Eyerys - Articles

'Gemini Omni' From Google Can 'Create Anything From Anything': Huge Gain But Not Quite There Yet

Aditya — Thu, 21 May 2026 14:52:31 +0000

The AI landscape has been defined by the relentless large language models (LLMs) war.

Ever since OpenAI's ChatGPT debuted and reshaped expectations for what conversational intelligence could achieve, the battle involving giants that shook the entire tech sphere. And Google, has astonished the world with Gemini.

Then, the powerhouse in search and foundational models, has also astonished the world with its Veo video generation models, particularly the Veo 3 and Veo 3.1, which delivered photorealistic clips with sounds that seemed poised to dominate the emerging field of generative media.

Those early demonstrations highlighted Google's deep expertise in rendering realistic motion and environments, positioning the company as a frontrunner in turning text and images into coherent video sequences that captured everyday physics and visual detail with surprising fidelity.

Yet the competitive dynamics shifted rapidly as specialized tools like ByteDance's Seedance 2.0 and Alibaba's HappyHorse 1.0 emerged from other developers, pushing the boundaries of multimodal video creation with superior motion fidelity, audio synchronization, and creative control that left many observers noting how Google's earlier leads had been eclipsed.

Then comes 'Omni,' which according to Google in the announcement, is literally an "any-to-any" model that can create "anything from anything," and "everything from everything."

Omni brings together an improved understanding of physics with Gemini's knowledge of history, biology, and culture, bridging the gap from photorealism to meaningful storytelling.

Actions have consequences, environments respond to events, and narratives evolve logically. pic.twitter.com/ajQ3purg0g

— Google DeepMind (@GoogleDeepMind) May 19, 2026

Seedance 2.0 excelled in handling combined inputs of text, images, video clips, and audio to produce cinematic sequences with director-level precision over camera movements, lighting, and narrative flow, smf Happyhorse 1.0 stood out for its top-ranked performance in blind human evaluations, generating dialogue-driven scenes with natural lip sync and immersive sound that felt remarkably lifelike.

These models raised the bar not just on output quality but on usability, allowing creators to iterate on complex, multi-shot stories without restarting from scratch, and they quickly became the benchmarks against which every new release was measured.

With the introduction of Gemini Omni, Google appears to be mounting a serious effort to catch up in this fast-evolving space, integrating the reasoning prowess of its Gemini models with advanced generative capabilities to enable more intuitive video editing and world-aware storytelling.

You can even reimagine the action in a video you took by asking Gemini Omni.

Transform your world instantly - change the environment, add new objects, or create something completely unexpected. pic.twitter.com/nMUyN7jNzW

— Google DeepMind (@GoogleDeepMind) May 19, 2026

And thanks to Gemini, this Omni includes features like conversational editing that let users reimagine a clip by describing changes to environments or objects while preserving the original intent.

Gemini Omni doesn't just build scenes that look real, it reasons about what should happen next. It combines an intuitive understanding of physics with Gemini's knowledge of history, science, and cultural context.

Rolling out today starting with video outputs to Google AI Plus,… pic.twitter.com/EkLjv5O0dN

— Sundar Pichai (@sundarpichai) May 19, 2026

And also thanks to Gemini, users can use Omni to create their own avatars.

Create videos with your own voice and likeness using avatars with Gemini Omni.

When you create an avatar, you have an AI digital version of yourself so you can easily generate videos that look and sound like you. No need to upload your image every time.

— Google Gemini (@GeminiApp) May 19, 2026

Announced as the first step toward a system that can create anything from any input, Omni Flash focuses on video for now but emphasizes native multimodality, letting users feed in text prompts, reference images, audio clips, or existing footage and then refine outputs through natural conversation.

It promises stronger world understanding, where actions follow logical consequences, characters remain consistent across scenes and lighting conditions, and cultural or scientific context informs how narratives unfold, bridging pure photorealism with something closer to meaningful storytelling.

The model is already rolling out in the Gemini app, Google Flow, and YouTube Shorts for subscribers, with API access planned soon, and

Gemini Omni can create anything from any input, starting with video.

This means you can combine images, audio, video and text as input and generate high-quality videos.

Or use drawings to create in a way that matches your vision.#GoogleIO pic.twitter.com/JIQW7uGfAZ

— Google (@Google) May 19, 2026

Early demonstrations show Omni Flash producing clips with improved physics simulation and continuity, such as characters interacting naturally with their surroundings or styles being applied seamlessly from reference inputs.

Gemini Omni is so fun - insanely great at editing videos! pic.twitter.com/KDGzcU78Wt

— Josh Woodward (@joshwoodward) May 19, 2026

Yet while these advances represent a clear step forward for Google in blending intelligence with media generation, they have not yet closed the full gap left by the current leaders.

Comparisons shared shortly after launch highlight that Omni still trails Seedance 2 in overall cinematic polish and motion stability during extended sequences, and it falls short of Happyhorse 1's benchmark dominance in realistic character animation and synchronized audio.

Some commenters even said that Omni also trails behind Kuaishou's Kling 3.0.

It's worth noting here that Seedance, HappyHorse, and Kling are all developed by Chinese companies.

Some outputs reveal lingering inconsistencies in complex physical interactions or subtle artifacts that more mature competitors have largely resolved, suggesting that Omni excels more at editable, context-aware refinement than at generating flawless video from the ground up on its first attempt.

Your browser does not support the video tag.

In a field moving at breakneck speed, Google's latest entry keeps the company firmly in the race.

While its debut claims no outright victory, it underscores how quickly the balance of power can tilt even among the biggest players.

Published:

21/05/2026

News

Google

Omni

Gemini

Trends

Review

WordPress 7.0 'Armstrong' Released: Offering Greater Flexibility In Web Development

Aditya — Thu, 21 May 2026 14:45:05 +0000

Today, in the era of social media and AI, the web has become more interactive, intelligent, and deeply integrated into everyday life.

And WordPress is still one of the most widely-used content management systems out there. With 'WordPress 7.0,' named "Armstrong" after the influential jazz musician Louis Armstrong, the platform wishes to provide users with greater flexibility to shape their online presence in personal and creative ways.

This major update introduces foundational AI capabilities into the core platform for the first time.

A new WP AI Client enables WordPress to connect with generative models from providers including OpenAI, Google Gemini, and Anthropic Claude.

WordPress 7.0 “Armstrong” is here.

This major release introduces foundational AI tools, a refreshed admin experience, expanded design controls, new blocks, and powerful developer APIs.

Explore what’s new, update when you’re ready, and start building with WordPress 7.0 today.… pic.twitter.com/eIVtWblxo5

— WordPress (@WordPress) May 20, 2026

These connections are managed centrally through a Connectors screen in the dashboard, where API keys can be added and organized securely.

The system supports features such as generating post titles and excerpts, suggesting alternative text for images, and working with image generation or editing tools. An accompanying AI plugin extends these options further, while a new client-side JavaScript package brings abilities directly into the browser environment, complete with a command palette interface for quick access.

The administrative interface has undergone a significant modernization.

The dashboard now features an updated color scheme, smoother transitions when navigating between sections, and a cleaner overall appearance that improves daily use. A global command palette, triggered by Ctrl+K on Windows or Cmd+K on Mac, allows users to search for and access tools from anywhere in the admin area without hunting through menus.

Font management has been unified into a dedicated section that works across block-based, hybrid, and classic themes, simplifying the process of installing, uploading, and organizing typefaces.

Revisions now include a visual scrubbing tool that displays changes at a glance, making it easier for editors to review versions side by side and restore earlier states with greater precision.

Several new blocks and design tools expand the possibilities within the editor.

The gallery block gains support for lightbox slideshows, allowing images to be viewed in an expanded, interactive format. Additional blocks for headings, breadcrumbs, and icons provide more granular control over common elements.

Responsiveness options have been enhanced so that blocks can be shown or hidden based on specific device types without impacting other viewports, and users can define custom breakpoints to fine-tune layouts.

Navigation overlays can now be constructed entirely with blocks and patterns, offering full styling control over columns, typography, and even custom close buttons.

Patterns themselves function more flexibly as cohesive units that can be inserted, detached, and modified individually. Custom CSS can also be applied directly at the block level within the editor, giving designers more immediate control over details.

On the development side, WordPress 7.0 introduces tools aimed at speeding up and simplifying building processes.

Blocks and patterns can now be created server-side using only PHP, with automatic registration through the block API. The Site Editor benefits from improved extensibility, including better routing support and a new boot package that allows plugins to construct custom pages within the editor environment. These changes, combined with various API enhancements and performance improvements, provide developers with a more capable foundation for extensions and custom solutions.

The release also raises the minimum PHP requirement, dropping support for older versions to encourage more secure and efficient hosting environments.

More than 875 contributors from countries around the world participated in bringing WordPress 7.0 to completion, among them over 200 first-time contributors.

Their combined efforts resulted in roughly 420 enhancements and bug fixes. Matias Ventura served as release lead, supported by a coordinated team handling release coordination, technical leadership, triage, and testing. The collaborative nature of the project remains central to its ongoing development as open-source software.

Sites running earlier versions can update through the standard process in the dashboard when administrators determine the timing is appropriate, ideally after testing in a staging environment to ensure compatibility with existing themes and plugins.

Published:

21/05/2026

From Consumer To Victim: The Man Who Loved Watching Voyeur Porn Discovers He And His Girlfriend Are Now The Main Attraction

Bess — Thu, 21 May 2026 10:12:26 +0000

06/02/2026

In the quiet Shenzhen hotel room, Eric and his girlfriend Emily thought they were sharing nothing more than a private escape.

They unpacked their bags, settled in for the night, and let the world fade away as they moved together in the kind of unguarded intimacy that most couples assume stays locked behind closed doors. But no. Unknown to them at the time, they were actually the stars of explicit content, streamed for paying subscribers and accessible to strangers worldwide.

Eric only realized this months later.

He was scrolling through the same Telegram channel he had long used to browse pornography, and suddenly he froze.

There they were, entering the room, setting down their luggage, undressing, and then having sexual intercourse, captured in crisp, unfiltered detail, then edited into an hour-long clip broadcast to thousands of strangers.

Eric and his girlfriend Emily. The couple have become victims of underground voyeurism in China. Traumatized, they wear hats in public out of fear of recognition.

"What drew me in is the fact that the people don't know they're being filmed," Eric later recalled of his old (and bad) habit. "I think traditional porn feels very staged, very fake."

Eric acknowledged, and confessed, that he was an avid viewer of this kind of voyeurism, and never knew that he and his girlfriend would become the victims themselves.

Now, realizing that he and his girlfriend are on the other side of the camera, the thrills had curdled into horror.

When he told Emily about this, her first thought that he was joking.

But after she watched the footage herself, she too was mortified. Emily became terrified that colleagues or family might stumble across it. The couple didn't speak to each other for weeks afterward.

When they finally reconciled, they went out wearing hats, constantly scanning faces in crowds, wondering who might recognize their most vulnerable moments turned into public entertainment.

Telegram is the go-to platform for this underground spy-cam trade and similar illicit content networks in China for a combination of practical, technical, and regulatory reasons.

What makes this harrowing is the fact that the footage wasn't a one-off hack or a bitter ex's revenge.

It was a snapshot of a sprawling, industrial-scale spy-cam operation thriving inside China's hotel industry, even though producing or distributing pornography has been strictly illegal there for years. An in-depth BBC investigation, published in February 2026, peeled back the layers on this underground trade, revealing at least 180 hotel-room cameras feeding live streams and archived clips across six different websites and apps.

Many of these operations are heavily promoted through Telegram groups, some swelling to as many as 10,000 members.

"Mask Park Treehole Forum" was one of the most prominent. The illicit Telegram chat channel that operated until it was exposed and shut down had over 100,000 members, predominantly Chinese-speaking men.

Subscribers in those Telegram chats aren't passive; they build entire communities around the thrill, trading tips on the best feeds, celebrating "good" performances, and bonding over the shared secret. It's a dark mirror of how mainstream social media works, except the currency here is stolen intimacy.

