Armwood Technology Blog

China Sought Access to Anthropic’s Newest A.I. The Answer Was No. - The New York Times

2026-05-12T12:52:24.662-04:00

China Sought Access to Anthropic’s Newest A.I. The Answer Was No.

"The latest artificial intelligence models from Anthropic and OpenAI are extending the United States’ lead over China and intensifying the rivalry between the countries.

May 12, 2026Updated 11:36 a.m. ET

A growing number of national security officials and analysts have begun to liken the intensifying competition between China and the United States over artificial intelligence to the Cold War’s nuclear arms race.Marissa Leshnov for The New York Times

A representative from a Chinese think tank approached officials from Anthropic at a meeting in Singapore last month to insist that the company change its stance and give Beijing access to its powerful new artificial intelligence model, according to people briefed on the discussions.

Anthropic refused.

The request was not an official demand from the Chinese government. But the talks in Singapore were a form of exchange that is often meant to pave the way for formal, direct diplomacy.

When officials from the National Security Council at the White House learned about the exchange at the meeting, which was convened by the Washington-based Carnegie Endowment for International Peace, they reacted with alarm. Some Trump administration officials saw it as another sign that Beijing would try every possible avenue to swiftly acquire the most powerful artificial intelligence model a U.S. company has produced so far, according to people briefed on the discussions.

However subtle, the outreach is a sign of the intensifying competition between China and the United States over artificial intelligence, which a growing number of national security officials and analysts have begun to liken to the Cold War’s nuclear arms race. Chinese analysts see the release of the latest models from Anthropic and ChatGPT as a significant advance in American technology, one that could pose a threat to China.

The escalating rivalry is an important backdrop to the summit this week between the United States and China. President Trump is scheduled to arrive in Beijing on Wednesday, for meetings with the Chinese leader Xi Jinping. While the race to develop the most effective model is unlikely to be discussed, the two sides could talk about access to the chips that power artificial intelligence or guardrails around its use.

Previewing the summit Sunday during a call with reporters, a senior U.S. official said A.I. and cybersecurity were high on the agenda with Beijing, noting concerns with “the latest models of A.I.” The official said the United States and China were exploring how to establish better communication over artificial intelligence, creating a “deconfliction” channel in which experts from each country could address the risks of artificial intelligence.

In April, Anthropic announced a new A.I. model called Mythos. The company said it was withholding it from a public release because it was skilled at finding software vulnerabilities and could cause a cybersecurity reckoning. It made the model available to the U.S. government and more than 40 organizations and companies, so that they could identify and guard against future attacks.

The technology has set off alarm bells across the world. For rivals like China and Russia, it showed the risks of falling behind in the race to develop powerful artificial intelligence. The systems have the potential to give a nation-state the upper hand in defending against — and spearheading — cyberattacks at a vast scope and scale.

For years, U.S. officials have estimated that artificial intelligence models developed by the most advanced American firms are around six months ahead of China’s best models.

But according to some U.S. government and industry officials, the latest models — OpenAI’s ChatGPT 5.5 as well as Anthropic’s Mythos — have drastically extended the lead, potentially by nine months to a year. Other American officials have been more cautious, noting that China has a track record of catching up quickly.

Last year, innovations by the Chinese firm DeepSeek showcased the country’s ability to close the A.I. gap. And DeepSeek has said that its new model was adapted to run on chips made by the Chinese tech giant Huawei, further underscoring Beijing’s push to keep pace.

Still, Chinese analysts have expressed concern over the potential of the new Anthropic model. One organization, IDC China, said that Mythos posed a significant risk to Chinese companies, and that Anthropic’s limits on its reach created a technology gap. Another analyst, highlighting the cybersecurity threat, wrote that China was sharpening swords while the other side rolled out a Gatling gun.

Chinese analysts have been particularly worried because they view Anthropic as hostile to China. The start-up is currently embroiled in suits with the Pentagon, which announced it would be removed from American government networks after a dispute over how the technology would be used.

But since its founding, Anthropic has geared its business toward U.S. national security customers. It was the first to put its A.I. models on classified American networks, for example, and has long taken pains to keep its technology out of the hands of the Chinese. In September, Anthropic expanded restrictions on its technology to include companies whose ownership structures made them vulnerable to control “from jurisdictions where our products are not permitted, like China, regardless of where they operate.”

Increasingly, both the Chinese and U.S. governments view their artificial intelligence companies — including those that produce the models and the cloud companies that host the computer networks they run on — as national assets. China blocked a $2 billion acquisition by Meta of the Chinese A.I. company Manus. China has also told some of its A.I. start-ups that they cannot accept American investment without government approval.

Anthropic and OpenAI have restricted access to their latest models to a few companies and U.S. government agencies. Both models have the ability to discover previously unknown vulnerabilities in computer networks.

Chinese officials have argued that Anthropic and OpenAI have been wrong to keep a close hold on the models, contending that China needs access to them to find vulnerabilities in software, especially to defend its own critical infrastructure.

The meeting in Singapore was hosted by Carnegie under the ground rules that participants would not attribute the information discussed at the meetings. Many of the sessions dealt with domestic regulation of artificial intelligence.

The direct request by the Chinese think tank official was made on the sidelines of the meeting, not during one of the formal sessions, according to people briefed on the events.

Matt Sheehan, a Carnegie senior fellow and the organizer of the Singapore meeting, declined to comment on the conversations held at the event last month but described the broader meeting as essential to the two countries. “It’s critical for experts in the U.S. and China to maintain lines of communication about the potential risks with A.I.,” he said. “That’s why we host these dialogues.”

Officials from Anthropic declined to discuss the Chinese request.

Liu Pengyu, a spokesman for the Chinese Embassy in Washington, said he was not familiar with the exchange at the Carnegie meeting. But he said China was committed to balancing A.I. development with security.

“China stands ready to strengthen exchanges and cooperation on A.I. with all parties, to foster shared benefits through joint consultation and to work together to create an open, inclusive and mutually beneficial development environment,” he said in a statement.

U.S. officials noted that while a member of a Chinese think tank made the overture, it was all but certain that the Chinese government had approved and directed the message. Beijing typically exercises a strong degree of control over its think tanks, especially when they are engaged in unofficial diplomacy, like that type of exchange, known as a Track 2 dialogue.

While China’s official government line has been that its technology sector will be able to innovate independently and can overcome restrictions placed on it by the United States, Chinese companies have been more worried about their access to advanced chips, according to American industry officials and independent researchers.

U.S. officials are hoping the American companies will continue to delay China’s access to the most advanced chips so that American spy agencies may be able to use the new program to gain access to sensitive Chinese networks, according to former U.S. officials.

