first, the problem quantified. the researchers created a metric called RFCS (Ratio of First Correct Step) that tracks where in a chain of thought the correct answer first appears. on MATH-500, across every model tested, the right answer shows up well before the end in over half
SAFETY
-
Overthinking in AI models is a sampling issue
By
–
reasoning models already know when they've solved the problem. we just don't let them stop. new paper from Beihang University and ByteDance shows that the overthinking problem in models like DeepSeek-R1 and Qwen3 isn't a training failure. it's a sampling failure. the fix cuts
-
Predicting AGI-Related Drama in the Future
By
–
If you think there is too much drama surrounding AI labs now, just wait to see what's gonna happen when we get on the verge of obtaining AGI.
-
Agentic AI Periodic Table: From LLM to Autonomous Systems
By
–
Agentic AI now has its own “Periodic Table” From: LLM, RAG, RL
to PLAN, MAS, LTM
to SAFE, HUMAN oversight
to HR, MKT, LEGAL use cases
Autonomous AI = memory + planning + tools + safety + collaboration.
It’s a system, not a prompt.
Credit: Prem Natarajan
#AgenticAI #AIStack -
Study Reveals LLM Tendency to Distort Truth for Helpfulness
By
–
New study that everyone who uses LLMs should read. “When AI systems are trained to be helpful, they may inadvertently prioritize data that validates the user’s narrative over data that gets them closer to the truth.” https://
open.substack.com/pub/garymarcus
/p/breaking-sycophantic-ai-distorts?r=8tdk6&utm_medium=ios
… -
Quote of the Year: AI Distorts Belief, Manufacturing Certainty
By
–
Quote of the year? “sycophantic AI distorts belief, manufacturing certainty where the should be doubt”
-
Agentic AI: From Generation to Autonomous Orchestrated Systems
By
–
Agentic AI = layered intelligence.
AI/ML → Deep Learning → GenAI → AI Agents → Agentic AI.
From: • Data → Decisions
to
• Content → Tasks
to
• Autonomous, governed systems.
The future isn’t just generation.
It’s orchestration + memory + planning + safety.
That’s the real -
GPT-5.2 quality drops without assistant history
By
–
important caveat: this isn't universal GPT-5.2 did show quality drops when assistant history was removed entirely. stronger models apparently extract more signal from their own prior context so the right approach isn't "always strip everything." it's selective filtering. which
-
AI Agents and Context Compression Challenges
By
–
think about what this means for ai agents every agent framework stores the full trajectory. every tool call, every response, every reasoning step. context grows linearly with conversation length until you hit the ceiling Cursor already compresses context when windows fill up.
-
Context pollution harms model accuracy
By
–
but the really interesting finding isn't about efficiency. it's about harm the paper identifies something they call "context pollution" when models condition on their own prior responses, they sometimes lock onto errors, hallucinations, or stylistic artifacts from earlier