The cameras themselves are ingeniously discreet: some no larger than a pencil eraser, wired professionally, and directly into the room's electrical system, and set to activate the moment a guest inserts their key card into the power slot.

One journalist tracked a device to a hotel in Zhengzhou, central China, hidden inside a wall ventilation unit with its lens aimed squarely at the bed. When the team disabled it, word rippled instantly through the Telegram channel.

"Zhonghua got taken down!" one subscriber wrote, referring to the camera by its code name.

The operator, known online as "AKA," replied: "It’s such a shame; that room has the best sound quality!"

But the mourning was short-lived.

Within hours, AKA posted that a replacement camera in another hotel was already live. "This is the speed of… [our livestreaming platform]," he boasted to his subscribers. "Impressive right?"

Photos shared online through the Telegram group "MaskPark Treehole Forum".

The mechanics are ruthlessly efficient.

Subscribers pay a monthly fee, which is around 450 yuan, or about $65, in order to access multiple live feeds showing several hotel rooms at once. Subscribers can rewind the streams from the moment the electricity kicks in or download edited clips from vast archives stretching back to 2017, one of which held more than 6,000 videos.

In the live chats, the audience doesn't just watch in silence; they gossip in real time, rating bodies, making lewd comments, speculating about the guests' conversations, and cheering like sports fans when a couple starts having sex.

Lights off draws groans of complaint.

Women in the footage are routinely described with crude slurs, like "sluts," "whores," "bitches."

This so-called business is profitable, and AKA, one of roughly a dozen agents the BBC identified, wasn't working alone. He answered to "camera owners" higher up the chain, the people who install the devices and run the platforms.

Based on membership numbers and fees, the BBC estimates AKA alone pulled in at least 163,200 yuan, or about $22,000, in just a few months. This is about four times China's average annual income of 43,377 yuan. I

t’s a business model built on volume, low overhead, and the relentless churn of new victims.

Hidden cameras, carefully positioned, capturing exactly what strangers are willing to pay for.

The demand driving this ecosystem isn’t some niche fetish confined to the shadows of Chinese social media. It’s a lucrative slice of the global adult entertainment industry, which continues to balloon in value even as laws tighten in certain regions.

Recent market analyses peg the digital adult content sector at around $44 billion to $70 billion in 2025-2026, with projections climbing toward $75 billion or even $200 billion by the early 2030s, fueled by subscription models, mobile access, and ever-more-personalized feeds. But what sets voyeur or "spy-cam" content apart is its promise of raw authenticity.

While studio-produced porn often feels polished and performative, and that they create scripted moans, perfect lighting, actors and actresses who know exactly where the cameras are. Hidden-camera footage sells the opposite illusion: unscripted, unaware participants going about the most private acts of their lives.

That sense of forbidden intrusion taps into something primal: a mix of curiosity, power, and the adrenaline rush of transgression.

Your browser does not support the video tag.

Long story short, what keeps the wheels spinning is the market's insatiable appetite for the "real."

Traditional porn may satisfy some, but voyeur content delivers a different high: the knowledge that these are ordinary people: flawed bodies, awkward moments, genuine reactions, and the fact that they're unaware that they're being watched performing

On major tube sites, hidden-camera and voyeur categories have ranked among the most searched for years, drawing hundreds of millions of views monthly.

The hardware side has exploded alongside it. Miniature spy cameras, disguised as smoke detectors, chargers, or air fresheners, sell for pennies on electronics markets like Shenzhen's Huaqiangbei, where anyone with basic know-how can stock up and turn a quick profit.

Global projections for the broader spy-camera and surveillance gadget market hover around $300 million to $550 million in the coming decade, but the real money is in the content it generates.

China’s operation stands out for its seamless integration of supply, distribution, and monetization, but the problem is hardly isolated.

Similar spikes in voyeurism crimes have been reported worldwide, including in the West, as well as neighboring countries, like South Korea, where the term "molka" has become an endemic. In many of these countries, the police had logged thousands of incidents in homes, Airbnbs, changing rooms, and public spaces, including toilets.

Victims span the spectrum: couples mid-act, women in showers, families changing clothes, and even inside toilet bowls.

In other words, what are supposed to be normal acts of human beings, assumed to be performed alone or with a trusted partner, are now objectified for strangers' entertainment.

A screenshot of one of the possibly countless of 'molka' voyeurism videos depicting women inside public toilet cubicles in South Korea.

And here, Eric and Emily aren't the only victims. They're far from alone, but are part of only a handful of people who shared their story.

While tech platforms bear heavy responsibility, and that Telegram insists it moderates non-consensual content and bans it explicitly, yet channels like these persist for months, sometimes years, raking in fees before any action is taken. If it's taken at all.

As for the hotels, they are known for their slow response, probably due to insiders who profit from this kind of business, despite new Chinese regulations introduced in April 2025 requiring regular sweeps for hidden devices.

Since enforcement lags, replacements go up within hours, and budget properties remain especially vulnerable, travelers are always at risk.

They too have created communities to share their survival tips: like urging others to sweep rooms with phone flashlights to catch lens glints, probe ventilation grates and outlets, download RF detectors or Wi-Fi-scanning apps, or simply avoid unverified spots in high-risk cities. Some couples even pitched the idea of setting up a tent in their rooms, or book hotels only at chains with verified privacy protocols.

Portable privacy screens and white-noise machines have become unexpected travel essentials.

Read: How 'Molka' Voyeurism Through Spy Cameras Has Become 'A Part Of Daily Life' In South Korea

The human cost is the part that rarely makes it into the glossy market reports.

For every Eric who once consumed this content and now recoils from it, there are countless others whose lives are quietly upended. Bodies that were meant to be shared only with a partner become public property, turned into fapping material, dissected in comment threads, replayed for profit long after the guests have checked out.

The industry literally doesn't care about consent or consequences. After all, business is business, and all it cares about is retention and recurring revenue.

It’s worth noting that, more often than not, the footage was recorded by men. They were frequently the ones uploading the videos online as well, distributing them across platforms like Telegram.

Telegram itself is blocked by the Chinese government, meaning users in China typically need a VPN to access the platform.

"This has heightened the concerns of many women, as voyeuristic incidents seem to be ubiquitous," said Huang Simin, a Chinese lawyer who specializes in sexual violence cases.

Vivaldi Introduces Version 8.0, Which Makes The Browser Feels Like A Different Kind Of Browser

Irvan — Wed, 20 May 2026 14:10:12 +0000

Web browsers, the gateways to the web, have been evolving since the first time they were introduced.

While major browsers like Chrome, Firefox, Edge and Safari command huge percentage of the market, Vivaldi standa apart primarily through its extreme focus on customization and built-in productivity tools designed for power users. While Chrome and Edge prioritize simplicity and integration with their respective ecosystems, Vivaldi treats the browser as a highly adaptable workspace.

Users can rearrange almost every element of the interface, including toolbars, buttons, tab positions, and panels, without relying on extensions.

This level of control extends to themes, layouts, and behaviors that adapt to individual workflows, something the major browsers offer only in limited ways.

Vivaldi also boasts tab management, privacy features, built-in apps, and more. And since it's a Chromium-based browsers, Vivaldi delivers full compatibility with Chrome extensions and modern web standards, but it avoids the data collection practices common in Google Chrome and Microsoft Edge.

And now, comes version 8.0, which is an update that looks more like a huge upgrade.

New Vivaldi just dropped.
And by new, we mean NEW pic.twitter.com/9LNRJnhcjr

— Vivaldi Browser (@vivaldibrowser) May 21, 2026

First off, Vivaldi version 8.0 for desktop brings its most substantial visual update in years.

According to the announcement, the browser now uses a Unified design approach where toolbars, tabs, panels, and other elements sit on a single continuous surface rather than in separate layered regions. This creates a more cohesive appearance with precise alignment and intentional spacing that makes the interface easier to navigate and read.

Themes integrate more deeply across the entire window. Backgrounds and colors flow without interruption from the tab bar through panels and edges. Translucency and blur effects blend wallpapers into the environment more naturally. New default themes include options like Zen, Soria Moria, Sunset Forest, and Kawaii Clouds, alongside refreshed light and dark variants.

Users with custom themes can choose the new Unified coloring mode or stick with previous styles through the theme editor settings.

One of the several themes introduced in Vivaldi 8.0.

Onboarding and layout options have expanded to help both new and returning users.

Six preset layouts are available right at startup and remain accessible later in Settings under Appearance.

Simple provides a clean top-tab experience with minimal distractions. Classic retains the familiar toolbar and panel arrangement now updated with the Unified look.

Vertical Left and Vertical Right shift tabs and address bar to the sides for better use of wide screens. Bottom places navigation at the lower edge, while Auto Hide keeps the interface out of view until the cursor reaches a screen edge, maximizing content space.

Vivaldi 8.0 introduces six preset layouts, available right from onboarding.

These layouts serve as starting points that users can further customize using Vivaldi's extensive tools. Tab management remains powerful with stacking, tiling, and grouping options that adapt to individual workflows. Panels, workspaces, and built-in features like mail and calendar stay fully accessible and integrated.

The update also includes backend improvements such as a rewritten tabs management system, combined site preferences and tracker blocker dialogs, and refinements to translation and permission handling.

Since Chromium itself has been updated to version 148 for better compatibility and security, Vivaldi is able to bring various fixes address mail functionality, tab behaviors, auto-hide interactions, and platform-specific issues across Windows, macOS, and Linux.

The result is a browser that feels more unified in structure while preserving the depth of customization that has defined Vivaldi over time.

While AI dominates the headlines and social media consumes ever more of people's attention, Vivaldi 8.0 aims to make the good old web feel "worth coming back to."

https://t.co/AJ1hcxEUir

— Vivaldi Browser (@vivaldibrowser) May 21, 2026

Published:

21/05/2026

Sophie Rain Allegedly Turning Down A $15 Million Offer For Her Virginity Reveals The Billion Dollar Economics Of Internet Fantasy

Bess — Tue, 19 May 2026 10:44:37 +0000

19/05/2026

Sophie Rain has built an internet empire not only from physical attraction or character, but also on contradiction.

The 21-year-old Florida creator became one of the most talked about names on OnlyFans after repeatedly claiming to be the most highly-paid creator, while still a virgin. Her statements transformed what could have been a private personal detail into the foundation of a massively profitable online brand.

Now, she says a billionaire allegedly offered her $15 million for the rights to her first time, an offer she claims she rejected immediately because she has "ethics and morals" to uphold.

The number sounds outrageous, but the business logic behind it is not.

Sophie understands the economics of scarcity better than most creators online.

In a digital landscape oversaturated with explicit content and unlimited access, rarity becomes valuable. The less accessible something feels, the more audiences obsess over it.

Sophie Rain, born Izabella Blair on September 22, 2004, is an American internet personality, OnlyFans creator, and co founder of the viral creator collective Bop House.

Raised in Florida, she first gained mainstream attention in 2024 after claiming she earned more than $43 million during her first year on OnlyFans while still being a virgin and a devout Christian. Before becoming one of the platform's most recognizable creators, she reportedly worked minimum wage restaurant jobs and has spoken publicly about growing up on food stamps.

Her online persona, built around the contradiction between religion, sexuality, exclusivity, and internet fame, quickly turned her into one of the most controversial and commercially successful figures in the modern creator economy.

That is why Sophie Rain is not simply selling explicit content.

She is selling anticipation, fantasy, and the illusion that some part of her remains unattainable. Her virginity is no longer just a personal detail. It has become part of the product itself.

The mystery surrounding her identity is what keeps people subscribed, debating, reposting clips, making reaction videos, and emotionally invested in her story.

As long as audiences continue believing that narrative and attaching value to it, the market value of her brand continues to grow.

View this post on Instagram

A post shared by Sophie Rain (@sophieraiin)

That is what makes the alleged $15 million offer so fascinating. It was not merely an offer for sex. It was an offer to destroy the very narrative that helped make her one of the most profitable creators on the internet.

From a purely financial perspective, accepting it may have secured instant wealth, but long term, it could have damaged the mythology surrounding her brand. Sophie herself has acknowledged that her virginity plays a major role in why fans subscribe to her content, and reports about enormous offers tied to it have circulated online for months.

