Artificial Intelligence

Someone finally did it. A developer named Andrew Vos published a GitHub plugin called Endless Toil that makes your coding agent emit human groans while it reads through your code. The worse the code, the louder and more desperate the sounds get. "Hear your agent suffer through your code," the repo reads. It works by running alongside agents like Claude or Codex in real time, scanning the code being processed and triggering escalating recorded groans based on how cursed things look. A mild mess g...

Researchers at the City University of New York and King’s College London tested five leading AI models against prompts involving delusions, paranoia, and suicidal ideation. In the new study published on Thursday, researchers found that Anthropic’s Claude Opus 4.5 and OpenAI’s GPT-5.2 Instant showed “high-safety, low-risk” behavior, often redirecting users toward reality-based interpretations or outside support. At the same time, OpenAI’s GPT-4o, Google’s Gemini 3 Pro, and xAI’s Grok 4.1 Fast sho...

The U.S. Department of Justice moved Friday to intervene in xAI’s lawsuit against Colorado, escalating a legal fight over how states can regulate artificial intelligence and whether companies can be held liable for “algorithmic discrimination.” In a press release, the DOJ said Colorado’s law, SB24-205, violates the Equal Protection Clause of the Fourteenth Amendment because it requires AI companies to prevent unintentional “disparate impact” based on protected characteristics such as race and se...

South Korean police arrested a man Thursday for posting an AI-generated photo of an escaped wolf—an image convincing enough to fool city officials and trigger an emergency alert to thousands of residents. According to police, the stunt delayed the capture of an actual wolf, which had escaped from a zoo two weeks prior, by up to nine days. Daejeon Metropolitan Police charged the unnamed 40-year-old with obstructing official duties by deception, specifically for "distributing fabricated wolf sigh...

Anthropic, the artificial intelligence company behind the Claude chatbot, announced Friday a set of new election integrity measures designed to prevent its AI from being weaponized to spread misinformation or manipulate voters ahead of the 2026 U.S. midterm elections and other major contests around the world this year. The San Francisco-based company detailed a multi-pronged approach that includes automated detection systems, stress-testing against influence operations, and a partnership with a...

DeepSeek is back, and it showed up a few hours after OpenAI dropped GPT-5.5. Coincidence? Maybe. But if you're a Chinese AI lab that the U.S. government has been trying to slow down with chip export bans for the past three years, your sense of timing gets pretty sharp. The Hangzhou-based lab released preview versions of DeepSeek-V4-Pro and DeepSeek-V4-Flash today, both open-weight, both with one million token context windows. That means you can basically work with a context roughly the size of...

On secondary share trading platforms, Anthropic has quietly flipped the AI power map—trading at roughly $1 trillion, overtaking OpenAI for the first time. On Forge Global, one of the leading private-share marketplaces, Anthropic is hovering around the $1 trillion mark, according to Forge CEO Kelly Rodriques, who confirmed the figure to Business Insider. OpenAI, meanwhile, is trading at approximately $880 billion on the same platform. just got offered a $1.05T valuation on my anthropic shares f...

Hollywood location scouts may soon be able to do more of their work from a laptop, thanks to artificial intelligence. Announced during Google Cloud Next on Wednesday, Google unveiled Maps Imagery Grounding. This new AI-powered tool lets filmmakers and creative agencies generate images and animated scenes based on real-world locations using Google Street View data, potentially cutting the time and cost of early-stage location scouting and storyboarding. “Generative AI has unlocked powerful ways...

Meta informed workers of sweeping layoffs Thursday, the New York Times reported, targeting roughly 8,000 employees—about 10% of its global workforce—while also eliminating 6,000 unfilled positions, as the Facebook and Instagram parent company funnels unprecedented resources into artificial intelligence development. The cuts were outlined in an internal memo from Janelle Gale, Meta's chief people officer, who framed the reductions as a necessary trade-off to fund the company's broader strategic...

The Microsoft-backed blockchain network Space and Time on Thursday launched Dreamspace, an AI-powered app builder designed to let users create and deploy on-chain applications and smart contracts without writing code. The platform has been in beta since August 2025, according to Space and Time co-founder Scott Dykstra, and the company said more than 34,000 apps were created before Thursday’s public release. “It’s about making it really easy to deploy an app that has a smart contract, where that...

