AI
6 items · 11 articles in this topic
The Riemann claims dominate, but the more durable items are measurement and safety: METR's updated time horizon plot, a paper on models rewriting their own weights, and OpenAI's distillation defenses.
- 1
METR updated its time horizon plot. It finds AI capabilities still improving exponentially but lagging on judgment-heavy tasks.
- 2
An arXiv paper examines whether language models can rewrite their own weights. The summary gives no results.
- 3
OpenAI published a post describing its approach to combatting adversarial distillation of its models. The summary gives no technical detail.
- 4
Thinking Machines released Inkling 2 Mini, a smaller model in its Inkling line.
- 5
A paper argues that concealed commercial incentives in LLMs erode epistemic rigour in their outputs, which it calls the 'Hungry Butler Effect'.
NotableHacker News (front page) ↗
- 6
The Dutch data protection regulator reports that AI smart glasses misinterpret medication instructions and pose safety risks.
More in AI · 5 articles›
The Economist reports AI may have solved a major unsolved problem in mathematics.
Hacker News (front page)aimathematicsresearchFeed excerpt
Article URL: https://www.economist.com/science-and-technology/2026/11/20/ai-may-have-solved-the-most-important-problem-in-mathematics Comments URL: https://news.ycombinator.com/item?id=50583445 Points: 5 # Comments: 2
An AI system and a human both reportedly produced solutions to the 167-year-old Riemann Hypothesis, drawing mathematicians' attention.
TechCrunchmathllmresearchFeed excerpt
Mathematicians are having their minds blown by the latest AI-driven result — and there's a competing human solution, too.
Anthropic mathematician Levent Alpöge claims Claude helped prove the Riemann hypothesis, a claim not yet independently verified.
The VergellmclaudemathematicsanthropicFeed excerpt
Anthropic went public on November 19th, 2026. | Photo: Getty Images Anthropic mathematician Levent Alpöge claims he has proved the Riemann hypothesis, one of mathematics's most notoriously difficult problems, with the help of Claude, cracking a puzzle that has eluded some of the world's best mathematicians for decades. If verified, it would mark yet another seismic achievement for AI in mathematics and add to the upheaval already unleashed by OpenAI's previous, often inelegant forays into the field. Alpöge's solution has yet to be verified independently. Anthropic has not commented or involved itself in Alpöge's announcement. In a document on his website, Alpöge says he texted Anthropic CEO Dario Amodei abo … Read the full story at The Verge.
A developer shrank ChatGPT-style language model software to run on a Super Nintendo.
The Registerllmretro-computingdemoFeed excerpt
Just don't expect complex, coherent, or always correct answers
ClawDOS is a hobby project running an AI coding agent on MS-DOS.
Hacker News (front page)ai-agentretro-computingcoding-agentFeed excerpt
Article URL: http://ricapar.net/ClawDOS/ Comments URL: https://news.ycombinator.com/item?id=50586292 Points: 53 # Comments: 15