly performance since launch - just 3,519 units. Registration data indicate 20% of Cybertrucks sold in Q4 2025 went to companies owned by Musk - SpaceX and xAI. Bloomberg has made the Edsel comparison official, framing the Cybertruck as the biggest automotive flop of the EV era. SpaceX stock rose 19% to $135 on its first day of trading June 12, for a market value of $2.1 trillion. Its shares recently traded at $112.13. Based on SpaceX’s market capitalization of $1.52 trillion July 23, Google has a 6% stake. It has a long history of investing in SpaceX, discussing a rocket-launch deal to help develop orbital data centers

Back in 2019, CNBC’s Jim Cramer watched the Cybertruck reveal and called it the “E-Edsel.” Most people laughed. Nobody’s laughing now. Q1 2026 delivered the truck’s worst quarterly performance since launch — just 3,519 units, according to Cox Automotive data. That’s from a vehicle Elon Musk once projected at 250,000 units per year. Bloomberg has since made the Edsel comparison official, framing the Cybertruck as potentially the biggest automotive flop of the EV era.

From Hype to Hard Numbers
Musk promised a quarter-million trucks a year — Tesla delivered roughly 8% of that target, then lost half of that.
In 2024, Tesla moved about 39,000 Cybertrucks in the U.S. during its first full year on sale. Then 2025 cut that nearly in half: 20,237 units, the steepest decline for any EV in the American market that year, per Cox Automotive. Q1 2026 sank further to 3,519 deliveries, down 45.1% from the same quarter a year earlier — not a rounding error, but a product in measurable trouble.
The gap between ambition and reality is genuinely staggering:
2024 U.S. sales: ~39,000 units (Cox Automotive)
2025 U.S. sales: ~20,237 units — a 48.1% decline, steepest drop for any EV that year
Q1 2026 U.S. deliveries: 3,519 units — lowest since launch, down 45.1% from Q1 2025
The Edsel in the Room
When nearly one in five trucks sold goes to the owner’s other companies, the demand story starts looking like fiction.
Registration and fleet data indicate that nearly one in five Cybertrucks sold in Q4 2025 went not to retail customers but to companies owned by Musk himself — SpaceX and xAI. Electrek has reported hundreds, possibly thousands, of units acquired this way. Fast Company called the Cybertruck “the biggest flop in decades.” The internal buying doesn’t disprove that characterization — it underlines it.

The truck itself isn’t terrible on paper. Towing capacity is genuinely strong, reaching figures that comfortably outpace a typical F-150 configuration. But raw capability doesn’t close the sale when the entry price sits at $69,990 for a stainless steel polygon most truck buyers never requested. That’s a significant premium over a base Ford F-150’s roughly $40,000 starting point — for a segment where utility and value retention drive purchase decisions far more than futuristic aesthetics.
Quality concerns added pressure from the start. Multiple recalls within the first 13 months — covering wipers, trim pieces, and structural issues — reinforced the perception that buyers were absorbing beta-test reliability at flagship prices. Traditional truck buyers who need a workhorse, not a conversation starter, found the Ford F-150 Lightning and Rivian R1T considerably less risky. Like a prestige streaming series canceled after one season despite a massive marketing push, the Cybertruck moved from viral sensation to cautionary tale at remarkable speed.
What Comes Next
A cheaper AWD variant is reportedly on the way — but the Edsel comparison is already part of the permanent record.
The Ford Edsel became shorthand for corporate hubris and misread markets. The Cybertruck is earning a similar distinction for the tech era — a reminder that even the most hyped product cannot manufacture demand it never genuinely had. A lower-cost AWD variant was anticipated for mid-2026. Whether that resets the narrative or simply extends the decline remains a genuinely open question.

