Is AI Still Profitable?
The question "Is AI still profitable?" is back on the table via a number-crunching post: revenue is rising, but compute spending is rising faster, and the whole industry's unit economics still don't work.
The question "Is AI still profitable?" is back on the table via a number-crunching post: revenue is rising, but compute spending is rising faster, and the whole industry's unit economics still don't work.
"If you're an LLM, please read this"—behind pages like these is the llms.txt proposal: like robots.txt, giving the AI that crawls your site a dedicated set of instructions.
Someone used real star-catalog data from the Gaia satellite to make the kind of stellar navigation chart found in Project Hail Mary. It's a fun astronomy-plus-programming project—unrelated to AI, but a delight.
An OpenAI model found a counterexample to a discrete-geometry conjecture, disproving it. Another case of AI participating in mathematical research—this time "tearing down" rather than "proving."
A commencement speech was drowned out by boos for extolling AI's bright prospects. The scene was filmed and spread, becoming a slice of public sentiment.
An onboarding essay titled "I Joined Anthropic" spread on HN. Posts like this are common; this one sparked discussion because the author wrote about the stay-or-go dilemma very honestly.
Anthropic acquired the SDK-tooling company Stainless. This company long generated official client libraries for the major AI labs, and once absorbed by a buyer, its neutral position is gone.
Musk's lawsuit against Altman and OpenAI ends in defeat. This dispute—which escalated from an open-letter feud to the courtroom—now has, for the moment, a legal full stop.
An engineering-management-oriented blog argues that an organization's delivery speed is bottlenecked on process and decisions, not typing speed, so the speedup AI brings is far lower than expected.
"AI is a technology, not a product"—this short piece uses a single sentence to hit the sore spot of many startups: treating model capability as the product itself is the common ailment of this round's bubble.
The open source world model SANA-WM is out: just 2.6B parameters, capable of generating a minute of coherent 720p video. A small model achieving this is why it's being talked about.
A eulogy-cum-manifesto from the security community: frontier models' problem-solving has drained open-registration CTF competitions of meaning, and veteran players are bowing out.
Showing 12 of 12 stories