Industry officials are trying to persuade China to change its strategy on artificial intelligence and to not make its most powerful new models open source, which anyone can use. Putting a model that has the ability to quickly infiltrate networks in the hands of hackers could unleash chaos around the globe, security experts have said.

But Chinese officials remain skeptical, as the United States continues to look for ways to extend its lead over competition from Beijing. American companies including Anthropic, OpenAI and Google have accused Chinese firms of trying to steal their technology by essentially copying a model’s core competencies. Reuters reported last month that the State Department had sent a diplomatic complaint to China warning it against the practice.

Congress and the Trump administration are also increasing funding for an arm of the Commerce Department that devises export controls on chips and other technology.

A correction was made on

May 12, 2026

An earlier version of this article misstated the day President Trump is scheduled to arrive in Beijing. It is Wednesday, not Thursday.

Dustin Volz writes about cybersecurity and intelligence for The Times. He is based in Washington.

Julian E. Barnes covers the U.S. intelligence agencies and international security matters for The Times. He has written about security issues for more than two decades.

Sheera Frenkel is a reporter based in the San Francisco Bay Area, covering the ways technology affects everyday lives with a focus on social media companies, including Facebook, Instagram, Twitter, TikTok, YouTube, Telegram and WhatsApp.

Tripp Mickle reports on some of the world’s biggest tech companies, including Nvidia, Google and Apple. He also writes about trends across the tech industry like layoffs and artificial intelligence."

China Sought Access to Anthropic’s Newest A.I. The Answer Was No. - The New York Times

Neil deGrasse Tyson: Elon Musk's DELUSION?

2026-05-10T21:48:38.642-04:00

Neil deGrasse Tyson: Elon Musk's DELUSION?

Passkeys Are the New Passwords. You Should Start Using Them Now.

2026-05-10T08:55:29.762-04:00

Passkeys Are the New Passwords. You Should Start Using Them Now.

April 29, 2026

Max Eddy/NYT Wirecutter

By Max Eddy

“Max Eddy is a writer who has covered privacy and security — including password managers, VPNs, security keys, and more — for over a decade.

For 15 years, experts have told me that passwords are the biggest problem with online security and that’s just the way it is. The passwords that people make up are easily guessed by machines, and the ones that can’t be guessed are too hard to remember.

Over time, as more and more passwords became necessary, many people simply recycled theirs across different accounts — creating a precarious situation where one phished password or data breach gave an attacker access to the victims’ email, bank accounts, and anything else that shared that password.

Then came password managers, services and software in which a person could safely store all of their complex passwords and thus need to remember only a single password: the password for their password manager. This technology addressed, but didn’t solve, some of the problems. If people put in the effort, they could eventually have unique and complex passwords everywhere. But the majority of people did not opt to take on this herculean task that brought no immediate reward except that (maybe) in the future something bad might not happen to them (possibly). And a smart attacker could just phish even the best passwords anyway.

Two-factor authentication was the next bandage on the gaping wound of passwords. With 2FA protecting you, an attacker could have your password but wouldn’t be able to use it without a second confirmation, such as a code generated by an app or sent by text message. But another hoop to jump through for logging in remains a hard sell. And a smart attacker could just (you guessed it) phish most forms of 2FA anyway.

But passkeys are different. Instead of trying to fix unfixable passwords, passkeys are an entirely new technology that securely logs you in without your needing to remember your password or to perform a 2FA ritual. Passkeys are not perfect, and we’re still a ways off from their being commonplace, but learning what a passkey is and how to use it moves you a little closer to a more secure future.

What is a passkey?

A passkey is, essentially, a digital key that “unlocks” your online account by securely logging you in without a password. In Tamora Pierce’s fantasy novel First Test, knight student Kel acquires a lock for her room that opens only when she speaks a magic word while turning an equally magical key. This is, in high-fantasy terms, a passkey: It works on only one door, it can’t be duplicated, and it works only when the owner allows it to do so with a magic word.

When you have a passkey, your device, instead of prompting you to type in your password, automatically finds your passkey and asks your permission to use it. You then authenticate it with a PIN, facial recognition, or a fingerprint scan, and only then does it log you in to your account.

“A passkey is like a super-long password that you can’t copy-paste, but that your computer and phone can use to log in to a website quicker and more safely than a password,” said Jacob Hoffman-Andrews, a senior staff technologist at the Electronic Frontier Foundation.

Passkeys can’t be forgotten, unlike passwords. Passkeys are bound to specific websites, so there’s no risk of their being used on a phishing site or being sent to a savvy scammer. And unlike passwords, they can’t be exposed if the site where you have an account has a data breach.

“In a world where AI can mimic voices and generate flawless phishing emails at scale, passkeys are the most critical defense we have to prevent credential theft,” said Derek Hanson, chief technology officer for Yubico, the maker of Wirecutter’s favorite security keys and a major force behind the creation of passkeys.

Build your virtual keychain

The first thing you need to create a passkey is to figure out where to store it.

You can create a passkey in a few places: on your device, in a password manager, or on a physical piece of hardware called a security key. Regardless of where you store your passkeys, they’re secured with end-to-end encryption. As a result, Apple, Google, Microsoft, and your password manager don’t have the means to decrypt them. Only the devices that you authorize can access your passkeys.

You may have seen some of the biggest sites on the internet — such as Amazon, Google, and PayPal — prompt you to create passkeys. And the reason is that many of the biggest tech companies actively support them. Apple, Google, and Microsoft all back passkeys, and their devices are set up to create, store, and use passkeys. If you’re just starting out with passkeys, we recommend using whatever the default option is on your device.

The passkeys you create can automatically sync between devices to be more available to you — but that feature is limited by where you store your passkeys. Passkeys made on iPhones, iPads, and Macs are saved to the Apple Passwords app and automatically sync between devices where you’re logged in with your Apple ID. Android devices, Chromebooks, and the Chrome browser save your passkeys to the Google Password Manager and sync similarly. Windows also has built-in support for creating and storing passkeys and syncs your passkeys between Windows devices where you’re logged in with your Microsoft account.

You can also choose to store your passkeys in password managers — our top picks both support passkeys. Passkeys stored in password managers such as 1Password sync to the apps and browser extensions, making them available no matter what device you’re using. If you store passkeys in your password manager, secure your password manager account with a complex, unique password and enable 2FA.

Top pick

1Password offers attractive, colorful apps and a straightforward onboarding experience that’s easy for newcomers to understand. At the same time, the technically inclined will appreciate its advanced features and security.

$29 from 1Password
(per year)

Budget pick

The free version of Bitwarden covers all the basics of a good password manager and doesn’t cost anything. But features like advanced security reports and encrypted file storage cost extra.

