Loading activity…
@xwang_lk
Professor @ucsantabarbara. Head of Research @SimularAI. Director @ucsbcrml @UCSB_AI. Build the Science of AI Agents. AI for Humanity in the long run.
AI
DeepSeek's README offers offline inference for DeepSeek-V4.1-Flash but lists additional systems needed for a complete service.
AI
In a quote-post, omarsar0 claims a PINOC MCP server could assemble an animated scene through one Claude Code prompt.
YouTubeAI
In August, the former worker said they had left xAI a while ago and praised the team's persistence through debugging.
AI
A post calls frontier-model use at every step expensive and relying on a small model alone risky.
AI
A post says the benchmark tracks how agents write, maintain, retrieve and use memory across 400 multimodal, multi-session tasks.
AI
Meta says the manuscripts span cardiology, oncology, infectious disease and emergency medicine, with experts retaining scientific control.
AI
A post describes separating what to investigate from how to execute, then learning a transferable value prior.
AI
A group member says the paper was Zhaotian's first project with them and congratulates the UCSB students.
AI
A user says Sai logged into a UCSB biobib website and found the Box folder link in an email thread.
AI
A contributor says the position paper is one of 11 Spotlight selections and is coming soon.
AI
A Simular AI researcher urges safety tests to assess agents’ actions with a mouse and keyboard, not just chat responses.
AI
Scale Labs says humans scored 93.1% across 522 tasks, versus 53.6% for top model GPT-6 Astra.
AI
OpenAI says the Decisions API is now in public beta for all developers. In launch posts, the company said the GPT-6 Luna-powered API accepts text and image inputs, supports predicates, choices, and scores, and can be up to 10x faster than GPT-6 Luna through the Responses API.
AI
A user celebrating COLM 2026 replied to their own post with a photo link labeled “Paper acceptance statistics.”
AI
A researcher says their group’s papers are titled Group-Evolving Agents and Length Value Models.
AI
A researcher says Lenvm is a spotlight paper at COLM 2026’s Efficient Reasoning Workshop on October 9.
AI
Leo Boytsov supports cleaning up the literature, while Susan Zhang questions whether AI can produce more elegant arguments.
AI
A post celebrates the students, calling them the university’s first-ever AI degree cohort.
AI
A user endorses Kevin Buzzard’s proposal to let machines explore math, then begin the journey toward wherever they get stuck.
AI
arXiv says the change aims to distribute moderator time fairly and support staff, volunteers, readers and authors.