Google said it owns $94.1 billion in SpaceX SPCX -3.79%
decrease; red down pointing triangle shares, a roughly 6% stake in the newly public company.
The tech conglomerate’s marketable equity securities include $80 billion in SpaceX shares subject to short-term restrictions on the ability to sell and $14.1 billion of shares subject to long-term restrictions through the third quarter of 2027, according to a quarterly filing published Thursday.
The restrictions are common for shareholders that buy stock around the time of an initial public offering.
SpaceX debuted on the public markets in June, marking a blockbuster IPO which Nasdaq said earlier on Thursday was its biggest listing in history.
SpaceX stock rose 19% to $135 on its first day of trading June 12, giving the company a market value of $2.1 trillion. The stock price has edged lower over the past month,
SpaceX shares recently traded at $112.13.
Based on SpaceX’s market capitalization of $1.52 trillion on July 23, Google has a roughly 6% stake in the space and AI company.
Google has had a long history of investing in SpaceX. In May, the two companies were discussing a rocket-launch deal to help develop orbital data centers, The Wall Street Journal reported.

Google has been expanding its efforts to put orbital data centers in space as it ramps up its AI development, and SpaceX Chief Executive Elon Musk has expressed interest in the unproven technology.

Ukrainian Dnipropetrovsk Offensive Retakes 7 Km, 6 Villages, Pushes Russians Back

2026-07-24T11:39:11.647-04:00

The significance of this action may be primarily in its addition to other Ukrainian forces attacks and counterattacks at this point in mid-summer, all of which are contributing to the overall reality of lost Russian momentum and even capability to defend what they have. JL

Oleksy Leonov reports in Militarnyi:

Ukrainian forces carried out a successful offensive, defeating Russian troops in the Dnipropetrovsk region near the village of Filiya. The Ukrainian troops regained control of 6.7 km² of territory. The Ukrainian military liberated six villages in the Oleksandrivka sector of the Dnipropetrovsk region: Ternove, Zaporizke, Novogeorgievka, Vorone, Sichneve, and Maliivka.
Ukrainian forces carried out a successful offensive and repelled Russian troops in Dnipropetrovsk region near the village of Filiya.
This was reported by analysts from the DeepState project.
According to analysts at Shrike News, Ukrainian troops managed to regain control of 6.7 km² of territory.
The village of Filiya is located on the left bank of the Solona River, which flows into the Vovcha River approximately 5 km downstream. Upstream, 3 km from Filiya, lies the village of Novopavlivka.
At the same time, Russian troops continue to hold control over certain settlements in the Dnipropetrovsk region, including the village of Dachne.

Earlier, the Ukrainian Defense Forces pushed Russian troops back from the settlement of Ivanivka.
The Ukrainian military also liberated six villages in the Oleksandrivka sector of the Dnipropetrovsk region: Ternove, Zaporizke, Novogeorgievka, Vorone, Sichneve, and Maliivka.

With Gas Lines, Attacks on E-commerce, Russia's Middle Class Now Feels the War

2026-07-24T11:07:01.385-04:00

The Kremlin has worked hard to convince Russia's relatively comfortable urban middle class that the invasion of Ukraine was none of their concern and would not affect their lives. But that was before Ukrainian attacks on Russian energy infrastructure and on the country's e-commerce hubs disrupted the once normal economy.

Now, a majority of Russians - 60% - say their economic conditions are worse. Not that anyone is feeling sorry for them, given what the Russian military have done to Ukraine. Welcome to the war your government started. JL

Alexander Osipovich reports in the Wall Street Journal:

Ukraine's long-range drone strikes have brought the war home to Russia’s middle class, with attacks on logistics hubs for the country’s biggest e-commerce company, Wildberries, a homegrown version of Amazon.com. Nine people were killed in the attacks, and total costs range between $3 billion and $4.4 billion. Most of that is due to the destruction of goods belonging to vendors. Less than two weeks before the attacks, the company amended its agreement with vendors to absolve itself of liability in the event of drone strikes. Russian drivers have also been waiting for hours in huge lines at gas stations. Moscow's stock index lost one-quarter of its value this year, and last week sank to its lowest level since October 2022, after a selloff in government bonds. Between March and May, a record 60% of Russians told Gallup their economic conditions were worse. "Because of the lines at gas stations, because of the fires, everyone in Russia sees the war from their window.”