In many ways, this is influencer marketing pushed to its absolute extreme.

Traditional influencers sell products through aspiration. Sophie sells identity through contradiction. She presents herself as religious, reserved, morally selective, and unattainable while operating on one of the internet’s most sexualized platforms. That tension fuels endless conversation. Every debate over whether she is genuine or performing for attention only strengthens the brand further.

The modern creator economy no longer revolves solely around content itself.

Fans are paying for emotional investment, perceived intimacy, direct interaction, and participation in an ongoing narrative. Sophie appears to understand this better than most creators in the industry. The longer the mystery survives, the more profitable it becomes.

why is it so hard for people to believe im a virgin https://t.co/aCxvX1F6Oy

— Sophie Rain (@sophieraiin) May 19, 2026

That is also why her comments about ethics and morals resonated online.

Whether people believe her claims or not almost becomes secondary. Morality itself becomes part of the marketing structure. In an era where oversharing dominates internet culture, withholding something becomes powerful. The refusal becomes content. The boundary becomes content. Even the speculation becomes content.

Ironically, Sophie's strategy almost exists as the complete opposite of creators like Bonnie Blue and Lily Phillips, whose brands exploded through excess, shock value, and the constant escalation of explicit content. Where Bonnie and Lily built attention through pushing limits further and further in public, Sophie built hers through withholding.

One strategy thrives on overexposure. The other thrives on restraint.

Yet all three creators reveal the same reality about the modern internet economy. Attention is currency, and personal identity has become the product.

View this post on Instagram

A post shared by The Iced Coffee Hour Podcast (@icedcoffeehour)

Virginity, relationships, breakups, pregnancies, friendships, and even mental health struggles have increasingly become monetized storylines in influencer culture. The line between personal identity and commercial strategy has nearly disappeared. What once belonged to private life is now repackaged into engagement, virality, and profit for an audience constantly demanding something more intimate and emotionally revealing.

Reports surrounding Sophie's earnings have only amplified that phenomenon. She has claimed to earn close to a hundred million dollars through OnlyFans, while online discussions continue suggesting that individual subscribers have spent staggering amounts on her content alone.

Whether those numbers are fully accurate or partially exaggerated almost becomes irrelevant. The attention itself is the product.

What separates Sophie Rain from countless other creators is that she understands restraint can often be more profitable than unlimited access. In an online economy flooded with explicit content, suggestion can become more valuable than revelation. Mystery creates anticipation, anticipation creates obsession, and obsession keeps audiences invested far longer than instant gratification ever could.

That may ultimately be the real reason she allegedly turned down the $15 million offer. Not necessarily because the amount itself was unbelievable, but because ending the narrative could potentially destroy the very thing that made her so valuable in the first place.

View this post on Instagram

A post shared by The Iced Coffee Hour Podcast (@icedcoffeehour)

Whether Sophie was protecting her personal values, preserving her self respect, or simply safeguarding the long term power of her brand, she understands something many creators eventually learn too late.

In the attention economy, mystery is one of the most expensive assets a person can own.

View this post on Instagram

A post shared by The Iced Coffee Hour Podcast (@icedcoffeehour)

8-Year-Old Spots NASA Website Typo That Went Unnoticed For Over A Year

Danny Krisnanto — Tue, 19 May 2026 09:46:35 +0000

19/05/2026

NASA's website at nasa.gov is one of the most impressive and high-traffic government websites in the world.

It's extremely popular, and regularly pulls in tens of millions of visits per month (over 95 million in April 2026 alone). That's more traffic than many major news or tech sites. Since it's regularly updated too, it always has been a go-to for reliable space news, stunning images, mission updates, educational resources, and raw scientific data.

In other words, the website is considered one of the most trusted and authoritative sources for space-related information.

In the era where the choir of content writing has been elevated with the help of AI, NASA shows that behind the massive popularity of the website, there are humans who can still make mistakes.

And an 8-year-old boy was the one who realized this.

hey @NASA

my 8-year old just found a spelling mistake on your website.

he insists this is a "critical mission update" :) pic.twitter.com/QD4QQFm4sP

— musarat (@musaratali) May 18, 2026

The post came from X user @musaratali, who shared a screenshot from NASA's Solar System page showing the Jupiter section. A link shows "Expore Jupiter" instead of the correct "Explore Jupiter," with the missing "L" neatly circled in red. The parent noted that their child insisted this was a "critical mission update," and the post quickly gained huge engagement.

NASA responded with perfect humor and humility. Their official reply read: "Oops! Thanks for pointing it out to us. We're on our way to retrieve that missing letter."

The space agency fixed the typo within the hour.

Oops! Thanks for pointing it out to us. We're on our way to retrieve that missing letter.

— NASA (@NASA) May 18, 2026

The replies flooded in with jokes about Pluto's planetary status, budget concerns from the eight-year-old's discovery, and appreciation that NASA is still run by real humans rather than flawless AI.

This small incident highlights exactly why nasa.gov feels special among major government and science websites. It draws tens of millions of visitors each month for stunning images, live mission data, and reliable information, yet it still has room for a simple human error and the personality to laugh about it.

In an era of polished corporate content, NASA's quick, witty response reminded everyone that space exploration remains a human endeavor filled with curiosity, mistakes, and quick fixes.

The real star of the story is that sharp-eyed eight-year-old, who now knows that paying attention to details matters, even at NASA. Somewhere out there is a future astronaut who just completed his first official mission: retrieving a missing letter from Jupiter.

Moments like this make NASA’s site one of the most human and engaging major science platforms on the internet.

It's worth noting that the typo had been around for more than a year, and can be traced back to at leat 2024, when the page's design and content were last updated.

NASA corrected the typo in about an hour after the original post.

Further reading: One Tweet, One Reply, One Lesson: The NASA Internship That Went Viral On Twitter

With 'Scheduled Tasks 2.0,' Manus Shows Why Persistent Context And Reliable Scheduling Matter In Agentic AI

Aditya — Tue, 19 May 2026 06:47:49 +0000

The large language models (LLMs) war intensifies to a degree that mere chatbots are no longer useful.

Since the arrival of OpenAI's ChatGPT in late 2022, the AI's conversational breakthrough prompted an intense race among tech giants and startups alike. Companies invested billions to scale models that could not only generate coherent text but also reason through problems, summarize documents, write code, and support creative work.

Within months, competitors released their own versions, each promising greater speed, accuracy, or specialized capabilities.

But then, the focus expanded from simple chat interfaces to tools that could integrate with workflows, analyze data, and assist across professional domains.

And as LLMs matured, attention turned toward agentic systems capable of taking meaningful action rather than merely responding to prompts.

Manus has approached these challenges with a focus on practical continuity in agent behavior.

Read the blog here: https://t.co/fHAGM27XSq https://t.co/fHAGM27XSq

— Manus (@ManusAI) May 18, 2026

Now operating as part of Meta, the platform has introduced an upgrade to scheduled tasks, known as 'Scheduled Tasks 2.0,' which targets the gap between timed triggers and sustained intelligence.

Rather than running isolated instances on a clock, the system emphasizes execution in the appropriate environment while preserving necessary background information.

A recurring process can resume directly within an existing conversation thread, drawing on accumulated instructions, attached files, prior decisions, and evolving results. This design supports more reliable outcomes for workflows that build progressively over days or weeks.

The upgrade extends scheduling beyond standalone tasks into broader project structures and custom web applications built on the platform. In a project setting, scheduled work reuses shared elements such as uploaded documents, specialized skills, data connectors, and predefined output formats.

For web apps, it enables background operations that keep interfaces current without manual restarts. A dashboard might refresh metrics each morning, or a reporting tool could compile and distribute summaries on a weekly cadence, all handled internally by the agent.

Users establish these automations through straightforward natural language instructions.

Within a task, project, or app, one might direct the agent to review open items every weekday morning and highlight priorities, or to compile customer feedback each Monday by referencing existing files and maintaining consistent formatting. The agent incorporates connected data sources where relevant and adheres to any established standards for presentation.

For well-understood routines, confirmation steps can be bypassed, allowing fully automatic progression while still providing visibility into outcomes.

Visibility improvements help users monitor activity without disruption.

A dedicated side panel displays active schedules and their associated runs. Calendar-style and schedule views clarify upcoming executions alongside completed ones. Each run appears as a traceable card or label that links directly back to its originating context, enabling quick review of results or adjustments as needed.

Editing options allow refinement of prompts, timing, or execution details after initial setup, including choices about whether runs continue in the same thread or generate fresh instances.

What sets this approach apart is the deliberate handling of context across repeated executions.

Many scheduling systems spawn independent sessions, requiring the agent to reload everything each time and risking outdated assumptions. Manus instead maintains continuity where it adds value, keeping the agent aligned with the history and resources that shape the work.

This reduces the common friction of context loss while still permitting independent runs when isolation makes sense. It reflects a step toward agents that function more like persistent collaborators than episodic responders.

Additional controls enhance flexibility. Connectors integrate external applications as live data inputs. Execution environments can leverage cloud resources for demanding tasks.

Projects serve as centralized hubs, letting schedules inherit configurations without redundant setup. These elements combine to support diverse applications, from daily personal summaries to enterprise-scale reporting and app maintenance.

The feature is now accessible across all user levels in tasks, projects, and web apps.

By addressing continuity alongside timing, it advances agentic systems toward more dependable automation. As the broader competition among AI platforms continues, developments like this highlight the growing importance of execution reliability over isolated intelligence.

The result is a tool that integrates more seamlessly into ongoing work rather than sitting apart from it.

Published:

19/05/2026

X Introduces 'Active Followers' Metric To Show How Many Followers Are Actually Online

Nia — Tue, 19 May 2026 04:17:19 +0000

Follower counts, likes, reposts, and impressions have long been the currency of online creators.

But one of the biggest unknowns has always been how many people in an audience are actually there at any given moment. A creator may have hundreds of thousands of followers yet only a fraction may be actively scrolling when a post goes live leaving many frustrated when engagement falls short of expectations.

X is now trying to make that reality more transparent with a new 'Active Followers' metric.

Available inside Account Analytics, the feature shows both the number and percentage of followers who were active on the platform within the last 24 hours giving users a clearer understanding of their audience’s real-time presence rather than relying solely on total follower counts that may include inactive accounts.

The new addition appears as a dedicated analytics tile alongside existing insights such as verified followers and includes a line graph tracking activity trends over recent days.

Shipped.

Credit: @LingyunGao4, @ZachWarunek https://t.co/GQG2fiQIUd pic.twitter.com/4XB61emlla

— Nikita Bier (@nikitabier) May 19, 2026

The rollout was announced by Nikita Bier, head of product at X, who said the feature was built after strong interest from users and credited engineers Lingyun Gao and Zach Warunek for shipping the update.

The update stems from a poll posted by Bier.

In that earlier message he asked whether users would find value in knowing how many followers had been active in the previous day. The positive response led directly to the implementation which Bier announced with a simple note that it had shipped crediting engineers Lingyun Gao and Zach Warunek.

In a follow up he explained that many creators overestimate the share of their audience online at any given moment and feel frustrated when a post fails to reach everyone on their list.

The new data is intended to clarify that reality and set more realistic expectations for post performance.

Early user feedback reflected a range of reactions.

Some accounts shared their personal figures with percentages landing anywhere from the mid 50s to the high 80s. Others expressed surprise at lower than anticipated activity rates among their followers and suggested the platform could add tools to identify or remove inactive ones from following lists. A few called for an expansion to shorter time windows.

People overestimate how many of their followers are active at a given time and get frustrated when their post doesn’t reach 100% of their followers.

When we add 3-hour data, it will be obvious how much time it takes for a post to reach all active followers.

— Nikita Bier (@nikitabier) May 19, 2026

Bier responded by confirming that data for the past three hours is planned for a future release which would further illustrate how long it typically takes for content to circulate through the currently active portion of an audience.

Account Analytics has long offered insights into impressions replies and other engagement signals. The active followers metric builds on that foundation by focusing specifically on recent platform activity rather than cumulative totals.

