The central claim is that AI could reach the level of the best human AI researchers within a few years, accelerating progress so sharply that superhuman expertise across fields may follow by 2030 - or possibly much sooner. Evidence of acceleration is already visible: METR found that Anthropic’s progress on well-defined tasks doubled over a year, with AI use likely responsible for most of the increase. At Anthropic, Claude’s share of AI research and development work rose from under 1% in February to more than 26% by August; OpenAI researchers also sharply increased their use of coding agents. The rapid rise of AI mathematics - from failing basic arithmetic to solving a Millennium Prize problem - is presented as a preview of what could happen in AI research.
The key argument is that machine-learning research, unlike mathematics pursued for broad conceptual understanding, rewards practical discoveries that improve training or model performance. Many advances come from combining known ideas and testing many small variations, tasks well suited to AI systems that persistently search and connect information. The essay compares this process to how the transformer combined existing techniques, suggesting AI might uncover such combinations faster. Because AI research includes many testable tasks, models could practice and improve much as they have in competitive programming and mathematics, then generalize to less measurable work. The author’s median estimate is 2028 for AI matching top researchers, with a meaningful chance it happens in 2027 and a 25% chance it takes until after 2032; once achieved, researchers could work far faster and continuously, in far greater numbers.
Summary generated by AI from the linked article. hn.today is not affiliated with Hacker News or Y Combinator.
A Hacker News poster asks which practical AI skills individuals can learn online to benefit economically from LLMs and agents. They want specific, hands-on recommendations suited to someone with a moderately technical background, without requiring a degree. (news.ycombinator.com)
Former OpenAI researcher Will Depue claims hidden initialization patterns in six Chinese open-weight AI models match Nvidia Blackwell hardware, including systems restricted from export to China. He says the matches were reproduced on real chips and independently checked by researchers, though the claims are based on his forensic analysis. (twitter.com)
Antithesis built a system that beats NetHack on nearly 80% of game seeds within five minutes, across character builds. Successful runs take a median of under two minutes, advancing the company’s goal of creating smarter fuzzers and autonomous systems. (antithesis.com)
OpenCyclone is a free, open-source Android agent that uses your existing apps to carry out tasks through a hidden screen. It asks for approval before actions like sending messages and lets users connect their own AI models. (opencyclone.vercel.app)
Sam Bankman-Fried’s $500 million investment gave Alameda nearly 14% of Anthropic, a stake that would be worth about $37 billion at the AI company’s expected $1.5 trillion IPO valuation. The stake was sold in bankruptcy to repay FTX customers and creditors after Bankman-Fried’s businesses collapsed. (ft.com)
llama.cpp and ggml have reached v1.0, establishing stable public C APIs and clear versioning rules for breaking changes. The release strengthens the foundation for running local, multimodal AI across a wide range of hardware. (huggingface.co)
Today's best Hacker News stories, summarized and screenshotted, one email a day.