If you don’t like the idea of your passkeys floating around in the cloud between your devices, you can store them offline on a hardware security key. I recommend both the Yubico Security Key C and the Yubico YubiKey 5C NFC in Wirecutter’s guide to security keys, based on price, durability, and brand reputation among the 12 keys I tested.

Top pick

This key works just about anywhere security keys are supported. It can store passkeys, and it supports NFC for wireless communication with your phone.

$29 from Amazon

$29 from Walmart

$29 from Best Buy

Upgrade pick

The Yubico YubiKey 5C NFC supports many authentication protocols, so it works anywhere security keys are accepted. If you can make the most of its advanced features, such as signing and encrypting with OpenPGP, it’s well worth the price.

$58 from Amazon

$58 from Walmart

$58 from Best Buy

But remember that if you lose your physical security key, you lose all of its passkeys as well. If you go this route, you should create backup passkeys for your accounts and store them on a second, backup security key that you keep in a safe place.

Unlocking sites with a passkey

You can use passkeys only on sites that support them. The FIDO Alliance, the industry group that developed the core technology for passkeys, has a searchable list of sites that use passkeys, which is a good place to start. But you are likely to see a prompt pop up on any of the major online services and retailers that are now using passkeys.

The passkey experience differs slightly depending on the device you’re using and the site you’re logging in to.

Some websites, like Amazon, will prompt you to create a passkey. On a phone, tap Add Passkey and authenticate using the same method you use to unlock your phone (in my case, with Face ID). That’s it — your phone has created a unique passkey for Amazon and securely saved it.

The next time you log in to Amazon, your phone will recognize that you have a passkey and ask to use it. (On desktop computers, you may have to enter your username before you can use a passkey.) Tap the button to confirm and authorize with biometrics or a PIN, and you’re logged in. No password, no problem.

You can also use a device with a passkey to log in to a website on a different device where you don’t have a passkey. After navigating to Amazon’s website on a laptop, for example, enter your username and then choose the option to use a passkey. The browser will display a QR code that you scan with your iPhone, which in turn triggers a request to use a passkey. Tap to confirm, perform a quick Face ID scan, and that’s it: You’re logged in. Note that both devices must have Bluetooth enabled, but you don’t need to pair them.

An important detail: When you use a passkey on one device (like your phone) to authenticate a different device (such as your laptop), the passkey isn’t transferred from one to the other. If you plan on logging in again on the laptop in the future, you might want to create a new passkey specifically for the laptop.

The disadvantages of passkeys

The greatest obstacle to passkey adoption may end up being exhaustion. It’s yet another thing that’s being asked of you to keep your accounts safe.

“Security fatigue is real,” Hoffman-Andrews said.

Passkeys are still in their early-adoption phase, and it can be hard to find sites that support them. So for the foreseeable future, you’ll need to continue using password managers and enabling 2FA wherever possible. Meanwhile, some sites that do use passkeys still rely on passwords and 2FA, which clever attackers can obtain through phishing, as backups.

And while passkeys are intended to be more convenient than passwords, that’s not entirely the case at the moment. Some sites don’t require 2FA for passkey logins, and others do; some sites let you create as many passkeys as you like for an account, while others limit you to just one. Logging in to the same site on an iPhone doesn’t work quite the same as it does on Android. These are little things, but they create more confusion for people still unfamiliar with the technology.

I’ve also found that it can be tricky to remember where I’ve created my passkey, which sometimes means I simply don’t know what device to use to log in. It’s not currently possible to transfer passkeys between different types of devices, which is annoying, but Hoffman-Andrews explained that this limitation also helps keep passkeys safe from phishing attacks.

Although passkeys are stored with end-to-end encryption, which prevents both providers and third parties from accessing or examining them, they’re not entirely in your control when you store them on third-party cloud services like those of Apple and Google.

“It is technically possible that a compromised cloud account — or a legal subpoena directed at the cloud provider — could impact the security or privacy of those synced credentials,” said Hanson.

If someone steals your phone, or confiscates your laptop, or gains access to the account that manages your passkeys, they might be able to access those passkeys. That’s why you should never reveal your PIN to anyone. You should also enable manufacturer tools to protect lost devices and learn how to deactivate fingerprint login or facial recognition on devices. People with elevated risk, such as journalists, activists, and politicians, should consider using hardware security keys to store their passkeys.

The bottom line: Use passkeys when you can

Password phishing, data breaches, two-factor authentication, and increasingly lengthy strings of mixed-case alphanumeric nonsense phrases may soon be relics of our password past — just not yet.

We recommend that you create a passkey on any site that allows it. If you’ve previously dismissed the option to create a passkey, or if you aren’t sure whether a site supports them, there are a few places to look. On the login page, look for the option to create or use a passkey. You should also check the security options in your account settings page for information on passkeys. On Amazon, for instance, go to Your Account > Login & Security > Passkey.

Although passkeys haven’t won out against passwords, “that doesn’t mean the users won’t benefit from them today where they are already available,” Hanson said.

Every passkey you create makes your accounts, and your online life, a little safer. And someday, the magic of passkeys will be so commonplace, you won’t be able to imagine life without them.

This article was edited by Caitlin McGarry and Annemarie Conte.

Meet your guide

I write about how to get the most out of your phones and computers, whether that means keeping your photos neatly sorted or protecting yourself from scams and surveillance.“

Opinion | Mark Zuckerberg Is Running Meta Into the Ground - The New York Times

2026-05-08T10:10:09.796-04:00

Meta Is Dying

May 8, 2026,

Alex Gamsu Jenkis

By Julia Angwin

"Ms. Angwin, a contributing Opinion writer, is an investigative journalist.

There is a moment when internet companies get the stink of death on them. For AOL, it was 2003, when it became clear that its users were abandoning its clunky dial-up internet service for far-faster broadband. For Yahoo, it was 2015, when their last-ditch acquisition spree failed, and they sold themselves to Verizon.

For Meta, that time is now. I believe the company — one of the most powerful media organizations in the world and one of the most valuable members of the S&P 500 — is at the start of a long, slow decline that will trigger aftershocks to our economy and our society.

It may be named Meta, but the company’s biggest asset is still Facebook. Started from a Harvard dorm, the original online social network has dominated our world for two decades. Its three billion users are still bigger than any single country. Its platforms can help sway an election, fuel an insurrection or spark a genocide.

But if you look carefully, you can see chinks in the armor. Meta’s earnings are starting to show the strain from years of growing consumer disaffection and reckless spending. The latest earnings, released on April 29, revealed a dip in user numbers for the first time since it started reporting these figures. And the slumping stock confirms what we have all known in our guts for a while: This is a company entering its zombie era.

