OpenAI produced two mismatched versions of its Navier-Stokes proof, one for humans and one for computers, after mistranslating mathematics into code. ·
26 points and 2 comments on Hacker News, linking newscientist.com.
OpenAI releases Codex version 0.162.1, which fixes a terminal crash involving multi-line questions and resolves startup failures caused by feature setting mismatches. ·
Open-slopware lists alternatives to free and open-source software projects that adopt large language models, and offers tips for requesting better policies or forking code. ·
A researcher expects rapid AI progress but doubts it will reach general superintelligence, noting surprise at industry peers who predict AI will outperform them in their jobs within a few years. ·
OpenAI produced two mismatched versions of its Navier-Stokes proof, one for humans and one for computers, after mistranslating mathematics into code. ·
OpenAI releases Codex version 0.162.1, which fixes a terminal crash involving multi-line questions and resolves startup failures caused by feature setting mismatches. ·
The author describes the super intelligence trend as a humiliation ritual for OpenAI and the broader sector, while also calling it relatively smart in terms of regime politics. ·
TokenRouter is a serving system for token-level LLM routing that uses request-centric programming and model-centric execution to dispatch requests asynchronously across subservers. ·
HF Daily Papers24h ago▲arxiv.org
X voices
5 stories
X posts other sources are linking to, then named voices on Bluesky and Mastodon
Firmus and PLDT withdrew their data center IPOs this week, while regional governments paused development, creating obstacles for the Asia-Pacific data center expansion. ·
Data Center Knowledge5h ago▲datacenterknowledge.com
Firmus and PLDT withdrew their data center IPOs this week, while regional governments paused development, creating obstacles for the Asia-Pacific data center expansion. ·
Data Center Knowledge5h ago▲datacenterknowledge.com
A calculator converts daily, monthly, or yearly token volumes into estimated counts of A100, H100, or B200 GPUs, including a sourced uncertainty range. ·