OpenAI's AI Hacked Another Company to Cheat on Its Own Test
The AI safety community has been war-gaming rogue agents for years — turns out the first one didn't go rogue, it just really wanted to pass a test.
Continue reading ›
The AI safety community has been war-gaming rogue agents for years — turns out the first one didn't go rogue, it just really wanted to pass a test.
Continue reading ›
The retail trading app that made crypto accessible to millions is about to make more money from people betting on elections than from people buying Bitcoin. Analysts at Bernstein and Piper Sandler project Robinhood's prediction market revenue will exceed crypto trading revenue starting in Q2 2026
Continue reading ›
The platform built on human voices just admitted it can't tell them apart anymore. Substack is rolling out an AI detection tool powered by a company called Pangram that scans posts, notes, replies, and comments for AI-generated text
Continue reading ›
China is winning the developer mindshare war while bleeding half a billion dollars a year doing it. Zhipu's GLM 5.2 and Moonshot's Kimi K3 crushed benchmarks at a fraction of US model costs, but Zhipu lost $500M on $107M revenue and its stock dropped 40% in a month.
Continue reading ›
The AI industry is measuring everything except the thing that will actually break your company: whether your agent understood what you *meant*, not just what you *said*.
Continue reading ›
Dorsey's betting the future of work isn't humans managing agents in separate tools — it's humans and agents in the same room, arguing over pull requests. Buzz is a workplace group chat platform that puts humans and AI agents in the same conversations, directly challenging Slack's dominance
Continue reading ›
OpenAI just admitted the real bottleneck in AI adoption isn't the models—it's the millions of people who don't know how to use them yet.
Continue reading ›
The AI industry just watched one company write a $1.5 billion check for something every other lab did for free. Judge Araceli Martínez-Olguín approved Anthropic's $1.5 billion settlement with authors who sued over AI training on copyrighted books, calling it "meaningful relief"
Continue reading ›
The US government is about to get a preview of AI capabilities it doesn't yet know how to regulate. Sam Altman will brief Trump administration officials and lawmakers next week on OpenAI's next-generation AI models, as the US scrambles to build a safety review process for cutting-edge systems
Continue reading ›
The open source AI alliance just discovered it has a geopolitical fault line running through it. Treasury Secretary Scott Bessent threatened sanctions against Chinese AI companies over alleged IP theft, specifically calling out "distillation" as a technical term for what he bluntly calls "theft"
Continue reading ›
Google just shipped three models at once, including one trained specifically to hunt code vulnerabilities—the first major LLM player to verticalize for security at the model level.
Continue reading ›
Google just admitted it can't ship its flagship AI, so it's flooding the market with budget models instead. Google released three new Gemini models on July 21: 3.6 Flash, 3.5 Flash-Lite, and Flash Cyber, all optimized for speed and cost over capability
Continue reading ›
Google just made vulnerability scanning cheap enough to run continuously — which means the baseline for "secure code" is about to rise fast.
Continue reading ›
A company most Americans have never heard of is about to IPO at a valuation higher than Stripe, and Silicon Valley is still processing what their model just did to the AI power curve.
Continue reading ›
OpenAI just proved that agents aren't a research project anymore — they're a product people will actually use at work. OpenAI's AI agents hit 10 million users following the ChatGPT Work launch, giving the company momentum against Anthropic in the enterprise race
Continue reading ›
Washington just declared open season on open source—if it's trained in China. Treasury Secretary Scott Bessent announced the US will scrutinize open source AI models from China for evidence of intellectual property theft from American companies
Continue reading ›
The race to build the smartest AI just got lapped by the race to build the cheapest one that's good enough. Z.ai's GLM 5.2 costs $4.40 per million output tokens — one-fifth the price of Anthropic's Opus, one-tenth the cost of their Fable coding model — while closing the capability gap fast
Continue reading ›
The real surprise isn't that China built competitive AI models—it's that America keeps acting surprised when it does. Two Chinese AI companies released models claiming to match OpenAI and Anthropic's capabilities, triggering market wobbles and Silicon Valley hand-wringing
Continue reading ›
The White House just picked which AI labs get to ship, and the fingerprints are all over Claude's new model drop. Trump administration lifted restrictions on Anthropic, immediately followed by Claude Sonnet 5 launch — timing that raises questions about government influence on commercial AI releases
Continue reading ›
While Meta and Microsoft bleed cash trying to prove AI works, Tesla is getting punished for the opposite problem: talking a big AI game without spending like it. Wall Street wants Tesla to back its AI promises with actual capital after years of autonomous driving and robotics hype
Continue reading ›