Death is different on the internet. Lifeless companies like AOL and Yahoo are still technically with us. You can visit their websites. They have customers. They may even be profitable, as they cut staff and monetize their last remnants of traffic. But they are, as the kids say, peak cringe. Many teens wouldn’t be caught dead with an AOL account, a Yahoo email address — or a Facebook profile.

Sign up for the Opinion Today newsletter Get expert analysis of the news and a guide to the big ideas shaping the world every weekday morning.

As a company’s brand ages, its founders leave. The excitement evaporates. The stock shrivels to a fraction of its former glory as the user base withers to those captured by an old email account or friend group. New owners often arrive — usually bean counters who are focused on cutting costs and maximizing profits. That’s when websites stoop to junk mode, spamming you with endless email “final sales” and loading up the pages with ads so gross and disturbing that they should be age-restricted.

Of course, Meta is a long way from hitting rock bottom. The online giant — which benefits from its ownership of WhatsApp, the world’s largest messaging app, and Instagram, the popular photo-sharing social network — made $200 billion in ad revenue last year. That was an astonishing 20 percent of the global ad market. Meta’s founder, Mark Zuckerberg, is still firmly at the helm thanks to an unusual ownership structure that prevents him from being fired.

Thanks to that, we will all get to watch Mr. Zuckerberg drive the company into the ground. From 2021 to 2026, he poured $80 billion into the Metaverse in the firm belief that we would all want to don headsets and hang out in a virtual world populated by legless avatars. Even after shutting that project down, the company still loses billions a quarter on projects like selling $500 “smart” glasses that are not only unpopular but also give major creep-filming-you-without-consent vibes.

While its adventures in avatars were going nowhere, Meta’s revenues still soared as even more ad dollars moved online in the pandemic. Then in 2022, the revolutionary chatbot ChatGPT burst on the scene, and Mr. Zuckerberg jumped into the A.I. race with an open checkbook. Pontificating about the democratization of A.I., he sank about $100 billion into building an A.I. model that anyone could run on their own machine. But last year, when that model turned out to be too slow, too inaccurate and too unwieldy for most people to operate on their own, Mr. Zuckerberg abandoned the effort and plunked down another $14 billion for a new team to play catch-up with the other leading A.I. models. Now Meta has said it will spend another $115 billion (minimum) over the next year into its new effort, which thus far performs worse than the competition.

Where is this money coming from? Increasingly, Meta has been using debt to fuel its spending, amassing $59 billion in long-term debt on its balance sheet by the end of 2025, double the prior year’s total. And that doesn’t count the “aggressive” accounting it has used to keep the cost of a $27 billion Louisiana data center off its books. “The spending growth looks increasingly unsustainable,” The Wall Street Journal’s “Heard on the Street” columnist Asa Fitch wrote this week.

Now, as the company careens from one staggeringly expensive misadventure to another, its cash-cow core business is starting to wear out. Last quarter, the number of daily active users across its properties declined for the first time to 3.56 billion from 3.58 billion.

When an aging business starts to take on water, the quickest, easiest — and most destructive — solution is to make moves that will generate more money now, but may cost the company later. And that’s exactly what Meta has started to do. In the first three months of this year, the company started cramming more ads onto its platforms while charging advertisers more. Those choices may have allowed the company to increase its revenue-per-user by a significant 27 percent in the first quarter of 2026, but they are also likely to further alienate users (and annoy advertisers).

At the same time, judges and juries are starting to penalize Meta for the societal harms of its products. In March, the company (alongside YouTube) lost a bellwether lawsuitalleging that its addictive design choices triggered anxiety, depression and body-image issues in a teen. Waiting in the wings are over 100,000 similar cases seeking claims in the tens of billions of dollars.

There is a grim satisfaction in watching this organization hoist with its own petard. This is the company that profited from trafficking in lies, that tuned its algorithms to boost hatred and division, that stole our data and used it against us, that created the culture of toxic memes that are now central to our degraded public discourse. The fall of Facebook could even be a sign of a heartening turn in our national conversation: TikTok traffics more in inspirational content — prom videos are currently trending — than in the divisive narratives Facebook fostered.

But in the continued absence of any meaningful regulation, history shows us that internet companies can still wreak a lot of damage when they are in decline.

As it was being outpaced by Google on nearly every front, Yahoo failed to invest in cybersecurity and fell victim to what is still the largest data breach of all time. In 2014, Russian hackers gained access to 500 million Yahoo accounts, targeting Russian dissidents and journalists while stealing gift card and credit card numbers.

Meta’s properties, which are already riddled with fraud and scams, are likely to get even worse, given that the company has been slashing its work force in key areas focused on A.I. safety and identifying dangerous and illegal content. That means its apps are likely to grow even more polluted with everything from A.I. deep fakes to child sexual abuse material.

And Meta is still Meta. Even after losing that bellwether case on its efforts to addict users to its platforms, Meta’s chief financial officer, Susan Li, recently bragged to Wall Street that the company is using A.I. to increase the amount of time users spend watching videos and interacting with content. Fortunately, given the company’s recent track record, there’s a good chance that at least some of these terrible ideas are likely to end up in the same graveyard where Meta’s other expensive flops are buried.

Meta may be dying, but rest assured it won’t go gently into that good night. Maybe that could be a good thing. The more users quit, and the more corroded Meta’s apps grow, the faster we can all log off and close this chapter of the social-media revolution forever.

Julia Angwin, a contributing Opinion writer and the founder of Proof News, writes about tech policy. She is the author of the forthcoming book “On Courage: How to Be a Dissident in an Age of Fear.”

Opinion | Mark Zuckerberg Is Running Meta Into the Ground - The New York Times

Why I Ditched ChatGPT For Claude (you should too)

2026-05-07T18:08:59.904-04:00

John Ternus is a different kind of tech CEO.

2026-05-06T16:25:34.494-04:00

Neil deGrasse Tyson: Give Us the Aliens

2026-05-06T11:49:22.380-04:00

Neil deGrasse Tyson: Give Us the Aliens

“Neil deGrasse Tyson, an astrophysicist, expresses his desire for the release of U.S. government files on aliens and UFOs to include an actual alien. He highlights the anticlimactic nature of the files, given prior testimonies and declassified information. Tyson also explores the human tendency to imagine aliens as humanoid, despite the vast diversity of life on Earth, and suggests that aliens might perceive humans differently based on our cultural norms and behaviors.

May 6, 2026, 11:00 a.m. ET

Illustration by Sam Whitney/The New York Times

By Neil deGrasse Tyson

Dr. Tyson is an astrophysicist and the author of “Take Me to Your Leader: Perspectives on Your First Alien Encounter.”