For users who rely on the platform for content distribution the information could influence decisions around posting schedules or audience management though it does not alter how the recommendation system prioritizes content.

The rollout reflects an ongoing effort to refine the tools available in the analytics dashboard without changing core platform mechanics.

Published:

19/05/2026

How 'Euphoria' Season 3 Turned Sydney Sweeney Into The Center Of A Cultural Firestorm The Internet Wasn't Ready For

Bess — Mon, 18 May 2026 08:33:53 +0000

18/05/2026

Euphoria has always been a raw unflinching look at the messy realities of youth addiction sexuality and fractured relationships.

Centered on a group of East Highland High School students navigating love trauma and self-destruction with Zendaya's Rue Bennett as the emotional anchor whose voiceovers guide viewers through the chaos of teenage life in a small California town. Sydney Sweeney stars as Cassie Howard a character who starts out in the first season as the sweet vulnerable blonde desperate for male attention and validation often putting herself in uncomfortable situations that expose her deep insecurities and people-pleasing tendencies.

By season two, Cassie spirals further becoming one of the show's most polarizing figures after she secretly hooks up with her best friend Maddy's boyfriend Nate leading to a brutal public showdown in the school bathroom that cemented her as the ultimate messy girl willing to sacrifice everything including friendships for fleeting affection and a sense of being wanted.

Across those earlier seasons Cassie embodied the quiet desperation of someone who measures her worth through how desirable she appears to others her arcs blending moments of genuine heartbreak with self-sabotage that made audiences both root for her and cringe at her choices.

But nothing prepared viewers for the escalation in season three, which premiered on April 12, 2026, after a four-year hiatus.

Season three introduced a five-year time jump, thrusting the characters into early adulthood.

Cassie is now married to Nate and grappling with the financial strains of their new life together. To keep up appearances and cover mounting debts from Nate's shady business dealings she launches an OnlyFans account a plot that quickly became the season's most explosive and divisive element turning her insecurity into a full-blown public spectacle of explicit content creation managed in part with help from old acquaintances.

Putting aside the simulated sex scenes, some of Cassie's scenes are indeed unapologetically graphic.

They include Cassie filming material that includes puppy-play sequences where she wears a brown corset dog-ear headband painted nose and collar with a leash Nate tugging it while calling her a bad dog and her responding with playful barks in a moment that blends humiliation with dark humor.

Even more shocking is the baby-themed content in early episodes where she dresses in a pink onesie diaper pigtails and sucks on a pacifier with her legs spread wide during the filming process a fetishistic age-play setup that lingers on the camera in ways designed to provoke.

Other sequences show her in polka-dot bodysuits jumping rope with toys ice cream dripping suggestively down her body.

There is also a scene where she can be seen sucking her own toes while topless or posed with strategic props all framed as her desperate bid for income and attention amid marital tension.

And a scene where she creates an ASMR content.

These are not subtle nods but extended visually intense moments that push the boundaries of what the show has depicted before and they are entirely true to the aired episodes rather than exaggerated rumors.

What sent the internet into absolute meltdown was not just the nudity or kink elements but the perfect storm of cultural timing and execution around that baby scene in particular which first leaked in a pre-premiere trailer prompting immediate outrage over its infantilizing fetish vibes and perceived mockery of real sex work.

OnlyFans creators publicly slammed the portrayal noting that such age-play content including diapers and pacifiers outright violates the platform's acceptable use policies against anything implying underage scenarios yet the show depicted Cassie churning it out as easy cash.

Critics like former adult star Maitland Ward called it disgusting and harmful while Megyn Kelly dedicated segments to ripping it as vulgar and unnecessary accusing the series of reducing Sweeney to shock value.

Social media erupted with accusations of a humiliation ritual aimed at the actress herself who by 2026 had become a bona fide movie star known for her curves and mainstream sex appeal making the on-screen degradation feel even more pointed to some.

HBO even deleted and re-uploaded the trailer after backlash digitally altering the baby shot to cover visible nudity which only amplified the frenzy as people accused the network of knowing they had crossed a line yet airing the full version anyway.

What makes the backlash feel especially pointed is how the producers led by Sam Levinson appear to have leaned directly into the internet's long-standing perception (and obsession) of Sydney Sweeney as a hyper-sexualized bombshell whose body and image are endlessly dissected thirsted over and criticized in equal measure.

Levinson has openly framed the show as a heightened reflection of teenage identity shaped by internet culture and in interviews he has defended the Cassie arc as metatextual commentary where the character's emotional vulnerability hyper-visibility and public scrutiny mirror exactly how audiences consume Sweeney herself in real life from her blockbuster roles and ad campaigns to the constant online fixation on her figure.

He has praised Sweeney for delivering brilliant chaotic performances when pushed into these extremes noting early conversations where she asked him to simply make Cassie crazy and the result feels uncannily like the show is winking at the very discourse that already surrounds its star turning her real-world sex-symbol status into fuel for the storyline's absurdity and self-destruction.

At its heart, the chaos stems from how season three weaponizes Cassie''s long-established fragility turning her need for validation into something cartoonishly self-destructive, while leveraging Sweeney's real-world image for maximum impact.

Fans split sharply between those praising it as bold commentary on modern influencer culture desperation and female agency in the face of financial ruin and those decrying it as tone-deaf exploitative trauma porn that prioritizes titillation over substance.

Clips of the puppy and baby moments flooded TikTok X and Reddit racking up millions of views in thirst traps moral panic rants and endless memes all while the show continued dropping new episodes weekly through late May.

Read: Sydney Sweeney Versus Internet's Body Shamers, And Getting The Last Laugh

By the time episode six hit screens on May 17 the discourse had already reached fever pitch with fresh waves of reactions to Cassie's escalating meltdowns and the looming fallout with Maddy ensuring the OnlyFans arc remained the inescapable talking point.

In the end Euphoria season three did what it always does best forcing uncomfortable conversations about desire shame and the cost of being seen even if this time the internet did not just watch but completely lost its collective mind over how far it was willing to go with Sydney Sweeney's Cassie.

It's worth noting that almost every female lead characters has been fetishized, either through ASMR, giantess, feet and breasts, systematic hyper sexualization and objectification through being a stripper, in sex clubs, being a sugar baby, OnlyFans creator, humiliation or degradation, pet play, and more.

Further reading: Sydney Sweeney Goes Viral For A Hotdog And A Seethrough Dress, And People Keep Seeing Her In Their Feeds

'Runway Agent' Is Turning Conversations Into Finished Films With End-To-End AI Video Creation

Aditya — Sun, 17 May 2026 15:03:50 +0000

The large language models (LLMs) war keeps on intensifying, and where things are heading it's clear.

Since the arrival of OpenAI's ChatGPT that demonstrated how conversational interfaces could reshape everyday tasks, tech companies race towards creating increasingly powerful models. Fast forward, the race quickly expanded into image and video generation, sparking competition across the industry. In that crowded space, Runway carved out a distinct position.

Not by sheer power, but by focusing exclusively on video tools from the start.

It stood out for developing models that turned text prompts and reference images into fluid video sequences with practical controls for timing, style, and consistency, features aimed at filmmakers and content creators who needed reliable outputs rather than experimental clips.Runway built on that foundation through successive generations of its technology.

And now, Runway releases 'Runway Agent' with the purpose of extending its approach into a more integrated system.

Meet Runway Agent. Your new AI creative partner that helps you ideate and execute fully finished, sound designed and edited videos. All with just a simple conversation. From ads to shorts to content for social, Runway Agent makes it easy to make more of what you need.

Get… pic.twitter.com/lFuN6uT6Oe

— Runway (@runwayml) May 13, 2026

The tool operates as a conversational partner that handles the full process from initial idea to finished video.

By bringing together existing tools, like Runway AI's ability to add custom voices and multi-shot and story panels, the company is positioning ‘Runway Agent’ as a more complete end-to-end AI video production system.

A user begins by describing the desired content in plain language, optionally uploading reference images for characters, products, or environments. They can also specify details such as aspect ratio, duration, and audio preferences. The agent then proposes a concept, breaks it into story beats, generates visual references, and suggests tones for dialogue and music.

Refinements happen through ongoing chat, after which the system assembles a multi-shot video complete with voiceover, sound design, and editing.

In other words, users can engage in a back and forth dialogue with the AI to refine the outline, adjust elements, or clarify creative choices until the plan aligns with their vision.

Once approved, the agent assembles a complete multi shot video that incorporates generated scenes, voiceover or dialogue where appropriate, background music, and sound design.

The process handles ideation through to final editing within the same interface, producing a polished output suitable for social media, brand campaigns, short films, or product demonstrations.

A built in timeline editor remains available afterward for any last minute changes to pacing, transitions, or specific frames.

In one demonstration, the agent produced a marketing clip in which a woman in a kitchen discusses the layered flavors of a chocolate bar, complete with synchronized dialogue and ambient audio.

The result compresses what once required separate tools for scripting, generation, and assembly into a single exchange. Users have noted that it streamlines ideation for brand campaigns, social shorts, or film previsualization, though some point out that final outputs may still benefit from manual polishing for complex projects.

This integration of conversational guidance with end-to-end video assembly forms part of Runway's approach to video production.

With the ability to combinie scripting, generation, editing, and sound design into one dialogue-based process, the tool should be able to simplify the ideation and production for routine tasks like social media content or brand clips. However, outputs for more complex projects often still need adjustments through the built-in timeline editor to refine pacing, transitions, or specific frames.

The feature is available now via the Runway web app.

In the expanding field of AI video tools, other platforms have introduced comparable agent-style functions.

These systems incorporate chat interfaces into more agentic workflows to move from initial prompts to structured videos. Each of them offer varying levels of automated storyboarding, voiceover, and editing support. As development continues across these tools, the competition around streamlined visual production workflows remains active.

Published:

13/05/2026

The Attempt To Kill 'Motherless' Once And For All: The Adult Website That Refused To Die

Bess — Sun, 17 May 2026 10:20:21 +0000

15/05/2026

When the attempt to shut down Motherless, the news sent ripples through online communities and beyond.

The move came after months of mounting international pressure triggered by a detailed investigative report from the network that exposed the platform as a hub for deeply disturbing content linked to drug-facilitated sexual assault and gender-based violence.

For years Motherless had operated with a hands-off philosophy, billing itself as a moral-free file host where anything legal could be uploaded and stored indefinitely, but major publications' probe painted a far darker picture of how its lax moderation enabled real-world harm.

The investigation itself pulled back the curtain on an ecosystem that extended well beyond a single website.

This however, is short-lived.

Motherless was put to sleep, only to be awakened about a week later.

Motherless.com is a long-running adult website that functions as a free, user-generated pornographic video and image-sharing platform, often described by its own operators as a "moral free file host where anything legal is hosted forever."

Launched in 2008 by Joshua Lange, who remains its sole owner and operator, the site was built on the model of early file-hosting services but specialized exclusively in explicit content. It started by importing hundreds of thousands of files from Lange's previous site, Hidebehind.com, and quickly grew into one of the internet's largest repositories of amateur pornography through crowdsourced uploads.

Unlike mainstream tube sites that partner with professional studios or pay for licensed material, Motherless relies almost entirely on everyday users uploading their own videos, photos, and clips.

In other words, there are no production deals, no payouts to creators beyond occasional credit systems that users can sometimes convert into small cash rewards or merchandise.

The homepage and navigation reflect a raw, unfiltered approach to adult material with categories and tags sprawl across hundreds of options, covering the full spectrum of human sexuality and then some.

This is why Motherless is considered the safe haven for truly unfiltered, often amateur and homemade material that other platforms might flag or remove more aggressively. And an investigation from CNN documented more than 20,000 videos tagged with terms like #passedout, #eyecheck, or #sleep.

Many of these clips showed women who appeared unconscious: eyes rolled back, bodies limp: while men performed sexual acts on them, often lifting eyelids to confirm unresponsiveness or filming in ways that suggested drugging or heavy intoxication.

These videos frequently garnered hundreds of thousands of views, and the reporting linked them to private Telegram groups where participants exchanged detailed guides on drug-facilitated sexual assault: mixing sedatives into drinks at low starting doses, combining them with alcohol, using anti-diarrheal medications like Imodium to counteract side effects, and even selling "sleeping liquids" or charging cryptocurrency for live streams of assaults.

