// Artifacts as Memory Beyond the Agent Boundary // An agent doesn't always need a bigger memory buffer. Sometimes the environment itself remembers on the agent's behalf. New research formalizes this intuition mathematically for the first time. The work introduces a formal
RESEARCH
-
Is exponential curve growth significance really expected?
By
–
its not significant because its expected on the exponential curce? pic.twitter.com/cyae6JSIPw
— Chubby♨️ (@kimmonismus) 14 avril 2026its not significant because its expected on the exponential curce?
-
Cheaper AI reasoning models secretly cost more reveals study
By
–
Your "cheaper" AI reasoning model is secretly costing you more! A new study by Lingjiao Chen, Chi Zhang, Yeye He, Ion Stoica, Matei Zaharia, James Zou from Stanford University, UC Berkeley, CMU, and Microsoft Research uncovers the "pricing reversal phenomenon" in LLMs. They
-
GPT 5.4 Pro Solves Erdos Problem in Mathematical Breakthrough
By
–
Otro problema Erdos que cae, a manos de GPT 5.4 Pro, resuelto tras 80 minutos trabajando
-
Chinese AI Models Six Months Behind US Frontier Labs
By
–
The question that's currently on my mind is this: Chinese models are about six months behind those of US Frontier Labs. Does this also apply to "Mythos"? Is it foreseeable that, for example, Qwen will release a similarly significant model as Claude "Mythos" in six months, or are
-
A Canticle For Leibowitz: Nuclear Holocaust and Knowledge Preservation
By
–
A Canticle For Leibowitz is a classic early (1959) post-apocalypse novel where an order of monks preserved the last remnants of learning (the memorabilia) after a nuclear exchange turned the remains of society into book and scientist burners. I first read it in the 80s as a mass
-
Analysis of data source selection patterns in AI models
By
–
On a analysé 1,7 million d’URLs consultées par des IA à travers 70 000 conversations, couvrant les principaux acteurs comme @OpenAI , @AnthropicAI , @perplexity_ai , @Google ou encore @xai . Objectif : mieux comprendre comment les modèles d’IA sélectionnent leurs sources et ce
-
P-tau217 biomarker predicts Alzheimer’s disease in older adults
By
–
Another study indicative of predictive power of p-tau217 for Alzheimer's disease in cognitively unimpaired older adults (mean age 71 at baseline) A very low p-tau217 denoted minimal risk https://
nature.com/articles/s4146
7-026-71269-3
… -
Dario’s Optimism Versus Lack of Concrete Scientific Breakthroughs
By
–
In a recent interview Dario was dismayed why he is considered an 'AI doomer' when he considers his AI outlook to be so positive. Then they build a jaw-dropping model, without a single mention of any new attempts at discoveries, scientific applications or health breakthroughs.