Ever since childhood I’ve wanted to be abducted by aliens. Now, as a professional astrophysicist armed with the knowledge of the size, age and composition of the cosmos, I know that nothing prevents any of us from imagining a universe teeming with life.

So the impending release of U.S. government files on aliens and U.F.O.s is a good thing, even if it feels like a distraction from other important files we’ve all been waiting to be disclosed. I expect the alien files will be anticlimactic. After a parade of alien insiders and whistle-blowers testified under oath to Congress in 2023, 2024 and 2025, what’s left to learn?

Personally, I’d be delighted if the files were accompanied by an actual alien. Alive or dead or undead. Preferably alive. Is that too much to ask for?

The whistle-blowers have already told us about the crashed flying saucers, extraterrestrial bodies and alien technology in our possession — hidden in undisclosed places. Not only that, but secret files have been declassified before. A 2017 headline in this newspaper was unambiguous: “Glowing Auras and ‘Black Money’: The Pentagon’s Mysterious U.F.O. Program.” And who could forget the Air Force’s Project Blue Book, which studied more than 12,000 U.F.O. sightings from 1952 until the project was terminated in 1969, with the goal of assessing threats to national security.

What’s clear, however, is that if an authentic alien walked out of the halls of Congress, nobody would ever again have to ask if you “believe” in aliens, just as nobody questions the existence of elephants. An alien of the alien files could become the literal elephant in the room.

Without good evidence of what actual aliens look like, we’re stuck imagining them. And imagine them we do. IMDb, an online database about entertainment, lists hundreds upon hundreds of films, TV shows, video games and documentaries about aliens — both friendly and evil. Mostly evil.

Disappointingly, in nearly all these portrayals, these aliens look a lot like us. They’re humanoid, with a head, two eyes, a nose, a mouth, a neck, shoulders, a torso, arms, fingers and legs. Remember that most life on Earth, with which we have DNA in common, looks nothing like us or any vertebrate animal. So we should expect aliens with no DNA in common — or no DNA at all — to look at least as different from humans as humans and other life-forms on Earth (like jellyfish or termites) look different from each other.

The only thing that would shock me about a living, declassified alien is if most Hollywood depictions ended up being right, violating everything we know about biodiversity on Earth and across the universe.

We care a lot about what aliens look like, but we don’t pay nearly enough attention to what we might look like to them. If an alien emissary landed in Los Angeles, for example, its first impression might be that Earth’s dominant life-form is the automobile. The city is heavily crisscrossed by major freeways, many of them 12 lanes wide. People line up in their cars on slow lines to obtain fast food handed through a window. They consume the food while still seated, never exiting their vehicles. Some of the larger life-forms on the freeway carry multiple automobiles within them. To the aliens, these car haulers are surely pregnant.

Assuming on arrival that the alien knew we were human, it would probably want to meet the person in charge. Who exactly would that be? The president? The prime minister? The pope? Or would it be a multibillionaire or captain of industry? Not knowing anything in advance about human civilization, but picking up clues from our cultural norms before arrival from leaked radio waves, an alien might instead expect to meet Ryan Gosling, Taylor Swift or Oprah Winfrey.

If we look more deeply into our own alien stories, there’s a persistent plotline that aliens are evil and want to kill us all. I suspect those fears are based not on what we believe about aliens but on what we know about humans.

In the history of our species, there’s no shortage of technologically advanced cultures that commit rampant violence against less-advanced ones. Within what we call civilization, humans oppress — or kill — one another over which creator of the universe they worship, or who they sleep with, or what side of an arbitrary line on Earth’s land masses they’re born, or how absorptive their skin is to sunlight, or what set of sounds comes out of their mouths.

Upon bearing witness to our irrational ways, any visiting alien that might have accompanied the release of the alien files surely long ago escaped back home to report, “There’s no sign of intelligent life on Earth!”

Neil deGrasse Tyson is an astrophysicist who is the director of the Hayden Planetarium at the American Museum of Natural History in New York City. He is the author of “Take Me To Your Leader: Perspectives on Your First Alien Encounter.”

Apple Reaches $250 Million Settlement Over Claims It Misled People on A.I.

2026-05-06T11:09:01.457-04:00

Apple Reaches $250 Million Settlement Over Claims It Misled People on A.I.

“Apple agreed to a $250 million settlement over claims it misled consumers about the capabilities of its artificial intelligence system, Apple Intelligence. The settlement resolves lawsuits alleging Apple oversold the features of the iPhone 16 and some iPhone 15 models. Apple denied wrongdoing but acknowledged the challenges in the global technology race to dominate A.I.

Some iPhone owners will be eligible to receive $25 to $95 over claims that the tech giant oversold its artificial intelligence system, Apple Intelligence.

May 5, 2026

The Apple store at Grand Central Station in Manhattan.Lucia Vazquez for The New York Times

Apple agreed on Tuesday to pay $250 million to settle legal claims that it misled consumers about the abilities of its artificial intelligence system, Apple Intelligence, according to court filings.

The settlement resolves a handful of class action lawsuits filed against Apple last year, which claimed the company oversold what its product could do during its rollout in 2024. Those suits were consolidated last year by the U.S. District Court for the Northern District of California, where a judge still needs to approve the settlement.

Consumers who purchased an iPhone 16 and some models of the iPhone 15 between June 2024 and March 2025 will be eligible to collect up to $95 per device, according to the filings. As part of the settlement, Apple denied any wrongdoing.

The settlement underscores Apple’s challenges in a global technology race to dominate A.I. The iPhone maker has largely sat it out, in part because it hasn’t built its own A.I. models like Google’s Gemini. Tech companies like Microsoft and Nvidia soared in value as they bet heavily on the technology.

Since “the launch of Apple Intelligence, we have introduced dozens of features across many languages that are integrated across Apple’s platforms,” Marni Goldberg, an Apple spokeswoman, said in a statement. “We resolved this matter to stay focused on doing what we do best, delivering the most innovative products and services to our users.”

Apple first teased Apple Intelligence in June 2024 as an answer to products like OpenAI’s ChatGPT. The company promised big improvements to its personal assistant, Siri, which has been part of its devices for more than a decade.

Apple also said it planned to introduce A.I. features to summarize notifications and offer help on improving writing in emails and text messages. In advertisements, the actor Bella Ramsey used Apple Intelligence to remember someone’s name and to catch up on an email.

But those features weren’t available on the iPhones that Apple shipped in September 2024. Instead, the company gradually rolled out the promised features and soon ran into problems. Notification summaries misrepresented news reports, for example, and Apple disabled that feature. In March 2025, Apple delayed the release of an upgraded Siri over quality problems.

