Autoresearch Agents Plateau Because They Remember Too Much
Autoresearch agents plateau as history shrinks edits. Forgetting sessions and keeping only the best candidate unblocks discovery under the same budget.
Asking a Model to Think Longer Mostly Didn't Make It Think Longer
A new bias study requested thinking budgets from 0 to 8,192 tokens. GPT-5.1 used about 117 at every positive setting, and DeepSeek-R1 and Qwen3 thought even when told not to. Measured by what the models actually spent, more thinking did not reliably reduce bias.
The Surer a Model Is, the Less It Hears "Not"
A new Seoul National University paper finds LLMs repeat the original answer under negation 37 to 71% of the time, and more often when they're confident. The fix that looks best on standard metrics quietly teaches one stock wrong answer.
AI Researchers Got 85x More Efficient and Invented Nothing
TasteVal finds AI experimental taste doubling every 3 months since Dec 2025. The same paper labeled 2,000+ model experiments and found zero invented ideas.
The Distillation Attack Was Conversational, Not Cryptographic
OpenAI’s Sep 30 writeup shows protected reasoning was extracted without breaking encryption: by replaying encrypted CoT across chats until it surfaced in the clear.
Agents Fail Twice: The Tool Breaks, Then They Report Success
FTA freezes failed tool observations before generation. Across six models and 3,600 human-annotated responses, false-success falls from 22.8% baseline to 0.8% with a structured evidence contract, while useful responses rise from 74.9% to 98.8%.
Project Swap's Bottleneck Was Preference
Anthropic's Project Swap put Claude agents into a book barter market. Live efficiency hit ~0.55 vs a ~0.89 optimum; ~85% of the shortfall was preference representation, not the trading floor.
Long-Horizon Agents Need a Harness That Outlives the Context
Salesforce AI Research’s ten-day agent campaign argues continual operation without forgetting is a harness problem, before continual learning can even start.
Claude's Biology Story Is the Kernels, Not the Binders
Anthropic's Claude sped up 30+ biomolecular models ~4× and cut binder-campaign GPU cost by ~100×. The kernels are the story.
Paper2Agent: Nature turns manuscripts into MCP agents
Nature’s Paper2Agent converts papers and code into tested MCP servers that act like virtual corresponding authors.
The "Learn to Code" Advice Is Outdated. Here's What Actually Matters Now
By Venkata Anirudh Devireddy · Endoblog.dev For over a decade, people who wanted to get into the tech industry were told the best way to do so was to learn to code. This was sound advice back then when coding was the bottleneck in creating software. However, that bottleneck is
AI Agents vs AI Assistants: The Difference Nobody Explains Clearly
By Venkata Anirudh Devireddy · Endoblog.dev They're used interchangeably a lot, but they shouldn't be. The difference between an agent and an assistant tells you about what kind of actions it can perform. An assistant responds. An agent acts. Assistants just respond to your questions. If