These stories underscored a broader pattern: drug-facilitated sexual assaults are chronically underreported due to memory gaps, embarrassment, and institutional biases, with conviction rates remaining low across Europe and the United States despite laws that technically cover such offenses.

Read: Husband Hired Over 50 Men Online To Rape His Wife Nearly 100 Times Over 10 Years

The porn website Motherless, which has faced international scrutiny over hosting content linked to gender-based violence and drug-facilitated sexual assault, has been taken offline by Dutch authorities following mounting pressure in the wake of a CNN investigation.… pic.twitter.com/RjEIPsEYJP

— CNN (@CNN) May 8, 2026

The legal landscape surrounding these platforms added another layer of complexity.

Motherless relied on jurisdictional loopholes, with servers in the Netherlands, a domain possibly registered elsewhere, and a parent company based in Costa Rica, all strategies common among adult sites seeking to minimize oversight. U.S. safe harbor provisions under laws like Section 230 had historically shielded platforms from liability for user-uploaded content, but European regulators were increasingly unwilling to tolerate the spillover effects.

The Dutch Public Prosecution Service acted decisively once evidence from the CNN collaboration with outlets like NOS and Nieuwsuur reached them, especially after the hosting provider NFOrce issued an urgent compliance review and gave the site a short window to respond.

The site was taken down, to only respawn about a week later.

In a statement issued shortly after the initial takedown, Motherless claimed the offline period was voluntary.

A dedicated page was created to create transparency about its auditing log.

By mid-May the site had resurfaced a 100%, as if nothing happened.

A prominent welcome message greets visitors with "We’re happy to be back and truly appreciate everyone who has been waiting." However, there are a few changes that followed.

First off, Motherless said that to comply with the law, moderation has been strengthened to better detect and remove illegal uploads, complete with a public audit log that tracks every report and its resolution so the community can monitor enforcement.

The company also promised to strengthen systems, ban repeat offenders, and review keywords before returning. While this is an improvement, it's also considered by many as a familiar cycle of cleanup that had followed previous scrutiny but never quite stuck.

The case stands as a reminder that online anonymity can shield both creators of content and the predators who exploit it, leaving society to weigh the costs of unrestricted access against the very real scars left on those caught in the crossfire. The attempt to kill Motherless once and for all had failed. Motherless refused to die.

'Maybe You're A Dog. I Don't Care': How Introverts Are Quietly Running The Modern Internet

Nia — Sun, 17 May 2026 09:38:12 +0000

Jean-Baptiste Kempf

computer engineer, entrepreneur, major contributor to the VideoLAN project and VLC software

FFmpeg and VLC form the invisible backbone of modern internet civilization, quietly driving the foundational infrastructure behind global platforms like YouTube, Netflix, Chrome, Discord, and billions of everyday consumer devices.

In a sweeping, multi-hour conversation on the Lex Fridman Podcast, developers Jean-Baptiste Kempf and Kieran Kunhya peeled back the curtain on this hidden digital machinery, revealing a world where the entire global video landscape relies almost entirely on the obsessive craftsmanship, intellectual focus, and quiet generosity of independent open-source volunteers.

They literally emphasized how introverts are literally running the world with their techs, despite not seen, and rarely under the spotlight.

The dialogue serves as a rare, unfiltered masterclass in how modern infrastructure is actually maintained away from the corporate spotlight, offering a stark contrast to the highly polished narratives pushed by contemporary tech conglomerates.

At the absolute heart of their engineering philosophy lies a radical, unyielding adherence to a classic, foundational internet ideal, one that is crystallized perfectly by the famous 1990s cartoon.

Jean-Baptiste Kempf is one of those people who leads the development of one of the most powerful tools on the web, ever.

"Most of us are introverts, right? You need to be more precise. You have extremely introverts, extremely, extremely introverts and introverts, right? It's just like a whole spectrum of different people. It doesn't matter. The important thing is, is your code good?"

"Is your code great? Is your technology great? We care about excellent code. We don't care who you are. Sorry, it's just like we have no idea to check. We cannot check, right? Like, maybe you're a dog. I don't care, right?"

"I don't care where you come from. I need to look at your code. And this is important because people don't understand that, and they come to the community and send some patches, and they get rejected, and they don't like that because, I mean, you're just like, “Sorry, it's not up to our standards."

"Oh yeah, but I'm an engineer at this very large company in Italy, in Germany, in the U.S.. We don't care."

"We care about the quality of your code because this is what defines our community, which means that we have a lot of people who contribute from very different backgrounds and are very introverted, sure. But that's okay, right?"

When the developers emphatically stated that they do not care who you are, where you come from, or if you are even a dog, they highlighted the fierce, uncompromising meritocracy that defines their development community.

Here, Jean-Baptiste Kempf is referencing a famous 1993 cartoon by Peter Steiner published in The New Yorker. The cartoon shows two dogs sitting in front of a computer, with one telling the other, "On the Internet, nobody knows you’re a dog." The line became one of the defining cultural symbols of the early internet era because it captured the belief that online spaces could act as a kind of equalizer.

In other words, one's appearance, age, race, nationality, social status, or background mattered far less than the ideas and information they shared online.

This literally means that in Jean-Baptiste Kempf's view, identity is secondary to contribution.

What matters is whether the code is good, whether the patch improves the project, and whether the technology meets the community’s standards. The statement is not meant literally. It is a way of saying that personal status or credentials should not influence technical judgment.

This mindset reflects the deeply meritocratic culture that shaped many major open source communities, including projects like FFmpeg and VLC media player.

In these communities, code quality matters more than reputation. A contributor with a prestigious degree or a job at a major technology company is judged by the same standards as a completely unknown programmer submitting their first patch. If the code is efficient, maintainable, and technically strong, it earns respect. If it fails to meet the project’s standards, it gets rejected regardless of who wrote it.

That culture also explains why many influential contributors in open source projects remain largely anonymous or use pseudonyms online.

Maintainers may never know what these contributors look like, where they live, or what their real names are. What defines a person inside the community is the quality of their work and the reliability of their contributions. In that sense, contributors become known less as personalities and more as trusted technical voices whose reputation is built entirely through the code they produce.

Since its first inception, FFmpeg has proven itself as one of the most useful tools out there. And now, it is easily one of the single largest consumers of CPU power in human history, featuring over one hundred thousand lines of highly specialized, handwritten assembly code where every single CPU cycle is optimized to exploit the specific flaws and limitations of human biological visual perception.

In a particularly gripping moment, they discussed the immense pressure of digital security and geopolitics, detailing how they flatly and impolitely rejected multiple backdoor requests from state intelligence agencies, cementing a fierce ethical stance that they would rather shut down the entire multi-billion-download project completely than compromise the privacy and trust of their global users.

Lex Fridman had to honor his conversation with Jean-Baptiste Kempf and Kieran Kunhya by wearing the VLC traffic cone hat the whole time.

Jean-Baptiste Kempf even said that he refused multiple offers of several million dollars for FFmpeg.

"I refuse dozens of millions of dollars, yes, several times. Yes, I could be a multimillionaire and be somewhere on the beach. But I did not do it because I thought it was not moral and it was not the right thing to do. And this is very important for myself, is to be like, I work for the greater good, I work for people, and I don't want. It's not just by myself," he said.

The podcast ultimately painted a profound, humbling portrait of a unique engineering subculture driven by creative passion, internal drive, and technical curiosity rather than public praise or financial profit.

It serves as a stark, powerful reminder that the tools people use to stream video and audio every single day are not merely corporate products, but the result of quiet, cross-border human collaboration (most of whom are introverts) operating completely outside the boundaries of traditional capitalism.

The developers also shared an intriguing psychological perspective on the nature of their long-term work, viewing regret not just as a standard moral feeling but as a literal cognitive tax or an attack on the mind that drains intellectual resources needed for future innovation.

By prioritizing the purity of the craft over individual human egos and emotional baggage, these volunteer engineers continue to build something incredibly useful, durable, and elegant for the rest of humanity, reminding the rest of the world of the classic closing wisdom from Linus Torvalds that most good programmers do what they do not for public adulation, but simply because it is fun.

A 10-Year-Old Girl's Testimony Sparks Internet Outrage: The Arrest Of Her Abusive Father

Bess — Sun, 17 May 2026 06:51:48 +0000

17/05/2026

A father is a man who should devote himself for his family. He is the first person in the family who is supposed to protect his family from every harm, and be that unbreakable guardian who asks nothing in return.

But a man in Natar in Lampung Selatan Regency has been accused of betraying that sacred role in the cruelest way possible.

Identified only by his initials NP and aged 48, he now stands as the prime suspect in the repeated sexual abuse of his own biological daughter, a 10 year old girl.

But unlike most similar cases, this one is unique because the victim has given her own remarkably detailed testimony.

She is able to recount the incidents in great detail, with specific dates and circumstances that she remembered clearly even amid her tears and profound trauma.

View this post on Instagram

A post shared by Lampung City Info (@lampungcityinfo)

The horrors allegedly began in mid January 2025 and continued for more than a year with the most recent incident occurring on the night of May 11, 2026 around 00:10 local time when the house was empty and the child was left alone with him.

Her mother had been working as a migrant laborer in Taiwan leaving the young girl under the sole care of her father at their family home in Natar.

According to the girl's own heartbreaking account given to investigators and her legal team, the abuse unfolded repeatedly when the two were alone in the house. The father would force her into the bedroom and that he would take off his pants. He then allegedly made her touch his genitals while instructing her exactly how to do it saying things like "do it like this."

While no penetration has ever occured, according to the girl's statement, these obscene acts include rubbing, fondling and squeezing.

The acts allegedly occurred at least five times inside their own home with the father exploiting every moment the house was quiet and no one else was around to witness or intervene.

When asked whether the girl knows anything about these kind of acts, like of her siblings taught her, or whether she heard from someone else, the girl said that she learned about them from no one, and feel violated, even though she did not fully understand what was happening.

The girl lived under constant threats and pressure from the man who was supposed to protect her never daring to speak until the weight of the trauma became unbearable.

Family members also suspect she endured repeated physical violence in the same household adding layers of fear to the sexual violations.

View this post on Instagram

A post shared by Paman Acong (@pamanacong_89)

When she finally gathered the courage to confide in her mother and relatives the truth spilled out in heartbreaking detail. Instead of returning to her fathers house the child sought refuge with relatives in Panjang Bandar Lampung.

On May 13, 2026 the family filed an official report at the Lampung Regional Police headquarters under case number LP/B/358/V/2026/SPKT/POLDA LAMPUNG.

The girl's testimony captured in emotional videos that quickly spread online showed her breaking into uncontrollable tears as she recounted the incidents.

She kept repeating how terrified she felt and how she never wanted to see her father again.

According to her account, she even wished him to be thrown in jail, beaten, and rot.

The entire account was then handed over to police as crucial evidence in the investigation.

View this post on Instagram

A post shared by Kabar Terdepan Mojokerto (@kabarterdepanmojokerto)

Investigators from the Directorate of General Criminal Investigation at Polda Lampung responded with speed and seriousness.

After collecting consistent statements from the victim conducted psychological evaluations and secured a medical visum et repertum that confirmed physical trauma in sensitive areas, an internal case review police officially named NP as a suspect issued an investigation warrant and took him into custody.

The case file was promptly sent to the Lampung High Prosecutors Office.

The demon now faces charges under the Indonesian Criminal Code and child protection laws which could result in up to 15 years in prison with the possibility of an additional one third increase because he abused his position as the victims own parent.

The Hotman Paris 911 legal assistance team played a crucial role from the start.

Local lawyers including Paman Acong and Putri Maya Rumayanti accompanied the family every step of the way providing full support and ensuring the childs safety.

They even facilitated the formal handover of the suspect to police.

During that moment the girl reportedly cried hysterically at the mere sight of her father underscoring the depth of her trauma.

Her mother who learned the devastating news while still in Taiwan broke down in tears of grief and relief thanking the legal team for standing by her daughter when she could not be there herself.