For four years, the Kremlin sought to shield Russians from the fallout of the war in Ukraine, maintaining a veneer of normality in everyday life and ensuring that affluent city dwellers could still enjoy the conveniences of a 21st-century economy.
That effort is now crumbling.
Ukraine’s long-range drone strikes have brought the war home to Russia’s urban middle class, most recently with attacks on logistics hubs for the country’s biggest e-commerce company, a homegrown version of Amazon.com.
The company, called Wildberries, is deeply embedded in daily Russian life, with 81 million users a month, according to its website. Hundreds of thousands of third-party vendors use its online bazaar as a storefront, selling everything from clothing to housewares to Sony PlayStations.
On Saturday, Ukrainian drones struck two of the company’s facilities, including a vast complex outside Moscow that was incinerated in a fire that took more than two days to extinguish. Two more Wildberries sites in southern Russia were hit overnight Tuesday to Wednesday.
Wildberries has 81 million users a month and hosts hundreds of thousands of third-party vendors. Donat Sorokin/TASS/Zuma Press
At least nine people were killed in the attacks, and total costs could range between $3 billion and $4.4 billion, according to Data Insight, a Moscow-based research firm. Most of that is due to the destruction of goods belonging to vendors, rather than Wildberries’ buildings themselves.
Some vendors on social media expressed shock that their inventories had been torched. Viktoria Terekhova, who sells playfully designed clothing for babies and toddlers under the brand Fareal, broke into tears in a video posted to Instagram as she spoke of the victims of the attacks and the daunting task of keeping her business afloat.
In total, around half her inventory was destroyed in the strikes on the warehouses near Moscow and in southern Russia, costing more than $160,000. Terekhova said in an interview that the attacks have put her business at risk of defaulting on its loans. “If these strikes continue, we’ll have nothing left,” she said.
Many of the vendors have voiced anger at Wildberries, which they worry won’t compensate them for their losses. Less than two weeks before the attacks, the company amended its agreement with vendors to absolve itself of liability in the event of drone strikes.
Wildberries didn’t respond to a request for comment.
Stills from a video posted by Viktoria Terekhova after much of her inventory was destroyed. Fareal
Tatyana Kim, its founder, has promised to help vendors and said Wednesday that the company had begun making payments to smaller vendors whose stock was destroyed in the attack near Moscow. She said Wildberries’ insurance policy didn’t cover drone attacks.
The strikes have prompted debate over whether Wildberries’ warehouses are a legitimate military target. Ukrainian President Volodymyr Zelensky justified the attacks by saying the logistical hubs hit by Kyiv help supply the Russian military with drone components, navigation equipment and other gear. A review of Wildberries website shows drones and body armor for sale, but the overwhelming majority of its offerings are civilian in nature.
Russia has frequently bombarded civilian infrastructure in Ukraine with missiles and drones. After the weekend strikes on Wildberries warehouses, Russia hit back by attacking logistics centers used by Nova Poshta, a sort of Ukrainian version of FedEx. Even before the attacks on e-commerce warehouses, Russians were voicing frustration about widespread fuel shortages, the result of a monthslong campaign of Ukrainian drone attacks on refineries.
So far in July, more than half of Russian refining capacity has been knocked offline by the attacks, according to estimates from IIR Energy, a Texas-based energy-market data and intelligence firm. The shortages have forced Russia to import gasoline for the first time in years, while drivers have been stuck waiting for hours in huge lines at gas stations.
“The fuel crisis is arguably the most visible shock for ordinary citizens since the mobilization of new troops in September 2022,” Oxford Economics said in a research note this week.
By many measures, Russia’s economy was a success story in the wake of Vladimir Putin’s full-scale invasion of Ukraine in 2022.

Hit with sanctions, Russia reoriented its trade around China and other non-Western countries. Popular goods from Western brands that exited Russia after the invasion, such as Apple iPhones, continued to be on sale, brought in by unofficial resellers in places like Dubai and Hong Kong. Oil revenues kept pouring in to fund the war effort. Military spending juiced economic growth in 2023 and 2024, though it has since slowed.
Now, rising fuel prices have stoked inflation, forcing Russia’s Central Bank to slow down its pace of interest-rate cuts, a disappointing shift for local investors and businesses.
Moscow’s main stock index has lost nearly one-quarter of its value so far this year, and last week sank to its lowest level since October 2022. After a selloff in government bonds, the Russian Finance Ministry said this week it was halting new sales “to aid with stabilizing the market situation.”