OpenAI launched GPT-5.5 on Thursday, pitching it essentially as a model targeted at agentic computer use. It writes and debugs code, browses the web, fills out spreadsheets, and keeps working through multi-step tasks without needing a human to babysit every move. The release is already rolling out today to Plus, Pro, Business, and Enterprise subscribers across ChatGPT and Codex, OpenAI said. “We’re releasing GPT‑5.5, our smartest and most intuitive-to-use model yet, and the next step toward a n...

OpenAI on Wednesday unveiled a free, specialized version of ChatGPT for physicians, nurse practitioners, physician assistants, and pharmacists, deepening the company's push into a healthcare sector grappling with staffing shortages, administrative overload, and burnout. The product, called ChatGPT for Clinicians, is designed to handle documentation, medical research and care consultations—time-consuming tasks that eat into the hours clinicians can spend with patients. Access is currently limite...

Tencent quietly dropped its most capable AI model yet on Thursday, and the benchmark numbers are hard to ignore. Hy3 preview, the company's first model after a full infrastructure rebuild, went open-source today across GitHub, Hugging Face, and ModelScope. It’s also available on Tencent Cloud’s official website, under a paid plan. My3 packs 295 billion total parameters (a measurement of a model’s potential breadth of knowledge) but only 21 billion active at any given time. That's the beauty of...

The White House warned Thursday that “foreign entities” are allegedly carrying out “industrial-scale” campaigns to copy the capabilities of American-based artificial intelligence models, using tactics including jailbreaking and networks of fake accounts to extract proprietary information and replicate their performance. In a memorandum titled “Adversarial Distillation of American AI Models,” Assistant to the President for Science and Technology and Director of the Office of Science and Technolo...

Artificial intelligence is expanding the government’s ability to analyze Americans’ digital records. A new bill aims to require a warrant before federal agencies can access that data. Kentucky Rep. Thomas Massie and Colorado Rep. Lauren Boebert introduced the Surveillance Accountability Act on Thursday. If passed, the legislation would amend Title 18 of the U.S. Code to codify a broad warrant requirement for government searches, close what supporters describe as the “third-party doctrine” looph...

Google unveiled two AI processors at its Cloud Next 2026 conference in Las Vegas on Wednesday, marking the company's eighth generation of custom silicon designed to challenge Nvidia's AI chip dominance. The training-focused TPU 8t delivers nearly 3x the compute performance per pod compared to its predecessor, with a single superpod scaling to 9,600 chips and delivering 121 ExaFlops of compute capacity. The architecture also offers 2.8x better price-to-performance, according to Google. The TPU 8...

OpenAI is pushing ChatGPT beyond the chat box with the launch of “workspace agents,” a new feature that lets businesses automate recurring tasks even when employees are offline. Announced on Wednesday, OpenAI said in a post that, unlike the custom GPTs users have built in the past, workspace agents are powered by OpenAI’s Codex model and run as persistent assistants that can connect to external apps, retain information across projects, and complete multi-step workflows without repeated prompts....

Xiaomi just launched a new AI model family. Again. A few weeks ago, the company dropped MiMo-V2-Pro—a trillion-parameter model that had been quietly circulating on OpenRouter under the alias "Hunter Alpha" before Xiaomi revealed its identity. It went from anonymous to top-tier overnight. We tested it, and it was impressive. Now Xiaomi is back with MiMo-V2.5 and MiMo-V2.5-Pro, a two-model family that adds something the previous generation never had in a single package: eyes, ears, and the abili...

A new investment product called USVC is designed to disrupt venture capital norms and provide retail investors access to some of the most successful private companies—like OpenAI, Anthropic, and Elon Musk’s xAI—for as little as $500.  The venture capital fund is offered by investment infrastructure firm AngelList, and is available for all U.S. investors.  “Go back to the 1500s, you set sail for the new world to find tons of gold—that was ‘adventure capital.’ Early-stage technology is the modern...

Even the Pope’s warnings about artificial intelligence may have been written by artificial intelligence. First reported by Wired, Pangram Labs, an AI-detection startup, said its updated Chrome extension flagged several posts from Pope Leo XIV’s official X account as likely AI-generated, including messages warning followers about AI’s effect on human thought and social structures. One of those posts warned: “When simulation becomes the norm, it weakens the human capacity for discernment.” “Pangr...