AI
New tools, big breakthroughs, and what artificial intelligence means for you.
New tools, big breakthroughs, and what artificial intelligence means for you.
New tools, big breakthroughs, and what artificial intelligence means for you.
Discussing a claimed FAA grounding, @robertskmiles praises aviation’s willingness to stop when systems behave unexpectedly.
Fradin disputes Fortune’s headline, saying robotics research joined the main team to improve future core models.
SonglinYang4 says the team landed NVFP4 in pretraining under the code name “yolo bundle.”
The method nearly matches on-policy self-distillation accuracy in reported tests, but faster training comes with a two-GPU catch.
Anthropic’s CEO says stronger growth could require pricier short-term computing capacity despite a reported $900 billion infrastructure spending plan.
OpenAI says ChatGPT Wallet is starting a limited US rollout for adult Pro subscribers, with approval required for every purchase.
ggml-org is committing to major-version bumps for incompatible library changes, not promising that breaking changes will disappear.
Ricursive says Ripple-65B Preview’s $32 million development bill includes experiments, debugging and final training—not just one big run.
The CapitalG-led Series C values LangChain at $3 billion as it pitches companies more control over agents through LangSmith.
NVIDIA’s SGLang Summit posts cite 60+ and 100+ “day-zero model launches,” with no explanation in their text.
@rohanpaul_ai says Vexo sends spoken tasks and conversation context to Claude, with phone approval before building.
Google says its U.S. rollout offers opt-in spending analysis and budget suggestions, plus optional financial-account linking through Plaid.
@arankomatsuzaki reports Drex Max leads Jev on the AA index, with scores of 72.6 and 66.2.
DeepSeek says developers can add search, files, links, safety and voice features to their own apps.
Liberty Times links the alleged group to anti-data-center messaging and efforts to ease chip export controls.
Sequoia put out a November 13 last call for its November 15 San Francisco poker fundraiser benefiting Khan Academy.
philhchen says Fable 5.2 lifts Claude Tag above agent rivals, while the same model’s API setup still scores higher.
After Cerebral Valley, @gokulr revisited a criticism he endorses: too many AI startups sell “picks and shovels.”
Google’s speed-focused model rolls out across Search, Gemini and developer tools, with free app access routed through automatic selection.
Arena reports visual-design net improvement scores of 45.34% for Fable versus 23.29% for Opus, both Max.