View this post on Instagram

A post shared by Kabar Terdepan Mojokerto (@kabarterdepanmojokerto)

Videos of the girls tearful interview and scenes of authorities confronting the suspect have fueled widespread outrage across social media and news outlets.

Indonesians from every corner have voiced their horror demanding the strictest punishment possible to serve as a powerful deterrent. Many have called for chemical castration or even the death penalty insisting that no one especially a parent should escape full accountability for harming a child. As the legal process advances with the suspect now detained the focus remains firmly on the young survivor.

She is receiving ongoing psychological support to begin healing from the profound betrayal she has endured.

This case exposes painful truths about families separated by economic necessity where parents migrate overseas for work and children are left vulnerable in the care of a single guardian.

Child sexual abuse often stays hidden behind closed doors where trust is supposed to be strongest and it remains tragically underreported despite growing awareness. Yet in the middle of this darkness the courage of one small girl who spoke out in such vivid and precise detail has sparked important conversations about vigilance family responsibility and the urgent need for stronger safeguards.

Communities authorities and society as a whole must stay alert so that no child ever faces such isolation and suffering again.

'Grok V9' Training Advances At xAI While SuperGrok Users Report Tight Generation Caps

Danny Krisnanto — Sun, 17 May 2026 02:50:19 +0000

xAI has completed the initial training run for its Grok V9 foundation model, which incorporates 1.5 trillion parameters.

This represents a threefold increase from the 0.5 trillion parameters that underpin the current Grok 4.3 version, itself based on the earlier V8 architecture. Early internal evaluations indicate that the new model has delivered strong results even prior to the next stages of development. The training incorporated enhanced data curation techniques, an updated training recipe, and optimizations tailored for Nvidia Blackwell GPUs.

Its plan is to conduct supplemental training that will integrate additional data from Cursor, the AI code editor. This step will precede supervised fine-tuning and reinforcement learning processes.

A public release of Grok V9 is expected to happen soon, and during this interval, the existing 0.5 trillion parameter model continues to receive incremental improvements on a frequent basis.

Statements from xAI leadership have described the upcoming version as a substantial advancement in scale and overall performance relative to its predecessor.

The version numbers are a little confusing and deserve some explanation.

Internally, we are working on version 9 of our new foundation model, which is 1.5T params. This is substantially better in every way than v8: data curation, training recipe, size, etc. It is also optimized…

— Elon Musk (@elonmusk) May 15, 2026

The update arrives amid active development of even larger architectures within xAI.

Multiple models are currently in training on the Colossus infrastructure, including variants aimed at 6 trillion and 10 trillion parameters. These efforts reflect a sustained focus on expanding model capacity while refining training methodologies across several concurrent projects.

The partnership with Cursor also provides mutual access to computing resources, with xAI supplying GPU infrastructure to support Cursor model training and exploring potential deeper collaboration that could include data sharing for Grok tuning.

We are improving the 0.5T Grok foundation model V8 (public version 4.3) every few days.

The 1.5T V9 just finished training (incorrectly called pre-training) and is a major upgrade. Next, we are adding the Cursor data in supplemental training (others call this mid-training),…

— Elon Musk (@elonmusk) May 17, 2026

At the same time xAI has rolled out Grok Build, an early beta coding agent available exclusively to subscribers on the SuperGrok Heavy tier.

The tool functions as a command line interface that supports planning, review, and execution of complex software engineering tasks.

It incorporates features such as parallel sub-agents for research and evaluation, along with an automated mode that ranks competing solutions before human review.

Early users have noted its responsive terminal interface and the speed with which the team has addressed initial feedback through rapid updates. Feedback has emphasized its utility for professional coding workflows, though availability remains restricted to the highest subscription level at present.

Elon on Grok v9.

Foundation model v9, 1.5T params, training run just finished. Cursor data still pending as supplemental.

Versus v8: 3x larger, rebuilt data curation, new training recipe, Blackwell native, v8 was Hopper.

Public Grok 4.2 is still v8 underneath. Elon says the v8… pic.twitter.com/0jlfhKqsYr

— tetsuo (@tetsuoai) May 16, 2026

While the public is eager for this Grok V9 and what it may introduce, particularly in coding and general reasoning tasks, developments have coincided with ongoing reports of usage constraints affecting many SuperGrok subscribers.

Individuals on tiers starting at $30 per month and extending upward have described reaching daily or session limits after generating only a small number of images or video clips, in some cases as few as 10 to 20 images or three videos. Voice interaction features have similarly been reported to cap out after 20 to 30 minutes of use for certain accounts.

These limits apply across both standard and higher paid plans and have prompted discussions about the gap between expected availability and actual experience.

Users have pointed to the lack of clear quota visibility in the interface as an additional source of inconvenience during active projects.

Inside the 100K GPU xAI Colossus Cluster that Supermicro helped build for Elon Musk to power xAI's Grok.

xAI has attributed the restrictions to rapid growth in overall demand and the resources required to support model training alongside user services. The company has stated that capacity expansions are in progress to ease these pressures, noting that similar periods of adjustment have occurred during earlier phases of expansion.

Some observers have characterized the situation as temporary growing pains associated with scaling large scale AI systems.

Others have expressed disappointment that paid subscribers continue to encounter such constraints on creative and productivity tools that were highlighted in earlier communications.

The current phase illustrates the challenges involved in advancing frontier AI models while managing the expectations and experiences of a growing user base.

Published:

17/05/2026

xAI Unveils 'Grok Build': An AI-Powered Terminal Built For Modern Developers Who Love Code, And Using A Mouse

Aditya — Sat, 16 May 2026 12:54:59 +0000

Developers and programmers have long favored the command line interface.

It delivers direct control through text based commands and allows users to maintain concentration without the interruptions common in graphical environments. Many developers report greater fluency with typing instructions than with mouse driven menus or visual layouts. The terminal has served as a reliable workspace for decades where scripts run efficiently and outputs appear in predictable formats.

This preference persists even as software projects grow more complex and involve larger codebases or multiple collaborating components.Recent advancements in artificial intelligence have introduced agents capable of handling coding tasks research and automation.

These agents often operate within chat interfaces or dedicated applications that can require frequent context switching.

Developers working with several agents simultaneously sometimes encounter visual disruptions such as screen redraws or layout shifts that interrupt focus. Tools designed for terminal use have begun to address these challenges by incorporating features that respect established command line workflows while adding capabilities for agent coordination.

xAI announced an early beta of 'Grok Build,' a tool described as an agentic command line interface intended for professional software engineering complex coding projects and workflow automation.

Try this early Grok Build (anything) beta and let us know what to improve.

Much appreciated! https://t.co/1GgGEQzv3E

— Elon Musk (@elonmusk) May 14, 2026

It operates directly within the terminal and is currently available only to SuperGrok Heavy subscribers.

The beta phase allows xAI to gather feedback and refine both the underlying models and the product based on real usage.Installation follows a straightforward process. Users run a single command curl -fsSL https://x.ai/cli/install.sh | bash to add the binary to the system.

After installation developers navigate to a project directory and launch an interactive session by typing grok.

The first launch typically opens a browser window for authentication using the SuperGrok Heavy account.

In environments without browser access users can set an environment variable with an API key obtained from the xAI console.

Syntax highlighting,, plan mode, clickable interface in Grok Build https://t.co/thwVY9auFq pic.twitter.com/gYQKJd0jAm

— Andrew Milich (@milichab) May 14, 2026

The tool then connects to the project files and prepares for natural language prompts.The interface takes the form of a terminal user interface that supports mouse interactions.

Users can click to navigate between plans subagents and parallel tasks.

The design maintains stable panes that avoid flickering or unexpected redraws during operation. This stability becomes relevant when coordinating multiple agents at once as developers can switch between reviewing implementation plans inspecting subagent outputs or selecting skills without visual interruptions. Keyboard shortcuts remain available including enter to send prompts shift tab to return to normal mode control h to return to the home view and control q to quit the session.

Grok Build coordinates subagents that handle distinct aspects of a workflow.

These subagents can run in parallel to perform tasks such as researching codebases proposing code changes executing shell commands managing dependencies or generating diffs for review. Before any modifications occur the system presents an implementation plan that users can examine and adjust.

this is an early beta. we still have much to improve, which is why its so exciting for people to start trying.

the goal is for Grok Build to be the best coding agent when it comes to solving real and difficult engineering problems. https://t.co/DHGzLSOMDN

— Jason Ginsberg (@JasonBud) May 14, 2026

The tool draws on Grok models to interpret natural language instructions and interact directly with local project files.

It supports validation of code edits to specific files and can ask clarifying questions to refine details during the process.Extensibility forms another aspect of the design.

Users can define custom skills plugins or extensions to adapt the tool to specific preferences or team requirements. A marketplace feature enables sharing of capabilities across projects or teams. The system also includes support for headless operation which allows integration into scripts bots or larger automation pipelines without relying on the interactive view.

Grok Build https://t.co/uWixUwweB3

— Elon Musk (@elonmusk) May 14, 2026

Through the Agent Client Protocol developers can build custom agents or incorporate Grok Build into other applications.

The early beta status indicates that xAI expects ongoing adjustments based on input from subscribers.

Within the CLI users can submit feedback directly using a designated command. Documentation outlines basic usage scenarios such as explaining a repository structure or guiding the agent through stepwise project development.

The release provides terminal based support for agent orchestration in coding and automation tasks. It reflects an approach that combines traditional command line efficiency with structured support for parallel agent work and interactive navigation.

Published:

16/05/2026

The Rich, The Powerful, And The Lonely Elon Musk: Viral During Beijing Banquet

Linda Sutisna — Sat, 16 May 2026 09:56:07 +0000

15/05/2026

In the glittering halls of Beijing's Great Hall of the People in China, a single one-minute video posted by Elon Musk managed to capture a moment that feels equal parts diplomatic theater and strangely intimate human drama.

The clip shows the Tesla and SpaceX chief casually chatting, smiling that familiar half-grin, glancing at his phone, taking selfies with others, and reacting with unfiltered charm while seated among presidents, premiers, and fellow tech titans during the state banquet honoring President Donald Trump's visit with Chinese President Xi Jinping.

It is a rare sight indeed: seeing so many tech leaders (who are also billionaires) gathered under one roof.

Most notably, these highly-influential people were representing the sharpest edges of American and Chinese innovation in the same formal setting, bridging two economies that have spent years locked in tense rivalry.

pic.twitter.com/QMgqquoGkS

— Elon Musk (@elonmusk) May 15, 2026

The full official White House list of U.S. business executives who traveled with President Trump and participated in the Beijing events (including the May 14, 2026 state banquet at the Great Hall of the People) includes the following.

Elon Musk: CEO of Tesla and SpaceX (tech/AI/electric vehicles/space).
Tim Cook: CEO of Apple (tech/consumer electronics).
Jensen Huang: CEO of Nvidia (tech/semiconductors/AI).
Cristiano Amon: CEO of Qualcomm (tech/semiconductors).
Sanjay Mehrotra: CEO of Micron Technology (tech/semiconductors/memory).
Chuck Robbins: CEO of Cisco Systems (tech/networking).
Dina Powell: President and Vice Chair of Meta (tech/social media/AI; executive-level leader).
Jacob Thaysen: CEO of Illumina (tech/genomics/biotech sequencing).
Jim Anderson: CEO of Coherent (tech/optics/lasers for semiconductors and industrial tech).
Larry Culp: CEO of GE Aerospace (advanced manufacturing/tech in aviation).
Kelly Ortberg: CEO of Boeing (aerospace/tech).
Ryan McInerney: CEO of Visa (fintech/payments tech).

Unlike the other CEOs, Musk and Huang directly travelled with Trump on Air Force One.

Non-tech/finance/agriculture members of the full delegation (also present) included Larry Fink (BlackRock), Stephen Schwarzman (Blackstone), Jane Fraser (Citigroup), David Solomon (Goldman Sachs), Brian Sikes (Cargill), and Michael Miebach (Mastercard).

These 18 executives formed the core high-powered business group; Trump personally introduced many of them to Xi Jinping.

