Posts

Three bombshells in 72 hours. OpenAI launched GPT-5.5. China’s DeepSeek dropped V4 at 85% less cost. Google committed $40 billion to Anthropic. The AI industry didn’t just shift — it lurched into a new phase where billion-dollar moves happen simultaneously and the gap between cutting-edge and affordable collapses faster than anyone predicted. GPT-5.5: The “Just Let It Work” Model OpenAI’s latest isn’t about raw intelligence gains. It’s about how the model works. ...

Your phone knows where you slept last night. Your email provider knows who you argued with this morning. Your browser knows what you’re worried about. Individually, these data points are mundane. Assembled by an AI system with access to government surveillance databases? They become the most detailed profile of your life ever constructed — and right now, Congress is fighting over whether to let that happen without a warrant. The FISA Fight Reaches Breaking Point Section 702 of the Foreign Intelligence Surveillance Act lets the government collect communications of foreigners abroad without a warrant. When an American contacts someone overseas, their communications get swept up too. The government can then search those intercepted American communications — no warrant required. ...

AI has an energy problem. A massive one. Data centers now consume roughly 1,000 terawatt-hours of electricity annually — about as much as Japan. Every ChatGPT query, every image generation, every AI-powered search burns through watts at a rate that would make your electric bill weep. So when a team at the University of Cambridge publishes a chip design in Science Advances that could cut AI energy consumption by up to 70%, it’s time to pay attention. ...

April 24, 2026 might be the most consequential single day in the AI model wars. Within hours of each other, two of the year’s most anticipated systems went live: DeepSeek’s V4 — an open-source behemoth running on Huawei chips with a million-token context window — and OpenAI’s GPT-5.5, which the company calls its “smartest and most intuitive model yet.” Neither was an incremental update. Both represent genuine leaps. And the fact that they dropped on the same day tells you everything about where this industry is right now. ...

There’s a moment in every tech revolution when the money stops making sense to normal people. For AI, that moment arrived this week. Google just committed up to $40 billion in Anthropic — the company behind Claude. This landed four days after Amazon announced its own $25 billion deal with the same company. Combined: $65 billion in potential investment from two competing tech giants, into a single startup, in one week. ...

The federal government is running a live experiment on 30 million Americans. An AI program called WISeR is delaying and denying medical care for seniors. The nation’s top health official says AI could make the FDA “irrelevant.” And somehow, nobody in charge sees the contradiction. The Quiet Launch That’s Ruining Lives In January 2026, CMS launched the Wasteful and Inappropriate Service Reduction Model — WISeR — introducing AI-powered prior authorization to traditional Medicare for the first time at scale. ...

A year ago, DeepSeek was a curiosity — a scrappy Chinese lab that somehow built a model rivaling OpenAI’s o1 for a fraction of the cost. Markets panicked. Think pieces multiplied. Then the hype faded. DeepSeek didn’t. Today the Hangzhou-based lab released V4 in two variants — V4-Pro and V4-Flash — and the numbers are hard to ignore. This is the largest open-weights AI model ever released, and it’s trading blows with the best closed-source systems from OpenAI, Google, and Anthropic. ...

If you were writing a dystopian tech thriller, you’d struggle to invent a plot more on-the-nose than what Meta announced this week. The company is cutting 10% of its workforce — roughly 8,000 people — while simultaneously rolling out software that tracks employee mouse movements, clicks, and keystrokes to train AI models. The models that will, eventually, do their jobs. The Tracking Tool Nobody Asked For Reuters broke the story earlier this week. Meta is deploying an internal tool called the Model Capability Initiative (MCI) to U.S.-based employees. It captures mouse movements, clicks, keystrokes, and periodic screenshots. The goal: teach AI models the messy, real-world ways humans interact with computers. Dropdown menus, keyboard shortcuts, the digital muscle memory that machines still can’t replicate on their own. ...

Less than two months between GPT-5.4 and GPT-5.5. Even OpenAI president Greg Brockman admitted during Thursday’s press briefing that “there are probably enough model releases that it’s getting hard to distinguish one from another.” He’s right. But GPT-5.5 deserves your attention anyway — not because of what it promises, but because of what it reveals about where this industry is heading at full speed. The Pitch: Less Hand-Holding, More Doing OpenAI’s headline claim is deceptively simple: GPT-5.5 can do more with less guidance. Hand it a messy, multi-part problem and it figures out the plan on its own — choosing tools, checking its work, iterating toward a solution. ...

Four days. That’s the gap between Meta announcing 8,000 layoffs and Reuters revealing that the company is recording every keystroke, mouse movement, and screen action its employees make — to train AI that does their jobs. You can’t make this stuff up. The Timeline That Says Everything April 17: Reuters reports Meta plans to cut 10% of its 78,865-person workforce starting May 20, with more cuts planned for late 2026. ...