Posts

There’s something deeply uncomfortable about the timing. OpenAI is barreling toward what could be the most anticipated tech IPO in a decade — a valuation north of $1 trillion, 900 million weekly users, the kind of numbers that make Wall Street salivate. And right in the middle of that momentum, Florida’s attorney general just dropped a bomb. On April 9th, Florida AG James Uthmeier announced a formal investigation into OpenAI and ChatGPT, citing concerns about child safety, national security, and the chatbot’s alleged role in a mass shooting at Florida State University. Subpoenas are coming. The stakes couldn’t be higher. ...

Mark Zuckerberg just played his most expensive hand yet. Meta unveiled Muse Spark, the first AI model from its Superintelligence Labs — the division built on a $14.3 billion Scale AI investment and a talent raid that cost hundreds of millions in individual engineer packages. After more than a year in the wilderness following the embarrassing Llama 4 launch, Meta says it’s back. But here’s the twist nobody expected: the company that championed open-source AI just went proprietary. ...

Remember when ChatGPT just answered your questions without trying to sell you anything? Those days are officially over — and OpenAI is counting on it to the tune of $100 billion. OpenAI has told investors it expects $2.5 billion in advertising revenue this year, scaling to $100 billion by 2030. That’s building an ad business in four years that took Google and Meta decades to create. This isn’t some side experiment. This is the company that kicked off the AI revolution fundamentally reshaping how the internet makes money. ...

Every few months someone claims AI is going to “accelerate science.” It usually means a marginally better protein fold. This week is different. This week, an AI system took a quantum algorithm that its own human authors had nearly thrown in the trash, rewrote it, and handed back a result that made the people who secure most of the internet publicly move up their doomsday clock by six years. If you read one AI story this week, make it this one. ...

The Nvidia Era Is Over — At Least in China DeepSeek’s next flagship model, V4, will run exclusively on Huawei Ascend chips. Not as a backup. Not as a proof of concept. As the entire inference stack. That’s not a press release talking point. That’s a tectonic shift. For years, China’s AI labs quietly depended on Nvidia silicon — H100s, A100s, whatever they could get their hands on through official channels or creative workarounds. That dependency is now ending, and it’s ending fast. ...

Forget another chatbot upgrade. The biggest story in AI right now is that the machines are starting to build themselves. Not in the Terminator sense — nobody’s assembling robot armies in a garage. But in a quieter, more consequential way: AI systems are writing the code, optimizing the training runs, and designing the infrastructure that powers their own successors. And the companies behind them aren’t hiding it. They’re putting it on product roadmaps. ...

A Chatbot With a Prescription Pad Utah just became the first state to let an AI chatbot renew psychiatric medications without a doctor signing off. The company is Legion Health. The product is a chatbot. The price is $19 a month. And starting this month, it can keep you on Prozac, Zoloft, Wellbutrin, Lexapro, and 11 other psychiatric drugs — no human physician required. This is not a drill. This is not a research paper. This is live policy in the United States. ...

There’s a moment every product hits where it has to choose between what made it famous and what keeps it alive. Cursor just made that call. On April 2, 2026, Cursor launched version 3 — codenamed “Glass” — and it’s not an update. It’s a philosophical coup. The default interface is no longer a code editor with AI sprinkled on top. It’s a mission control dashboard for fleets of AI agents. The file explorer? Gone from the default view. The code you write yourself? Optional. ...

Micron just lost over $100 per share. Samsung shed 5%. DDR5 prices dropped 30%. The culprit? A compression algorithm from Google Research that makes AI models need dramatically less memory. And it hasn’t even shipped as a product yet. The DeepSeek Sequel Nobody Expected TurboQuant does something deceptively simple: it compresses the key-value cache — the short-term memory AI models use during inference — by 6x while making inference 8x faster. Zero accuracy loss. No retraining. No fine-tuning. Just plug it into your existing pipeline. ...

OpenAI’s next flagship AI model is codenamed after a potato. And it might be the most important thing the company has built since ChatGPT. Greg Brockman, OpenAI’s president, confirmed on the Big Technology podcast that Spud has finished training. It’s not an incremental GPT-4o update. It’s an entirely new base model — nearly two years of research condensed into what Brockman calls a “major step toward AGI.” The timing is no accident. Spud lands in the same week as OpenAI’s $122 billion fundraise, the death of Sora, and Anthropic’s alarming Mythos leak. Welcome to the most consequential week in AI this year. ...