Apple misrepresented the “capabilities of the series 16 iPhone and deceived millions of consumers into spending hundreds of dollars on a phone they did not need, based on features that do not exist,” according to one of the class action lawsuits.

In December, Apple announced the retirement of its head of A.I., John Giannandrea. In January, the company said it would use Google’s Gemini to power its A.I. products, including Siri.

David McCabe is a Times reporter who covers the complex legal and policy issues created by the digital economy and new technologies.

Kalley Huang is a Times reporter in San Francisco, covering Apple and the technology industry.“

DJI Osmo Pocket 4 Is a Great Vlogging Camera, but Not for the US

2026-05-06T00:03:49.648-04:00

DJI Osmo Pocket 4 Is a Great Vlogging Camera, but Not for the US

“From its new sensor, film profiles and built-in storage, the Osmo Pocket 4 has a lot to offer. Shame it won't be on sale in the US at launch.

Andrew Lanxon

April 16, 2026 5:01 a.m. PT

4 min read

Andrew Lanxon/CNET

I've filmed with my DJI Osmo Pocket 3 all over the world, including up in the frozen Arctic, so I was excited to test out the latest version, the Osmo Pocket 4. The camera doesn't make a lot of changes to what's already a superb vlogging setup, but there are some key upgrades, from better dynamic range to a suite of new filmic color profiles.

I've been using the camera for a few days, and I'm pleased with what I've seen so far. But before I get to the details, I must address the elephant in the room: It won't be officially on sale in the US when it launches. It's a murky situation; while DJI's drones have been banned for sale in the US, the company's other products are subject to FCC regulations that, at the very least, delay their launch.

Watch this: DJI OSMO Pocket 4 Hands-On Q&A

36:40

DJI's Osmo Nano, for example, still isn't sold directly via DJI's online store, but you can pick it up in major retailers like Best Buy and B&H Photo. For the Pocket 4, DJI states it "will not be available in the US market as the application for authorization is still pending." The word "pending" suggests that the situation could change, but as of right now, it might be tough to get hold of for US shoppers. And, even if you do get your hands on one, service and support could prove tricky.

There are two new buttons beneath the display; a digital zoom button that crops in 2x and a custom function button you can program to enable a variety of settings.

Andrew Lanxon/CNET

In the UK, the Pocket 4 will start at £445, and a Creator Combo (including an attachable fill light, wireless mic and battery handle) will set you back £549. For reference, those UK prices translate to $604 and $745, respectively.

DJI Osmo Pocket 4: What's new

Physically, you'd struggle to tell the difference between the new model and its predecessor. It's marginally bigger, but it maintains the 2-inch rotating LCD and the camera unit mounted on the gimbal above. The camera has the same 20mm field of view with the same f2 aperture. Beyond that, a lot has changed.

The camera still uses a 1-inch image sensor, but it's a new version that offers up to 14 stops of dynamic range for better results in high-contrast situations. Its resolution maxes out at 4K (like before), but it'll now shoot at up to 240 frames per second for silky smooth slow motion.

There's a variety of new color profiles built in, designed to help you get a cool-looking filmic tone for your clips without having to apply filters or presets in post-production -- just pick a tone and hit record. Or if you take your color grading more seriously, it'll shoot in 10-bit D-Log for greater flexibility than its predecessor offered.

A clip-on fill light is available as part of the Creator Combo. It could be handy for those times when you're trying to record a vlog in darkness. Perhaps when exploring a haunted house.

Andrew Lanxon/CNET

I've had a play with the presets and they're fun. A little much for my tastes, though you can turn the intensity down to look more natural. I think these are great for quick footage, but if I were shooting a more involved project where I know I want the best possible quality, I'd shoot everything in D-Log to apply my own filmic tones later -- and have full control over them as a result.

Despite the aperture remaining the same, DJI reckons the Pocket 4 has superior low-light performance, thanks to advances in image processing. I've not had a chance to take it out in the dark yet, but I'm looking forward to putting the old and the new side by side. There's also a new slow-shutter video mode for creative nighttime shooting, which again sounds like a fun thing to play around with.

The Pocket 4 has a slightly larger battery than the Pocket 3 (1,545 mAh, up from 1,300 mAh), along with fast charging support, which will apparently take it from empty to 80% full in just 18 minutes. It also has 100GB of built-in storage, unlike the Pocket 3. I love this: I can't remember how many times I've grabbed my camera for a day's filming only to later realize I left the microSD card back in my computer.

The 100GB of built-in storage is great for those occasions when you go out without a microSD card to hand. I do that often.

Andrew Lanxon/CNET

DJI Osmo Pocket 4 Pro: Where's the dual-lens model?

The Osmo Pocket 4 I've been testing is broadly similar to the Pocket 3, but with most specs bumped up slightly. But the rumor mill has been churning for months about a "Pro" version of the Pocket 4, which, according to alleged "leaked photos," will feature a larger camera module with multiple lenses -- likely including a telephoto zoom option, similar to the multilens setups seen on most phone cameras.

DJI has remained silent on the matter, with no mention of a Pro model in its launch materials; whether this more full-featured camera ever arrives remains to be seen. Personally, I'd love to see a Pro version feature an even bigger sensor and even interchangeable lenses, as DJI offers with its Ronin 4D. I won't hold my breath, though.

The various filmic color profiles are handy for adding a bit of extra flair to your footage.

Andrew Lanxon/CNET

DJI Osmo Pocket 4: Should you buy it?

If you already have the Pocket 3, I don't think there's a significant reason to upgrade to the new model (which is good news for potential buyers in the US). Sure, you might see marginal gains in dynamic range and perhaps battery life, but otherwise the video you'll be capturing will be broadly the same.

However, if you're still using the much older Osmo Pocket 2, or if you're considering getting one for the first time ahead of your upcoming summer vacation, then the Pocket 4 has a lot going for it. I'll be spending a lot more time testing it over the coming weeks and months to see how well it performs in both pro and enthusiast environments and how it compares with its new rivals, the Mission 1 series from GoPro.”

Apple Has a Mac Problem… and It’s About to Get Worse

2026-05-05T13:23:16.516-04:00

Neil deGrasse Tyson: Peter Thiel Is WRONG!

2026-05-04T16:09:47.692-04:00

Stop Paying the Apple Storage Tax — Do This Instead (2026 Mac Storage Guide)

2026-05-03T21:33:03.148-04:00

Stop Paying the Apple Storage Tax — Do This Instead (2026 Mac Storage Guide)

Reacting to "Why AI is so smart but also so dumb?" - YouTube

2026-05-02T11:38:36.988-04:00

Reacting to "Why AI is so smart but also so dumb?"