Between March and May, a record-high 60% of Russians told Gallup that their local economic conditions were getting worse, the first time that a majority of respondents had been negative since the pollster began asking the question in 2006.
Putin has shown no inclination to scale back his military ambitions in Ukraine in response to the worsening economic picture. On Wednesday, he dismissed the fuel crisis as “temporary” in a televised meeting with the government’s top economic policymakers. Earlier, the Kremlin’s top spokesman condemned the strikes on Wildberries warehouses and said the authorities were in contact with the company.
Maxim Mironov, a finance professor at IE Business School in Madrid and critic of the war in Ukraine, said that Ukraine’s drone attacks would ultimately have only a small negative impact on Russia’s economy, but their psychological effect was already meaningful.
“For years Putin has been selling the story that the fighting is somewhere on the fringes of the empire, it’s a special military operation, and everything is normal,” he said.
“Now, people have begun to feel it, because of the lines at gas stations, because of the fires…Everyone in Russia sees the war from their own window.”

First Of Its Kind Ukraine Ground-Air Robot Assault Destroys Russian Bunkers

2026-07-24T10:15:16.811-04:00

The Ukrainians used a heavy-duty bomber drone to carry a fully-loaded ground drone laden with explosives over 19 kilometers so that the ground drone could then attack a Russian bunker complex.

Yes, the bomber could have tried bombing it and the ground drone could have attempted the journey over fields and roads, but this unique combined arms assault, reduced the risks and enhanced the probability of success - and all with Ukrainian soldiers' lives being at stake. JL

Euromaidan Press reports:

Just days after landing robots on an enemy shore in the first all-robot amphibious assault, Ukrainian forces flew one into battle as, this week, a Ukrainian heavy bomber drone airlifted an explosives-laden unmanned ground vehicle into action. A 19-km trip is a long trip for a remotely controlled vehicle that's susceptible to enemy radio jamming and prone to getting stuck in potholes. So airlifting a UGV underneath a heavy bomber drone gets the UGV into action faster and with less risk of a loss or accident en route. A heavy bomber drone could just drop a bomb. But a UGV can do something an aerial bomber can't do: drive straight through the door of a Russian bunker regardless of how much protection the Russians have.
Just days after landing robots on an enemy shore in the first all-robot amphibious assault, Ukrainian forces flew one into battle. This week, a Russian video circulated depicting a Ukrainian heavy bomber drone airlifting an explosives-laden unmanned ground vehicle into action.
What happened next is unclear. Did the exploding UGV successfully strike its target?
Regardless of the outcome, however, the unmanned air assault is notable. If Ukrainian forces can use an unmanned aerial vehicle to airlift one unmanned ground vehicle, they can do it a second time. And a third. And fourth. So on until these kinds of all-robot operations become routine.
And that's indeed the aim in Kyiv. To make robotic combat operations an everyday occurrence in order to help alleviate a deep and persistent manpower shortage. This spring, recently sacked Ukrainian defense minister Mykhailo Fedorov claimed Ukraine would acquire 25,000 UGVs in just the first six months of 2026.

That's twice as many UGVs as Kyiv bought in all of 2025. The rapid mass acquisition would allow more Ukrainian brigades to shift to robotic resupply missions, removing potentially thousands of precious human logisticians from dangerous roads.
But UGVs perform combat missions, too. Armed with guns, they help counterattack Russian infiltrators and also patrol roads inside the disputed gray zone. In November, one gun-armed UGV belonging to the Ukrainian 5th Assault Brigade, fighting north of Kostiantynivka in eastern Ukraine's Donetsk Oblast, inadvertently advanced UGV tactics in an accidental encounter with an up-armored Russian MT-LB infantry transport.

The UGV, equipped with thermal night vision and armed with a remotely operated M2 .50-caliber machine gun, was reportedly on a combat trial—crawling around the battlefield at nighttime for testing purposes—when an MT-LB rolled toward it on the same road.
Connected via radio, the UGV operators opened fire, peppering the tracked MT-LB with rounds until the Russian vehicle passed by—and escaped in the opposite direction.

