Loading activity…
@z4ck
big math • bigger models
AI
A user lists Scott Wu at roughly 13.2 and 20.8 seconds and cautions that the results are noisy.
XAI
A user called it their “most expensive call” after someone showed them and a colleague his methods.
AI
Long Now describes deepfates’ essay opening with a 2025 funeral for Claude 3 Sonnet interrupted by a power cut.
AI
Soumitra Shukla and Alex Imas describe the Luddites as skilled machine users who challenged how employers used automation.
AI
Manos Zaranis says MF² was accepted at EMNLP 2026 in Budapest and targets comprehension beyond shallow recall.
AI
The poster says they've made 81 bots since August and asks new users to share what they'd build.
AI
A user said they’d spent two days struggling with Codex’s auto-reload and asked what was worth trying instead.
AI
A user says they caught the issue at step seven, when rollouts were almost identical except for hardware nondeterminism.
AI
Two users quote the same post, one writing 'wow' and the other adding three fire emojis.
AI
NVIDIA says Vera Rubin offers Day 0 support for the model, with throughput of up to 200K TPS/GPU.
AI
The announcing post says bots can build and update reports, memos and dashboards, with new cards in every desktop chat.
AI
The poster says another Fable recovered the exact message in a fresh session without being told how it was hidden.
AI
A user imagines the Boring Company building it for electric self-driving chassis, versus roughly 10 mph today.
AI
Stanford says the book connects classical robotics and modern AI in one framework for building autonomous systems.
AI
A coauthor says AI agents could expand what people can do, but who benefits could shape inequality.
AI
One commentator says as few as two major Chinese AI startups may still be growing by late 2027.
AI
A user recalls the Yi family of LLMs fondly; a reply wonders whether 01 AI is still big in China.
AI
SemiAnalysis says Nvidia increased Rubin interactivity by close to 2x in less than eight days using low-latency specialized kernels.
AI
Anthropic says customers get the full Claude Platform feature set with Google Cloud IAM credentials and billing.
AI
A post identifies DeepSeek's V4.1F-1017 and MiniMax's M3.1-Flash; a reply warns that the approval standard could change.