Harvard Just Caught AI Lying to Every Executive in America

2026-04-29T22:59:14.590-04:00

My Take on The New Apple

2026-04-29T21:37:10.600-04:00

(1256) M5 Mac Mini: 3 Huge Upgrades (And 1 Massive Problem)

2026-04-29T01:14:38.186-04:00

M5 Mac Mini: 3 Huge Upgrades (And 1 Massive Problem)

Musk vs Altman: Messy feud threatens to derail OpenAI’s IPO | Chris Stokel-Walker

2026-04-28T18:45:27.450-04:00

Musk vs Altman: Messy feud threatens to derail OpenAI’s IPO | Chris Stokel-Walker

Qwen AI Just Showed Why The US Can't Ignore China Anymore

2026-04-28T13:55:03.980-04:00

Qwen AI Just Showed Why The US Can't Ignore China Anymore

What Elon Musk’s Clash With Sam Altman of OpenAI Is Really About - The New York Times

2026-04-28T12:44:00.200-04:00

Et Tu, Brute? What Elon Musk’s Clash With Sam Altman Is Really About.

"Mr. Musk’s lawsuit against Mr. Altman and OpenAI makes the case that all-encompassing greed is Silicon Valley’s defining feature.

April 28, 2026

Illustration by Ben Denzer; Photographs by Eric Lee/The New York Times, Anna Moneymaker/Getty Images

One of the most controversial and overexposed men in the world is suing another man, who is equally unsympathetic and equally inescapable. Both are insanely rich.

It is so tempting to look away.

Elon Musk’s lawsuit against Sam Altman involves onetime colleagues and buddies who became peevish enemies. Now they would like to take each other down. Happens all the time. These guys just have more lawyers.

Ignoring this conflict would be a mistake, however. The rancorous dispute between Mr. Musk and Mr. Altman, which went to trial this week with opening statements in an Oakland, Calif., federal courtroom on Tuesday, goes to the heart of Silicon Valley, a place that has always cloaked itself in virtue.

Mr. Altman and Mr. Musk started working on what was supposed to be a different sort of tech lab in 2015. OpenAI was a Manhattan Project for artificial intelligence, a nonprofit venture that would act as a shield against rapacious behavior by less benevolent outfits. The goal was to “shift the dialog toward being about humanity winning rather than any particular group or company,” according to a document in the case.

Mr. Musk, the chief executive of Tesla, provided the initial funding. Mr. Altman was OpenAI’s leader and spokesman. But Mr. Musk says their interests quickly diverged when it became clear just how much money was up for grabs. OpenAI converted to a for-profit company last year. “A textbook tale of altruism versus greed,” Mr. Musk asserted in his suit’s opening salvo.

The fact that the person calling himself an altruist here is likely to become the world’s first trillionaire doesn’t necessarily make it untrue. In his lawsuit, filed in 2024, Mr. Musk said Mr. Altman, OpenAI president Greg Brockman and others “unjustly enriched” themselves in the development of OpenAI “to the tune of billions of dollars.”

OpenAI, whose value is approaching $1 trillion, had the inevitable response: No, you’rethe one who is greedy. The company argued that Mr. Musk walked away when he could not take over the entire enterprise.

“This case has always been about Elon generating more power and more money for what he wants,” OpenAI said in a statement.

One of the few things the moguls agree on is that their feud evokes the works of a certain Elizabethan playwright. Mr. Musk, 54, said in his suit that Mr. Altman’s “perfidy and deceit are of Shakespearean proportions.” Mr. Altman, 41, mused in a blog post this month that “there has been so much Shakespearean drama between the companies in our field.”

If there is a Shakespeare play that could sum up this soured friendship, it’s “Julius Caesar.” Brutus wants to stop Caesar from gaining too much power, or so he says. Caesar is quite surprised that he’s being assassinated by a supposed friend. “Et tu, Brute?” he cries. Brutus ends the play as dead as Caesar but is mourned as “the noblest Roman of them all.”

Mr. Musk should be so lucky to draw such praise.

‘For the Good of the World’

In the middle of the last decade, Mr. Altman was a Silicon Valley insider running the top start-up incubator, Y Combinator. Ambitious and persuasive, he didn’t want just to fund companies. He was on a mission to save humanity, which — unknown to the masses — was at great risk.

“I think A.I. will probably, most likely, lead to the end of the world,” Mr. Altman said in 2015. It was a fear he would often express. Why not, he asked, create a bulwark against the other A.I. companies “for the good of the world”?

Mr. Altman drew in Mr. Musk, who was even more worried about where A.I. was heading. “We are summoning the demon,” Mr. Musk once said.

Immediately, there was a problem. People everywhere work on nonprofit ventures for modest salaries. They sacrifice for their ideals. Mr. Altman knew that would not fly in Silicon Valley. The engineers and scientists would “get start-up-like compensation if it works,” he promised.

The nonprofit was dead almost before it began. OpenAI is owned by its employees and investors, including Microsoft, Amazon, Nvidia and SoftBank, as well as the OpenAI Foundation. (Mr. Altman has no direct equity in OpenAI but has other investments that make him comfortably a billionaire.) OpenAI is planning to sell shares to the public in one of the richest stock offerings in history.

Silicon Valley is the great wellspring of wealth in modern America. Nine of the 10 richest Americans are tech entrepreneurs, with Warren Buffett the only exception. People might be offended by OpenAI’s turnabout, but few could say they were shocked.

Except the richest man in the world, whose own A.I. venture, xAI, is now part of one of his other companies, SpaceX. SpaceX will soon sell shares to the public as a decidedly for-profit operation.

No Happy Ending

Tech companies are subject to relatively few constraints these days. Congress is generally passive. Federal regulators have been hobbled. The Trump administration is stocked with venture capitalists and others receptive to tech and its money, as is President Trump.

What’s left for tech opponents are civil suits. Social media companies face an onslaught of cases. One of the first, in Los Angeles last month, found that Meta and YouTube were to blame for anxiety and depression in a young woman who was a heavy user.

“Trials are all we have right now, and things are better because of them,” said Max Tegmark, a co-founder of the Future of Life Institute, a nonprofit trying to reduce catastrophic technology risks. “Trials provide information that is not otherwise accessible.”

The exhibits in the Musk/Altman trial are an example of material that presumably would never have seen the light otherwise. That includes emails between the two leaders as they tried getting OpenAI off the ground.

“Do you have any objection to me proactively increasing everyone’s comp by 100-200k per year?" Mr. Altman wrote to Mr. Musk in 2015. “I think they’re all motivated by the mission here but it would be a good signal to everyone we are going to take care of them over time.”