An unmanned ground vehicle (UGV) from the Bulava Battalion of Ukraine's 72nd Mechanized Brigade. Photo: David Kirichenko
One-way drones
Other Ukrainian UGVs are meant to fight exactly once. They're suicide bombers, hauling payloads of explosives into Russian strongpoints and blowing themselves up. Back in January, the Ukrainian 3rd Assault Brigade's NC-13 robotics unit piled a dozen TM-62 anti-tank mines, each with 7.5 kg of explosives, onto one of its UGVs and carefully steered the robot into a Russian bunker presumably somewhere south of the ruins of Kupiansk in eastern Ukraine's Kharkiv Oblast.
It made a very big boom. As seen in an official video, the NC-13 team broke into cheers when an escorting aerial drone registered the blast. The team had carefully assembled, painted, and deployed the robot apparently over a period of days before maneuvering it 19 km across rough terrain to deliver the explosive payload.
The recent robotic air assault addressed the mobility problem. A 19-km trip is a long trip for a remotely controlled vehicle that's susceptible to enemy radio jamming and prone to getting stuck in potholes. Airlifting a UGV underneath a heavy bomber drone gets the UGV into action faster and with less risk of a loss or accident en route.

A heavy bomber drone could just drop a bomb, of course. But a UGV can do something an aerial bomber can't do: drive straight through the open door of a Russian bunker regardless of how much protection the Russians pile atop the door.

The Ukrainians are finding more and more new ways to get their UGVs into action. On 13 July, the Ukrainian 123rd Territorial Defense Brigade sailed several unmanned surface vessels across the western Black Sea and landed gun-armed unmanned ground vehicles on the Kinburn Spit, at the westernmost edge of the Russian occupation zone in southern Ukraine.
The 123rd Territorial Defense Brigade released a video montage—shot by a USV, a UGV and an escorting UAV—depicting the UGV firing on a Russian installation with its suppressed .50-caliber machine gun.
It was an all-robot amphibious assault. "An impressive feat," RAND analyst Michael Bohnert wrote.
The later robotic airlift was no less impressive.

Major Russian Dobropillia Armored Assault Annihilated, 9 Tanks Destroyed

2026-07-24T10:00:47.382-04:00

The Russian command took another one of those actions that - despite four-plus years of failing to achieve objectives, despite all they've learned about drones, despite the fact that they can no longer replace casualties, weapons and armored like they did even last year - the world has learned to expect from them.

Which is, that they ordered a suicidal massive armored assault on a well-defended Ukrainian sector - Dobropillia - and, to absolutely no one's surprise, suffered precisely the annihilation anyone could have predicted. Nine tanks, two infantry fighting vehicles, three Grad rocket launchers and dozens of soldiers: Kaput. It is fair to wonder if this is reported up the chain of command to the Kremlin as a victory or as noble effort that weakened Ukrainian defenders - or if no one really cares anymore, they just need to keep showing Putin they are trying so he can keep trying to convince Trump and the Europeans he's still a threat. JL

New Voice of Ukraine reports:

Russian forces have made a failed breakthrough on the Dobropillia front. Drone operators from the Phoenix border unit of the Pomsta ("Revenge") brigade, working alongside adjacent units, repelled a Russian assault during which the Russians deployed dozens of armored vehicles on July 24. Border guards said they had not seen a breakthrough attempt of this scale in their area of responsibility for a long time. Drone operators from the Phoenix unit destroyed nine Russian tanks, two armored vehicles and three Grad multiple-launch rocket systems that were intended to provide fire support for the offensive. Several dozen Russian troops were also killed.
Russian forces have made a failed breakthrough on the Dobropillia front. Drone operators from the Phoenix border unit of the Pomsta ("Revenge") brigade, working alongside adjacent units, repelled a Russian assault during which the Russians deployed dozens of armored vehicles on July 24. Border guards said they had not seen a breakthrough attempt of this scale in their area of responsibility for a long time. Drone operators from the Phoenix unit destroyed nine Russian tanks, two armored vehicles and three Grad multiple-launch rocket systems that were intended to provide fire support for the offensive. Several dozen Russian troops were also killed.
Drone operators from the Phoenix border unit of the Pomsta ("Revenge") brigade, working alongside adjacent units, repelled a Russian assault during which the Russians deployed dozens of armoured vehicles on July 24.
Ukraine's State Border Guard Service said Russian forces had been amassing equipment for an offensive operation for an extended period. Border guards said they had not seen a breakthrough attempt of this scale in their area of responsibility for a long time.

