AGI is a few years away. World Models are an important step towards it.
RESEARCH
-
Improving AI stability via deterministic RNG and low-rank perturbations
By
–
of course but EGGROLL’s deterministic RNG exactly reconstructs every low-rank perturbation from seeds making evolutionary paths more auditable/replayable than backprop and Could strengthen persistent AI stability
-
MIT’s “Insum” speeds up einsum for sparse datasets
By
–
MIT researchers developed “Insum,” a technique for speeding up computations on datasets replete w/zeros. It rewrites Einstein summation (“einsum”) operations to avoid inefficient handling of zeros, improving memory efficiency & performance: https://
bit.ly/4upJM5s -
Nvidia Announces LongLive-2.0 for Long Video Generation
By
–
Nvidia presents LongLive-2.0
— AK (@_akhaliq) 19 mai 2026
An NVFP4 Parallel Infrastructure for Long Video Generation pic.twitter.com/kGTa0gegb9Nvidia presents LongLive-2.0 An NVFP4 Parallel Infrastructure for Long Generation
-
Researcher joins Anthropic to work on frontier LLM R&D
By
–
Personal update: I've joined Anthropic. I think the next few years at the frontier of LLMs will be especially formative. I am very excited to join the team here and get back to R&D. I remain deeply passionate about education and plan to resume my work on it in time.
-
Researchers propose CodePercept to improve AI on STEM diagrams
By
–
Why do AI models fail at STEM diagrams? Is it weak reasoning or poor perception? Researchers from Shanghai Jiao Tong University and Alibaba’s Qwen Team present CodePercept. Instead of scaling reasoning, they scale perception using code as a precise medium—generating
-
AI reveals humans’ poor complex decision-making skills
By
–
Humans are bad at making complex decisions. #AI can call them out
by Patricia Waldron @TechXplore_com Learn more: https://
bit.ly/3PM4rkV #MachineLearning #ArtificialIntelligence #ML -
ElevenLabs launches Einstein voice agent to bring archives to life
By
–
Voice agents add a new dimension to learning.
— ElevenLabs (@ElevenLabs) 19 mai 2026
Today, we're introducing Albert Einstein's voice to ElevenLabs, and launching an Einstein agent, to bring his written archives to life in his iconic voice. pic.twitter.com/UD57Sfklr7Voice agents add a new dimension to learning. Today, we're introducing Albert Einstein's voice to ElevenLabs, and launching an Einstein agent, to bring his written archives to life in his iconic voice.
-
Research shows more ‘thinking’ time does not always improve AI answers
By
–
The assumption that "more thinking = better answers" was never tested until now. This paper tested it. The answer is no. For operators: if your AI output feels overengineered, bloated, or keeps second-guessing itself, the model isn't being thorough. It's losing confidence.