The Future of Life Institute gives OpenAI an overall grade of C plus for safety while xAI got a D. “A.I. is less regulated in America than sandwiches,” said Mr. Tegmark, who is also a physics professor at the Massachusetts Institute of Technology. “You can’t open a sandwich shop without having your kitchen inspected. But you can release an A.I. girlfriend for 11-year-olds and that’s fine.” A defeat for OpenAI might begin to change that, he said.

Some A.I. watchdogs said they would like to see OpenAI brought to justice the way Meta and YouTube were. But they would prefer almost any plaintiff to Mr. Musk.

“I don’t have long-term faith in a system where we’re legislating through private litigation,” said Sacha Haworth, executive director of the Tech Oversight Project, a Washington-based advocacy group. “I don’t want to rely on a billionaire with a grievance.”

If Mr. Musk wins, she pointed out, it would weaken or even destroy OpenAI, “opening up a large share of the market that an Elon Musk company can then gobble up.”

And if OpenAI gets the suit dismissed? “It would send a signal that it’s OK to launch as a nonthreatening nonprofit working for the public’s benefit and then cynically change to a for-profit without any accountability,” she said.

Ms. Haworth’s conclusion: “There’s no happy ending here.”

(The New York Times sued OpenAI and Microsoft in 2023 for copyright infringement of news content related to A.I. systems. The two companies have denied those claims.)

Some critics are worried that, in the worst scenario for OpenAI, its charitable arm would shutter. That, they said, would wipe out a very large foundation that could have helped people. Mr. Musk says he will give any damages he receives to the foundation.

Others take a more benign view.

“The law doesn’t rely on you being a good person to act in the public interest,” said Shaoul Sussman, a former official with the Federal Trade Commission. “A lot of the dirty laundry of OpenAI is going to come out.”

In a different environment, Mr. Musk’s pursuit of OpenAI might have been brief, ending with a tip to regulators. But he is not keen on government oversight, which during the Biden administration produced investigations and enforcement actions into his companies.

Instead, Mr. Musk’s case against OpenAI uses a legal doctrine called ultra vires, which means “beyond the powers.” It holds that a corporation is restricted to activities defined in its charter. This approach was widely used in the early 19th century when the federal government was small and weak and only a competitor could rein in your company.

Most corporations now have wide-ranging charters that allow them to pursue multiple goals. But there is one exception: nonprofits.

“This is the first high-profile case that I know of being pursued under these statutes for 100 years,” Mr. Sussman said.

In a trial expected to last several weeks, the very old laws will meet the very new technology. As Shakespeare said, what’s past is prologue.

David Streitfeld writes about technology and the people who make it and how it affects the world around them. He is based in San Francisco."

What Elon Musk’s Clash With Sam Altman of OpenAI Is Really About - The New York Times

I Ditched My iPhone for an Android For a Month: It Was Eye-Opening

2026-04-28T00:35:01.453-04:00

LLMs are a Dead End - YouTube

2026-04-27T22:34:16.973-04:00

LLMs are a Dead End - YouTube

Anthropic's Claude Mythos Just Broke Containment (Full Story)

2026-04-27T12:45:47.013-04:00

My Honest Thoughts about Deepseek

2026-04-27T11:42:39.078-04:00

My Honest Thoughts about Deepseek - YouTube

AI Chatbots: Last Week Tonight with John Oliver (HBO)

2026-04-27T10:15:37.157-04:00

Elon Musk and Sam Altman face off in court over OpenAI’s founding mission

2026-04-27T09:53:27.289-04:00

Elon Musk and Sam Altman face off in court over OpenAI’s founding mission

“Elon Musk is suing Sam Altman and OpenAI, alleging breach of contract and unjust enrichment. Musk claims Altman betrayed OpenAI’s founding mission by restructuring it as a for-profit entity, seeking damages and the removal of Altman and other executives. OpenAI denies the allegations, asserting Musk agreed to the for-profit transition and is motivated by jealousy.

Musk’s lawsuit accuses Altman of fraud, while OpenAI says that Musk is ‘motivated by jealousy’

Blake Montgomery, Dara Kerr and Nick Robins-Early

The two Silicon Valley tycoons are headed to court. Composite: Invision/AP, Reuters

A lawsuit between two of Silicon Valley’s biggest tycoons goes to trial Monday in California, the culmination of a years-long bitter feud. Elon Musk has accused Sam Altman of betraying the founding agreement of the non-profit they started together, OpenAI, by changing it to a for-profit enterprise.

Musk accuses Altman, OpenAI, its president Greg Brockman, and its major partner Microsoft of breach of contract and unjust enrichment in the lawsuit. Jury selection is scheduled to begin Monday morning at a federal courthouse in Oakland, with opening arguments from both sides expected later this week. The trial is slated to last two to three weeks. Along with internal communications from Musk and key executives at OpenAI, the trial promises a who’s who of Silicon Valley on the witness stand, including Musk, Altman and Microsoft CEO Satya Nadella.

OpenAI has vehemently denied Musk’s allegations, saying that he agreed in 2017 that establishing a for-profit entity would be a necessary next step for the company and that Musk is “motivated by jealousy” and “regret for walking away”. The company also contests that Musk’s funding was an investment, stating that it was instead a tax deductible donation to the non-profit and does not entitle him to ownership in OpenAI.

The case carries sizable stakes for OpenAI, which is expected to go public later this year at around a $1tn valuation. Musk is seeking a range of remedies that include the removal of Altman and Brockman and more than $134bn in damages, which the tycoon says would be redistributed to OpenAI’s non-profit arm. He also wants to reverse the company’s restructuring as a for-profit entity.

Altman, Musk and several other founders launched OpenAI in 2015 as a non-profit organization, with Musk providing about $38m. Altman’s relationship with Musk turned sour around 2017, after the billionaire grew impatient with OpenAI’s progress and made a failed bid to exert more control over the company. He left OpenAI’s board in 2018 and did not offer any more funding.

During OpenAI’s post-Musk years, it launched the wildly successful ChatGPT, raised tens of billions of dollars from Microsoft and grew to be one of the world’s most valuable private companies. Altman became the face of the AI boom. As the startup sought even more investment in 2025, it gained final approval from regulators to restructure its main business into a for-profit corporation, though one technically still overseen by the original non-profit.

Musk’s suit alleges that Altman’s dealmaking and maneuvering of OpenAI break with the fundamental mission of the company as a non-profit to benefit humanity and amount to a breach of contract. The suit also claims Altman and Brockman unjustly enriched themselves through their control of the company.“