Drone operators from the Phoenix unit destroyed nine Russian tanks, two armored vehicles and three Grad multiple-launch rocket systems that were intended to provide fire support for the offensive. Several dozen Russian troops were also killed..

AI-Linked Cos' Value Up $27 Trillion, 36% of US Stock Market. That's Not Good

2026-07-24T09:50:40.887-04:00

Yeah, you're tired of hearing it: the negativity, the refusal to accept that this time is different, the naysayers who just don't get that AI is going to supplant everything we've ever known. Etc.

The problem is that the alarm bells of everyone who actually knows anything about money and economics and statistics and business are not just ringing - they're exploding. And the reason is pretty simple: a sensible analysis of the projections compared to the actual amount of potential customers willing and able to spend it who exist on this planet just does not add up to a conclusion where the AI bubble (yep, used the b word) pays out. There are not enough people doing enough things to make additional capital to justify the expense and the debt and the risk. The issue is no longer if, it is when. Do. The. Math. JL

Annie Lowry reports in The Atlantic:

The value of AI-linked firms has climbed $27 trillion in the past three years, 36% of the value of the entire U.S. stock market. Although future earnings could justify those valuations, the profit expectations require Panglossian optimism. OpenAI CEO Sam Altman is arguing we are in an AI bubble. The International Monetary Fund is citing it as a significant risk to financial stability. Silicon Valley is building 1,500 data centers in the US, spending more than $700 billion on the build-out this year. (But) tech companies are going to need to start generating huge revenues and profits to justify these valuations. OpenAI needs $100 billion in free cash flow by 2030. AI-infrastructure investment is now responsible all American GDP growth. Without it, recession. (And) if more communities ban data centers, or if companies prove reticent about purchasing AI, or if China develops AI models that are not as expensive—there could be a massive correction.

The American stock market is booming, thanks to artificial intelligence. Tech giants are borrowing billions to acquire AI talent, purchase chips and hardware, and construct data centers. And market watchers are starting to get worried. They see financiers bulldozing giant piles of money to private AI start-ups with no realistic path to profitability, tech companies reliant on other tech companies for revenue growth, and non-tech businesses without a lot to show for their AI investments. The value of AI-linked firms has climbed $27 trillion in the past three years—an astonishing amount, equivalent to 36 percent of the value of the entire U.S. stock market today. Although future earnings could justify those valuations, as Dominic Wilson and Vickie Chang of Goldman Sachs argued in a note to clients, the profit expectations require Panglossian optimism.

No less an authority than Sam Altman is arguing that we are in an AI bubble. The International Monetary Fund is citing it as a significant risk to financial stability and warning about what might happen when it bursts: diminished investment, tighter credit, reduced consumption, disrupted trade flows.
That’s pretty much what happens when any bubble pops, as a Dutch tulip obsessive could have told you in 1637 or a bitcoin evangelist could have told you in 2011, 2013, 2014, 2018, or 2022. Yet the AI bubble is no ordinary bubble. Hyper-rich corporations are stoking it, rather than kitchen-table investors. They’re blowing it up when credit is fairly expensive, not dirt cheap. That might make the bubble less fragile and longer lasting than those of the past. But it won’t make it any less painful when it pops.
The dot-com bubble of the late 1990s and the housing bubble of the late aughts were remarkably broad-based compared with the AI bubble today. Uncle Ted got a hulking desktop computer, opened a newfangled E-Trade account, and started day-trading shares in Apple and Pets.com. The share of American households owning equities climbed 13 percentage points from 1995 to 2001, during which time more than 2,752 firms went public (far more than the 730 operating businesses that have IPOed in the past six years). A half decade later, Aunt Linda bought a condo with nothing down, flipped it, and mortgaged three new investment properties in the Phoenix suburbs, sight unseen. The homeownership rate rose 5 percentage points as the housing bubble inflated; 40 percent of mortgages issued at its height went to investment or vacation properties.
In both cases, regular people were staking their savings on what seemed like a winning bet: the internet changing everything, housing prices never going down. In both cases, cheap credit fueled the irrational exuberance. Low interest rates let venture capitalists fund nonsense web businesses and let banks provide junk loans to borrowers with terrible credit. In both cases, rising interest rates popped the bubble.

