Academic papers, technical breakthroughs, and scientific progress in artificial intelligence.
AI Weekly Issue #537: AI models may soon copy themselves onto smart fridges, Anthropic says
Anthropic's Frontier Red Team says small AI models may be months away from being able to copy themselves onto internet-connected devices such as smart…
Experts adopt AI agent's random opinions and think they're their own, Anthropic study finds
Anthropic ran an experiment with 50 experts in causal inference. Each worked on a data analysis task about a fictional farming cooperative and its vol…
The PHP Incident
Wiz Research reported two remote code execution vulnerabilities in PHP, called PHP2Shell, to the PHP team on August 24. The PHP team began working on …
AI Weekly Issue #523: AI ethics is nobody's job now. The labs prefer it that way.
# AI Ethics Roles Disappearing at Major AI Labs as Companies Restructure Four frontier AI laboratories removed ethics personnel and oversight structu…
Nous Research confirms it hit $1.5B valuation, launches AI agents for business users
Nous Research, the AI company behind the open-source Hermes Agent model, has raised $90 million in Series B funding, confirming a $1.5 billion valuati…
How to Build Reliable AI Agent Systems for Production
# How to Build Reliable AI Agent Systems for Production AI agents operating in production environments face critical reliability challenges that can …
EmTech Future 2026: When AI Meets Everything
Google Research VP Yossi Matias argues that artificial intelligence's most transformative effects emerge at the intersections of AI with established f…
The Download: a biological de-aging contest and why LLMs don’t reason
MIT Technology Review reports on two emerging developments in AI and biotech this week. A new biological de-aging contest challenges competitors to r…
When can we say AI made a scientific discovery?
# When Can We Say AI Made a Scientific Discovery? Anthropic launched a molecular biology lab earlier this year where Claude AI agents operate alongsi…
Not About Navier-Stokes
# AI's Approach to Mathematical Problems Reveals Deeper Shifts in How We Solve Hard Science OpenAI's recent work applying AI to the Navier-Stokes pro…
AI agents do more of the work in model development, but humans still make the decisions
# AI Agents Handle More Model Development Work, But Humans Retain Decision-Making Control Researchers examining the actual mechanics of AI-assisted m…
Insurers claim AI is already increasing healthcare costs
Blue Cross Blue Shield released data showing that hospitals using artificial intelligence tools drove an additional $942 million in healthcare spendin…
AI access makes people almost entirely unwilling to say "I don't know," study finds
A new study reveals a troubling behavioral shift when people gain access to AI assistants. Researchers found that having an AI readily available nearl…
Nvidia's SoL-Pi system cuts coding agent token usage nearly in half by optimizing the harness
Nvidia researchers have developed SoL-Pi, a system that reduces token consumption in coding agents by nearly 50 percent without sacrificing performanc…
Intelligence doesn't come cheap as AI drives up costs for the NSA, hospitals, and insurers
# Intelligence Doesn't Come Cheap as AI Drives Up Costs Across U.S. Government and Healthcare The computational appetite of advanced AI models is for…
Young organs may not be a fountain of youth for recipients
A viral moment at last year's aging conference hinted at an emerging obsession among world leaders: the possibility that biotechnology might eventuall…
Top AI experts badly underestimated how fast the field is moving, study finds
# Top AI Experts Badly Underestimated How Fast the Field Is Moving, Study Finds The Forecasting Research Institute has documented a troubling pattern…
AI performance costs are falling faster than those of any previous technology
AI performance costs are plummeting at rates never seen in computing history. Epoch AI, a research organization tracking machine learning trends, meas…
The Download: why AI’s latest breakthroughs and fears may be more hype than reality
# The Download: Why AI's Latest Breakthroughs and Fears May Be More Hype Than Reality The tech industry enters each summer cycle with predictable fer…
The Post-training Process OpenAI Used for ChatGPT
# The Post-training Process OpenAI Used for ChatGPT OpenAI's approach to post-training represents a deliberate engineering choice that shapes how mod…
OpenAI forms math advisory group as its AI resolves more than 100 open problems
OpenAI has assembled a mathematics advisory group to guide its AI research as the company's models solve increasingly complex mathematical problems. T…
Research context
Academic papers, technical breakthroughs, and scientific progress in artificial intelligence.