Chinese executives who attended the state dinner and mingled/rubbed shoulders with the U.S. delegation (per on-site reporting and CCTV footage) were not always direct one-to-one matches but represented key sectors like tech, electronics, autos, and manufacturing.

The confirmed attendees included:

Yang Yuanqing: Chairman and CEO of Lenovo Group (tech/PC and devices; direct counterpart to U.S. hardware leaders like Apple and Dell).
Lei Jun: Founder and Chairman of Xiaomi (tech/smartphones, EVs, consumer electronics).
Liang Rubo: CEO of ByteDance (tech/social media/AI; TikTok parent).
Jia Shaoqian: Chairman of Hisense Group (electronics/appliances/displays).
Zhou Qunfei: Chairwoman of Lens Technology (tech/display glass for screens).
Lu Weiding: Chairman of Wanxiang Group (auto parts/EV components).
Cao Hu: Chairman of Fuyao Glass Industry Group (auto glass manufacturing).

Additional aviation/industrial leaders include Liu Tiexiang (China National Aviation Holding) and He Dongfeng (Commercial Aircraft Corporation of China).

These Chinese leaders sat near or interacted with U.S. counterparts at the banquet tables, creating the rare mixed U.S.-China tech billionaire scene captured in Musk’s viral video.

No single exhaustive "pairing" list exists, and that seating was diplomatic rather than industry-matched.

These powerful business figures from the U.S. did not travel halfway around the world simply for ceremonial handshakes, or business matters, or photo opportunities with their Chinese counterparts.

After all, the event was more about Trump-Xi.

However, the groups above represent the full set of named executives, attending a gathering considered one of the largest concentrations of U.S. and Chinese tech power under one roof in recent years.

Their purpose in China is crystal clear and relentlessly practical: to secure deeper footholds in the world’s second-largest market at a pivotal moment.

For the U.S. CEOs, the banquet and surrounding talks represent a chance to press for expanded market access, smoother regulatory approvals for advanced AI chips and electric vehicles, stabilized supply chains, and fresh investment pipelines that could protect and grow their massive stakes in everything from consumer electronics to autonomous driving technology.

Chinese leaders, in response, have emphasized that their doors will open even wider to American businesses, framing the gathering as a mutual win that deepens economic ties while signaling stability amid global uncertainties.

President Trump's decision to bring this high-powered delegation along on the state visit carries its own calculated purpose.

Surrounded by executives whose companies represent trillions in combined value, Trump is wielding corporate muscle as a tool of diplomacy, pushing Beijing to deliver concrete commitments on trade openness, reciprocal investments that could bring jobs and capital back to the United States, and resolutions to long-standing frictions over technology transfers and tariffs.

By personally introducing the CEOs to Xi and spotlighting their presence at every key event, the president aims to score tangible economic victories, project strength to his domestic audience ahead of midterm elections, and reframe U.S.-China relations as a partnership of mutual prosperity rather than pure competition.

It is business-first diplomacy in its purest form, where the presence of the world's richest and most influential executives becomes leverage for national interests.

But behind the glamour, all that orchestrated grandeur, and trillion-dollar deal-making, Musk's viral clip peels back the curtain to something far more relatable.

Musk, the richest man in the world, whose fortune exceeds that of many attendees combined, sat among heads of state at the banquet table.

Yet in a quiet moment, phone in hand and seemingly lost in thought, he appeared almost solitary: a reminder that immense wealth and influence do not necessarily shield one from the quiet isolation that can accompany extraordinary success.

The contrast is magnetic: the rich and the powerful assembled in one of the most exclusive rooms on Earth, yet the video's charm lies in how it humanizes the loneliness that can linger even at the center of history.

In the end, what started as a state banquet has become something far more memorable—a fleeting glimpse into the very human side of those shaping our technological and geopolitical future.

I am the WiFi

— Elon Musk (@elonmusk) May 15, 2026

Inside Indonesia's 'Chromebookgate' Scandal: Google, Politics, Hundreds Of Millions Of Dollars Of Alleged Corruption, And Nadiem Makarim

Danny Krisnanto — Sat, 16 May 2026 08:25:23 +0000

16/05/2026

Indonesia, the world's largest archipelagic nation, is a home to over 270 million people and hundreds of ethnic groups.

Greatly defined by its extraordinary cultural diversity, abundant natural resources, and strategic position between the Indian and Pacific Oceans, Indonesia has emerged as a major regional power, driven by expanding infrastructure, a growing middle class, digital innovation, and strong domestic consumption.

As a member of Association of Southeast Asian Nations and the Group of Twenty, the country plays an increasingly influential role in global trade, diplomacy, and sustainable development. Despite challenges related to geography, inequality, and environmental conservation, Indonesia continues to position itself as a dynamic nation with significant economic potential and cultural influence on the world stage.

But in this wealthy nation, wealth gap exists, and corruption thrives.

Then comes the 'Chromebookgate' scandal which involves everything between Google, one of Indonesia's biggest startups, politics, favoritism, and internet connectivity.

Former Indonesian Education Minister and Gojek co-founder Nadiem Makarim...

It all began as a high-profile corruption trial involving former Indonesian Education Minister Nadiem Anwar Makarim reached a critical juncture when state prosecutors demanded an 18-year prison sentence along with a fine of one billion rupiah or an additional 190 days in detention if unpaid and restitution payments totaling 5.68 trillion rupiah or nine more years behind bars if not fulfilled.

This demand which effectively could result in up to 27 years of imprisonment stems from allegations tied to the procurement of Zyrex Chromebook M4322 Chromebook laptops and associated Chrome Device Management software for the Ministry of Education Culture Research and Technology between 2020 and 2022 as part of a broader digitalization program launched during the COVID-19 pandemic.

Prosecutors from the Attorney General's Office claim that the initiative which had an overall value exceeding 9.9 trillion rupiah caused state losses of 2.18 trillion rupiah through overpricing unnecessary software licenses and specifications allegedly tailored to favor Google's Chrome operating system ecosystem over alternatives such as Microsoft Windows.

... is accused of high-profile corruption and graft tied to a multi-trillion rupiah government project to procure Chromebook laptops and management software for schools between 2019 and 2022.

They further allege that the devices proved ineffective in many remote areas due to heavy reliance on stable internet connectivity which remains limited across parts of Indonesia and that the process involved favoritism potentially linked to prior investments by Google in companies associated with Nadiem before his entry into government service.

Based on official company financial history and the formal indictments presented by state prosecutors in the corruption trial, Google's total investment in Gojek (later GoTo) between 2017 and 2021 reached $786.99 million.

And this Chrombookgate scandal first came under formal investigation by the Attorney General's Office on May 20, 2025, following reports of irregularities in the tender process for information and communication technology equipment intended to support remote learning and the Minimal Competence Assessment in schools.

Initial probes examined around 40 witnesses including special staff from the ministry and led to raids on private apartments and even the headquarters of GoTo the parent company of Gojek.

High school students holding the Zyrex Chromebook M4322 Chromebook laptops.

By mid-July 2025 four individuals were named as suspects including ministry officials Sri Wahyuningsih and Mulyatsyah as well as special staff members Ibrahim Arief and Jurist Tan with some facing immediate detention.

Nadiem himself was questioned, initialy as a witness, in late June 2025 and had his overseas travel restricted before being formally named a suspect and arrested on September 4, 2025, after which he spent time in Salemba Detention Center before later transitioning to house arrest on health grounds.

The trial opened in the Jakarta Corruption Court in January 2026 with co-defendants such as Ibrahim Arief receiving sentences of up to four and a half years in related proceedings earlier this year.

Throughout the hearings prosecutors have outlined what they describe as an orchestrated scheme involving the review and analysis of technology needs that steered procurement toward Chromebooks despite earlier internal assessments in 2019 by the ministry's communications and information technology center warning of their limitations in regions with poor connectivity.

Prosecutors allege that Nadiem's education digitalization program caused Rp 2.18 trillion (~$125–$132 million) in total state losses.

Key allegations by prosecutors:

Favoritism and rigged specifications: Nadiem and associates allegedly tailored tender specs to favor Google's Chrome OS ecosystem, sidelining alternatives like Windows. This allegedly made Google the "sole controller" of Indonesia's education tech. They claim he ignored or overrode internal advice that Chromebooks were unsuitable due to heavy internet dependency. Indonesia has poor connectivity in many remote/rural areas.

Overpricing and unnecessary purchases: Laptops were allegedly bought at inflated prices, plus unneeded software licenses (Chrome Device Management). This caused direct state losses of Rp 2.18 trillion.

Personal enrichment: Prosecutors accuse Nadiem of gaining Rp 809 billion (~$46 million) personally, plus holding Rp 4.87 trillion in unexplained wealth. They allege a quid pro quo: Google got the contract as payback for prior investments in Nadiem-linked companies (before he entered government). He allegedly used associates (including former Gojek-linked figures) to maintain indirect control.

Abuse of authority: Meetings with Google executives in 2020, issuance of related regulations (e.g., Permendikbud No. 5/2021), and bypassing proper procurement rules. Co-defendants include ministry officials like Sri Wahyuningsih and Mulyatsyah (some already sentenced to 4–4.5 years).

Nadiem and attorney Hotman Paris Hutapea.

Five vendors were involved in supplying the devices including local manufacturers Advan Digital Axioo and Zyrex along with Evercoss and PT Supertone and the program ultimately distributed over 1.1 million laptops modems and projectors to more than 77,000 schools according to ministry records from the period.

Prosecutors contend that the specifications were manipulated to lock in the Chrome operating system and Chrome Device Management licenses which they argue provided minimal additional value while inflating costs by hundreds of billions of rupiah.

They also point to alleged personal enrichment on Nadiem's part amounting to around 809 billion rupiah plus holdings of unexplained wealth estimated at 4.87 trillion rupiah claiming these flows traced back through entities linked to his pre-government business network.

In their closing arguments on May 13 prosecutors emphasized four main unlawful acts including deviations from proper procurement planning violations of principles in government purchasing rules and neglect of educational quality priorities in favor of personal interests.

Some co-defendants had already been convicted on related charges highlighting what the prosecution calls a collective conspiracy within the ministry.

Nadiem Makarim co-founded Gojek, one of the Indonesia's largest startups ever.

Nadiem Anwar Makarim who was born on July 4 1984 in Singapore to Indonesian parents with a background that includes an activist lawyer father and connections to prominent families grew up between Singapore and Jakarta attending elite schools before pursuing higher education abroad.

He earned a bachelor's degree in international relations from Brown University in 2006 and later an MBA from Harvard Business School.

After a stint as a management consultant at McKinsey and Company in Jakarta he ventured into entrepreneurship co-founding Zalora Indonesia and serving briefly as a Chief Innovation Officer at a payments firm before making a name for himself when he launched Gojek in 2010.

What began as a simple call-center service for motorcycle taxis and deliveries rapidly expanded into a super app encompassing ride-hailing food delivery logistics and digital payments becoming Southeast Asia's first decacorn with a valuation surpassing 10 billion dollars and attracting major investments including from Google.

Nadiem Makarim stepped down from his executive role at Gojek to become Indonesia's Minister of Education and Culture.

In October 2019 Nadiem resigned as Gojek's chief executive to join President Joko Widodo's cabinet as minister of education and culture later overseeing the merged ministry of education culture research and technology until October 2024.

During his tenure he championed reforms such as the "Guru Penggerak" teacher training program and a major push for digital tools in education arguing that technology could bridge gaps in access especially amid the pandemic's disruptions to in-person schooling.

He has consistently maintained that the Chromebook initiative was a legitimate response to urgent needs for remote learning tools and that prices were competitive with Chrome operating system costs being free compared to licensed alternatives.

Prosecutors claim that Nadiem's wealth increased exponentially following the scandal.

And in this Chromebookgate scandal, Nadiem has denied all charges asserting that he never personally signed any procurement documents and that decisions on implementation rested with director-general level officials in line with standard ministerial delegation of authority.

His legal team which includes prominent attorney Hotman Paris Hutapea has highlighted the absence of direct evidence showing any illicit financial flows to him personally noting that his declared wealth actually decreased by 51% during his time in office after he divested his Gojek shares upon entering government to avoid conflicts of interest.

