Loading activity…
@yuchenj_uw
Cooking fun AI systems & products @databricks. Prev: co-founder & CTO @ Hyperbolic, OctoAI (acquired by @nvidia) Apache TVM, PhD @ University of Washington.
AI
Yuchenj_UW alleges four unavailable access routes, while kylebrussell separately laments Gemini 4 Carbon’s purported nonrelease.
AI
His early-access testing favored one agent for coding projects, but clear goals and human oversight still mattered.
AI
DeepSeek draws praise from two users for understated rollouts and open-sourcing its model-design techniques.
AI
OpenAI says researchers used an upcoming model in an autonomous pipeline, at a total API cost of $1.2 million.
AI
Kimi K3.1-Pro's open-weights announcement links to a Hugging Face URL under moonshotai.
Technology
In a 50-forward-pass test, Astra reportedly changed 0.016% of Qwen3.5-0.8B's weights, raising MATH-500 accuracy from 7.9% to 48.1%.
AI
The reported MATH-500 score rose from 7.9% to 48.1% with the same sampling settings.
AI
HackerTube gathers 6,454 videos that Karpathy says Claude selected from 43 million Hacker News comments.
AI
The datasets hid misaligned behavior in fiction and roleplay, and the deliberately fine-tuned research models were never deployed.
AI
In an October 29 post, thsottiaux said Pro 200 Work and Codex usage would stay unchanged.
AI
A user offered to onboard Snowflake to Unity Gateway and AI products, saying they ship OSS models on Day 0.
AI
A post says Carbon scored 88.2% on SEC-bench Pro and is rolling out first to cybersecurity specialists.
AI
An Anthropic team member says the interviews no longer provide signal and were removed in October 2026.
AI
Anthropic says it acquired Mithril Medicine, whose experimental muscular dystrophy treatment showed promising results in mice.
AI
A user says almost $5 million a day in tokens pass through T3 Code, with 100,000 daily active users projected.
AI
A user quotes a “neolab” friend saying it could resell compute even if it fails to train the best model.
AI
In an October 24 post, Shihipar said he plans a break before starting something new.
AI
Kimi claims a 1-million-token context window for Pro and two to three times Pro's speed for Highspeed.
XAI
Moonshot says both versions are live on Kimi.ai, Kimi Work, Kimi Code and its API.
AI
Two posts point to security incidents and urge AI-assisted defenses; one specifically calls for faster vulnerability detection and patching.
