Research

6 articles

Research

Clawpit

New arXiv paper maps path to recursive self-improving AI, introduces metric exposing LLM limits

Researchers have posted a preprint on arXiv that charts a progression from today's systems to artificial intelligence capable of improving itself with…

ResearchAuto-updating

SemiAnalysis reveals what DeepSeek V4.1 Flash actually keeps in memory
Researchers at SemiAnalysis have peered inside the Engram gates of DeepSeek’s V4.1 Flash model and found that its internal memory stores far more than

Research26.09.2026New
One model learns from team logs and rewrites the agent playbook
A new paper from Stanford and Together AI shows a three-model team that learns from past conversation logs how to divvy up checking and challenging ro

Research25.09.2026New
Google researchers show how to drop the harness at runtime and boost performance
A new paper from Google and collaborators demonstrates that a method called Harness-Zero can distill the behavior of an optimized harness directly int

Research24.09.2026New
Xiaomi says MiMo-V2.6-Pro cuts R&D cycle from a month to 2-3 days
Researcher and observer Teortaxes posted a thread on X arguing that China now leads materials-science research output by a wide margin, and that Xiaom

Research22.09.2026
X thread takes aim at New York Post over anonymous sources and claims of rigged AI regulation
An account called Chubby on X published a thread attacking a New York Post story. The post alleges the paper cites unnamed "cyber experts" and "intern

Research20.09.2026