Defense witnesses including former Google executives have testified that there was no connection between the company's earlier investments in GoTo and the education ministry's procurement choices underscoring that Google's involvement was limited to standard commercial supply without any quid pro quo.

Additional expert testimony has challenged the state loss calculations arguing that audit methodologies were flawed and that Chromebooks were procured at market-competitive rates with 85% of the devices still active and in use as of 2025.

Nadiem's supporters point out that the program aligned with a national digital education agenda and that any specifications were driven by policy goals rather than favoritism while some legal analysts have suggested the matter might better fit administrative rather than criminal proceedings. On the day of the prosecutors' demand Nadiem appeared emotional in court embracing his wife Franka Franklin and later described the hearing as profoundly disappointing while expressing hope that the facts presented would lead to his full acquittal.

"He's my boss," a Gojek driver said. "Back in the day, he showed up and opened up tons of income opportunities for millions of people. I'm one of those who felt it directly."

That sentence is simple, but it hits hard.

Because for a lot of Gojek drivers, Nadiem isn't just some entrepreneur. Instead, he is a figure who reshaped many people's lives. Some shared how they can pay rent every month thanks to Gojek orders, and even afford a house. Some can send their children to school, even college, from years of driving. Others claimed that they can cover their parents' medical bills.

Nadiem Makarim embraced his supporters, the Gojek drivers who see him as more than a founder, but a man who built a platform that created opportunities for millions. To them, he became a symbol of hope and economic mobility, a hero in the eyes of many who earned their livelihoods through the app.

Some used to do random odd jobs with no stability, then came Gojek, and many of drivers suddenly received a much more reliable daily income.

Even during the COVID-19 pandemic when tons of people lost formal jobs, some still held on because of driving and delivering food.

Now, Nadiem gets hit hard. Many of those people are confused. Some of the drivers are even willing to step in and take his position if he really has to go to prison.

When prosecutors demand 18-year sentence plus restitution, Nadiem expresses shock, noting it's harsher than some murder/terrorism penalties and that his assets are far under Rp 500 billion.

"What hurts even more, and I still don't understand, is that I have dedicated myself for 9-10 years serving this country, yet they still demand restitution money. So it's not enough for me to just go to prison? They used Rp 4 trillion + Rp 809 billion (total Rp 5 trillion). My wealth at the end of my term as minister wasn't even Rp 500 billion," Nadiem said.

"They took the peak value from my Gojek IPO/share in 2015, fictitious wealth that is not real, and used it as replacement money. They know I don't have that money. So why was it thrown at me? What's even more surprising: there is no connection to real wealth. The money I made created millions of jobs through Gojek. All the proof is there. Yet they still use it as a weapon of law to frighten me, to pressure me. I don't understand what the real reason is."

The case has sparked widespread debate across Indonesia with reactions ranging from concern over potential politicization of anti-corruption efforts to fears that it could discourage talented private-sector leaders from public service.

Critics including some political analysts and international observers have raised questions about selective prosecution and its implications for investor confidence in the country's tech ecosystem particularly given Google's explicit denial of any wrongdoing and the lack of charges against the company or its representatives.

Nadiem embracing his wife Franka Franklin.

Nadiem's father is also present during the trials..

Supporters of the investigation on the other hand emphasize the need for accountability in large-scale government procurements especially those involving billions in public funds during a national crisis.

The defense is scheduled to present its full arguments around early June 2026 with a verdict anticipated later that month. As the proceedings unfold they continue to draw attention not only for their impact on one of Indonesia's most prominent entrepreneurs but also for broader questions about governance innovation and the balance between oversight and progress in one of the world's largest democracies.

The outcome will likely influence future policy on public-private partnerships in education and technology while testing the resilience of Indonesia's judicial processes amid intense public scrutiny.

Codex Finally Goes Mobile As OpenAI Expands AI Coding Agent To ChatGPT App

Irvan — Sat, 16 May 2026 07:45:08 +0000

The competition among AI companies, often referred to as the LLM war, intensifies to productivity.

The launch of ChatGPT by OpenAI in 2022 marked the beginning of an intense period of competition in large language models. Within months, the tool gained over 100 million users and demonstrated the broad appeal of conversational AI. This prompted rapid responses from other technology companies. Google introduced its Bard model, later rebranded as Gemini. Anthropic released Claude, while Meta made its LLaMA models open source. Microsoft integrated advanced AI into its products, xAI launched Grok, and more that followed.

By 2024 and 2025, the focus shifted toward reasoning capabilities, with OpenAI advancing models like the o series that emphasized step by step problem solving. The competition extended to specialized applications, including software development tools.

OpenAI first introduced Codex in 2021 as a specialized model trained on code to generate programming snippets from natural language descriptions. It powered early versions of GitHub Copilot and demonstrated the potential for AI in software development.

However, OpenAI deprecated the original Codex models in its API. Newer general purpose models such as the GPT 4 had surpassed it in coding tasks, leading developers to migrate to those alternatives. The name Codex went on hiatus for roughly two years while broader model capabilities caught up.

Codex reemerged in May 2025 not as a simple code completion tool but as an autonomous coding agent.

This version functions as a software engineering teammate capable of handling complex tasks across codebases, including writing features, fixing bugs, and managing workflows in local or remote environments.

It operates through a dedicated desktop app, command line interface, and integrations, allowing it to run independently while providing updates and seeking approvals. Over the following year, Codex gained features such as background computer use, parallel agent execution, and support for hooks that customize its behavior with scripts for validation, logging, or repository specific rules.

In other words, OpenAI transformed Codex into a command center for building with agents.

Now, OpenAI announced an update that extends Codex access through the ChatGPT mobile application on iOS and Android.

You've been asking for this one...

Now in preview: Codex in the ChatGPT mobile app.

Start new work, review outputs, steer execution, and approve next steps, all from the ChatGPT mobile app. Codex will keep running on your laptop, Mac mini, or devbox. pic.twitter.com/9i2Jckjt9z

— OpenAI (@OpenAI) May 14, 2026

Users can now connect their phones to instances of Codex running on laptops, Mac minis, or development boxes.

From the mobile app, it becomes possible to initiate new tasks, review outputs such as code changes or test results, steer ongoing execution, and approve subsequent steps.

The actual computation remains on the connected machine, with the phone serving as a remote interface that syncs real time information including screenshots, terminal outputs, and file diffs via a secure relay system. The feature rolled out in preview across all supported regions and all ChatGPT plans, including free tiers. Support for connecting to Codex apps on Windows is scheduled for a future release.

This development builds on earlier capabilities such as general availability of remote SSH connections for managed environments and programmatic access tokens for scoped credentials on business and enterprise plans.

For eligible ChatGPT Enterprise workspaces, Codex also supports HIPAA compliant use in local setups.

The mobile integration allows developers to monitor and direct long running tasks without remaining tied to a desk, while files, credentials, and project context stay on the host machine.

Rolling out today as a preview on iOS and Android in all supported regions.

Support for connecting your phone to the Codex app on Windows is coming soon.https://t.co/jGNH2CUXih

— OpenAI (@OpenAI) May 14, 2026

Reactions on social platforms reflected mixed views. Some users welcomed the ability to manage coding agents on the go, sharing screenshots of task lists and live updates.

Others noted that the announcement coincided with ongoing requests for other model improvements, such as restoring access to certain GPT variants. The update positions Codex as a more flexible part of the broader AI coding ecosystem that has evolved since its initial 2021 release and 2023 deprecation.

Rivals like Anthropic, launched Remote Control for Claude Code in February 2026.

This feature, similar to this Codex feature, allows users to run a local Claude Code agent session on their machine and control it remotely from the Claude mobile app or web interface.

Published:

16/05/2026

Opera One Gets Major Ad Blocker Overhaul With Native uBlock Compatibility And A Focus On Performance

Danny Krisnanto — Fri, 15 May 2026 15:00:35 +0000

Opera has offered a built-in ad blocker for a decade.

This makes it one of the earliest major browsers to include the feature natively without any need for separate extensions. But an update released in mid-May 2026 for Opera One brings the ad blocker to a whole different level.

This is because it has been completely rebuilt engine that improves speed and efficiency while adding compatibility with uBlock lists and filters.

This change aligns the native tool more closely with industry-standard blocking capabilities that users previously sought from dedicated extensions.

The upgrade focuses on better performance during everyday browsing, especially for video streaming on platforms like YouTube or Twitch.

Start browsing without getting interrupted!

Opera’s Ad Blocker is upgraded:

• Rebuilt engine
• uBlock compatibility
• Fully Native
• No extensions needed

With this upgrade, Opera gets even better for Video Streaming. pic.twitter.com/65cNGkC649

— Opera (@opera) May 14, 2026

Ads and trackers are handled more effectively at the browser level, which can lead to faster page loads and fewer interruptions when sharing or watching content.

Because the blocker is native, activation remains straightforward: users toggle it on through the privacy shield icon or settings, with options to whitelist specific sites if desired. No installation or configuration is required.

Compared to the dominant browsers from Google, Microsoft Edge, and Apple Safari, Opera occupies the position of an underdog in market share.

Yet it has maintained a distinct identity through features like this built-in ad blocker, alongside other tools such as a free VPN and sidebar integrations that prioritize user control and convenience.

So why do you need to use an Ad Blocker? We’ll tell you:

• Better video streaming experience
• Faster page loading
• Fewer annoyances
• Better privacy

Start your day with Opera, turn on Ad Blocker, and browse without annoying distractions.

— Opera (@opera) May 14, 2026

The rebuilt ad blocker also supports long-term compatibility by operating within modern web standards, including aspects of Manifest V3, while preserving the simplicity that has defined Opera's approach.

For users tired of managing extensions or dealing with slowdowns from heavy ad-laden pages, the update offers a practical, low-effort improvement.

Testing the feature on the latest version of Opera One shows noticeable differences in load times and cleanliness on ad-heavy sites.

Overall, this refinement keeps the browser's longstanding ad-blocking strength relevant in a landscape where privacy and performance demands continue to grow.

Published:

14/05/2026

Not Willing To Provide Google A Phone Number? New Accounts Are Getting Only 5GB Storage

Danny Krisnanto — Fri, 15 May 2026 10:00:07 +0000

Number is just a number. Yes, partly true, but entirely crucial when it comes to storage space.

Google has quietly adjusted its free storage policy for new accounts, shifting from the usual 15GB to a default 5GB limit when users create an account without linking a phone number. This can be devastating for a lot of users since that amount of storage space covers pretty much everything a user has.

From storing emails on Gmail, files on Google Drive, and photos and videos on Google Photos.

Without a phone number, users will be capped at just 5GB.

But upon adding a phone number during or after signup will restore the full 15GB allocation at no extra cost.

Before, Google said that free users get 15GB...

The change first gained attention in early May 2026 through user reports on forums, particularly from regions in Africa and other select areas.

During account creation, a prompt now appears offering two paths: continue with 5GB or verify a phone number for the higher limit. Google explains that the phone requirement helps prevent abuse by limiting multiple free accounts per person while also strengthening account security and recovery features. In many setups, phone verification is already mandatory, so the 5GB option mainly applies when that step can be skipped.

It's worth noting that the change only affects new users.

Existing accounts remain untouched and continue to receive the standard 15GB.

Google has framed the adjustment as a limited test in specific regions rather than a full global policy shift. The company has also updated some support documentation to describe free storage as “up to 15GB,” signaling that the amount may depend on verification. This reflects ongoing challenges in managing cloud infrastructure amid rising demand.

Now, Google says that free users get "up to 15GB."

Public responses have been mixed.

Some users see the phone verification as a reasonable trade-off to curb spam and fake accounts, while others worry about the increased collection of personal information.

Those who prefer not to share a number can still create and use accounts with the reduced 5GB cap, though they may need to be more diligent about managing storage or consider Google One subscriptions for additional space.

The rollout is not uniform, so the experience can vary based on location and how the account is set up. New users are advised to review storage options directly during signup or in account settings afterward.

Published:

15/05/2026

News

Google

Review