Andrej Karpathy has joined Anthropic and returned to LLM R&D.
SuperGrok and X Premium subscribers can now use Grok models on OpenClaw.
Gemini 3.5 Flash has been added to Vercel AI Gateway, with medium thinking set as the default.
Vercel has launched a Flat Rate CDN limited beta for Pro teams.
It introduced a way to organize repetitive instructions in Lovable into reusable Skills.
Manus has unveiled Scheduled Tasks 2.0, allowing recurring tasks to continue within the same context.
LLM Wiki v2 grows into an operational knowledge base by adding lifecycle management, graphs, and automation.
A California jury dismissed all claims in Musk's lawsuit against Sam Altman.
During pretraining, LMs repeatedly oscillated between pattern imitation and generalization.
HRM-Text has released a method for pretraining a 1B model for about $1000.
Qwen3.5-9B's political censorship was found to be a small circuit that decomposes into 3 axes.
Qwen3.7 Preview launched on Arena's Text and Vision categories, ranking 13th and 16th respectively.
Hugging Face released 6 Ettin-based rerankers along with the training recipe.
An Indic corpus spanning 11 languages and 8.4B tokens has been released on Hugging Face.
Claude Managed Agents can now run on Vercel Sandbox.
Aggressive investment by China's CXMT to expand memory production is expected to cause RAM prices to fall in the second half of 2027.
At Dell Technologies World, Dell and NVIDIA unveiled new Vera Rubin-based AI factory products and an on-premises agent strategy.
MLRC 2026 submissions have opened, and this year it will run as an official NeurIPS 2026 track.
NVIDIA has delivered its first Vera CPUs for agentic AI to major AI labs and OCI.
In a 6.9B LLM experiment, AI discourse increased misalignment while aligned discourse reduced it.
Grok Skills has launched on web, iOS, and Android, supporting document automation and personalized workflows.
Vercel now supports a Consolidated Commit Status on GitHub for monorepos.
Vercel has exempted traffic blocked by WAF from CDN and transfer billing.
Former OpenAI staffer Leopold Aschenbrenner is backing a strategy that leverages crypto mining infrastructure to secure AI infrastructure.