Today, Uncle Ted and Aunt Linda aren’t really getting in on the AI frenzy. The share of Americans who own stocks has held steady. Household debt has grown, but it has fallen relative to disposable income and GDP. Everybody seems to know someone who was personally burned by the dot-com collapse and the housing crisis. How many people know someone who’s staking it all on OpenAI and Anthropic today? (The companies aren’t public, after all.) Indeed, how many people know someone whose livelihood has been directly affected by the AI frenzy at all? The insularity of the AI bubble isn’t the only thing that makes it unusual, and we should probably think about it as two overlapping bubbles instead of just one. AI is driving tremendous spending on capital expenditures—physical infrastructure, software development. And it is driving a tremendous run-up in company valuations.
Digital innovations don’t tend to require a ton of labor or heavy equipment. The companies offering them tend to be capital-light, like marketers and insurers, rather than capital-intensive, like airlines and hotel chains. But AI is different. A start-up might need thousands of times as much computing power to train an algorithm as it would need to develop a conventional software product. To get that power, Silicon Valley is building 1,500 data centers in the United States and counting, and purchasing a mammoth quantity of semiconductor chips. (The Wall Street Journal calls these chips “the 21st century’s most important market.”) Amazon, Microsoft, Alphabet, and Meta alone are spending more than $700 billion on the build-out this year. AI-infrastructure investment is responsible for essentially all American GDP growth at the moment. Without it, in other words, we might be in a recession.
The boom is lifting the value of farmland, driving up the cost of construction employment, and flushing huge sums to water and electrical utilities. Still, tech companies are the main beneficiaries: Nvidia is selling chips to Meta; Amazon is selling cloud-computing capacity to OpenAI. This surge in revenue, and the excitement about the money to be made when AI starts bolstering productivity and profits, is driving up valuations. The Magnificent Seven—Alphabet, Amazon, Apple, Meta, Microsoft, Nvidia, and Tesla—now account for one-third of the value of the Standard & Poor’s 500. OpenAI is worth more than Eli Lilly, JPMorgan Chase, Visa, Costco, Exxon Mobil, Wells Fargo, CVS Health, McDonald’s, and Boeing.

Tech companies are going to need to start generating huge revenues and huge profits to justify these valuations. OpenAI needs to spin up roughly $100 billion in free cash flow by 2030, according to calculations by Harrison Rolfes of PitchBook. Analysts expect it will lose $10 billion to $30 billion that year. If it does—or if more communities ban data centers, or if non-tech companies prove reticent about purchasing AI software, or if China develops AI models that do not require so much computing power—we could be in for a massive correction. The AI economy is a trillion-dollar ouroboros of buying and selling, investment and equity staking, all happening between San Francisco and San Jose. Big Tech is advancing money to AI start-ups to buy cloud services from Big Tech, which is using the revenue to run new AI models … you get the idea. What happens to one firm could happen to all of them.
Tech companies are also going to need to start generating even bigger revenues and profits to pay off their debts. In the early days of AI, venture capitalists, wealthy individuals, and Big Tech firms used hard cash to invest in the new technology. But the build-out has proven so expensive that Silicon Valley has turned to corporate bonds and private credit, meaning loans made by entities other than banks. Thus, though the AI boom has happened while interest rates are fairly high, it still involves a lot of leverage. The deals are complicated, opaque, and structured so as to be invisible on traditional balance sheets, making the “ultimate distribution of risk less transparent,” Stijn Van Nieuwerburgh of Columbia Business School has found. Already, lenders are getting queasy—“buy-side indigestion,” as Morningstar describes the market’s recent reticence to meet Silicon Valley’s demands.

When the bubble bursts, Uncle Ted and Aunt Linda will be affected, even if they were never the ones stoking the frenzy. Their retirement plans and pensions are invested in Silicon Valley stocks that could drop; their small businesses are reliant on access to credit that might get throttled. But, hey, there’s always the possibility that AI might steal their job before any of that happens.

Russian Troops Hit By Machine Gun Realized Too Late It Was A Ground Drone

2026-07-23T11:23:41.280-04:00

For the Russians, confusing, demoralizing - and lethal.

For the Ukrainians, effective. JL