Stanford just turned the Meta-Harness paper into open source code. It's a framework that automatically optimizes the scaffolding around a fixed base model. Think memory, retrieval, and context decisions. The approach is unusually direct. A coding agent acts as the proposer.
@alphasignalai
-
Game Theory Study Proves AI Layoffs Harm Economy
By
–
Researchers just mathematically proved AI layoffs will destroy the economy. Using game theory, two economists from UPenn and Boston University mathematically proved something uncomfortable. Every company replacing staff with AI is also firing its own customers. Laid-off
-
Kimi Proposes Solution to LLM Prefill-Decode KV Cache Transfer Problem
By
–
A new paper just exposed how much money AI labs waste on GPUs. Running LLMs at scale hits a wall when prefill and decode share one datacenter. The KV cache transfer between them is massive. This forces expensive RDMA networks and identical hardware everywhere. Kimi proposes
-
DeepSeek V4 Launch: 1.6T MoE Model With 1M Context Window
By
–
While everyone watched GPT-5.5 launch, DeepSeek quietly shipped V4 the next morning. V4-Pro: 1.6T total / 49B active, MIT license.
V4-Flash: 284B total / 13B active.
Both with native 1M-token context. At 1M tokens, V4-Pro runs at 27% of V3.2's FLOPs and 10% of the KV cache. -
Sim2Reason Trains LLMs on Physics Without Human Annotation
By
–
AI can now learn physics the way Newton did — by experiencing it.
— AlphaSignal AI (@AlphaSignalAI) 24 avril 2026
Training LLMs on physics problems hits a wall fast.
Human-labeled question-answer data is scarce and narrow.
Less than 2% of DeepSeek-R1's training pairs touch STEM.
Sim2Reason skips annotation entirely.… pic.twitter.com/SSPs33du51AI can now learn physics the way Newton did — by experiencing it. Training LLMs on physics problems hits a wall fast. Human-labeled question-answer data is scarce and narrow. Less than 2% of DeepSeek-R1's training pairs touch STEM. Sim2Reason skips annotation entirely.
-
AiScientist Runs Autonomous ML Research for Hours or Days
By
–
The best AI research agent doesn't think harder — it just never forgets. A new paper introduces AiScientist, a system that runs ML research autonomously for hours or days. Setup, coding, experiments, debugging. The full loop, unattended. The core idea is simple: Instead of
-
Stanford Paper on Legal RAG Hallucinations in AI
By
–
Paper: https://
law.stanford.edu/wp-content/upl
oads/2024/05/Legal_RAG_Hallucinations.pdf
… Check out http://
AlphaSignal.ai to get a daily summary of top models, repos, and papers in AI. Read by 280,000+ devs. -
Stanford Study Exposes Hallucinations in Legal AI Platforms
By
–
Stanford proved "hallucination-free" legal AI is a marketing lie. Researchers ran 202 legal queries through LexisNexis and Thomson Reuters' legal AI tool platforms and compared them against GPT-4. Every response was manually reviewed by legal experts. Lexis+ AI hallucinated
-
AlphaSignal.ai announces paper 2604.15034 and daily AI summary
By
–
Paper: https://
arxiv.org/abs/2604.15034 Check out http://
AlphaSignal.ai to get a daily summary of top models, repos, and papers in AI. Read by 280,000+ devs. -
Autogenesis: AI agents rewrite themselves without human help
By
–
AI agents can now rewrite themselves without human help. Most AI agents stop improving the moment they ship. New tools arrive, environments shift, and the agent stays frozen in time. ' Retraining is expensive and human patches are brittle. A new paper called Autogenesis