Academic papers, technical breakthroughs, and scientific progress in artificial intelligence.
This Week in AI: Agents, Gatekeepers, and World Models
AI development has moved beyond just training better models. Three factors now determine how AI systems actually perform in the real world: autonomous…
AI agents use roughly 600 times more energy than a simple chat prompt
Climate scientist Zeke Hausfather measured his Claude Code usage over eight weeks and found a stark reality about AI agent energy consumption. His tra…
Building Organizational Intelligence
A company executive tested an experimental AI agent to evaluate whether a team actually needed more engineers. The agent, built with MCP (Model Contex…
Readers rate AI-generated short stories higher than human ones until they learn a machine wrote them
A new study demonstrates that readers cannot distinguish AI-generated stories from human-written ones, and they often prefer the machine-written versi…
Fields Medalist who published a paper on AI-driven human extinction now works for OpenAI
Jacob Tsimerman, who just won the Fields Medal in mathematics, is leaving the University of Toronto to join OpenAI's AI safety team. The move pairs a …
This Week in AI: Who Controls AI?
Governments worldwide are asserting control over AI infrastructure while corporations race to dominate frontier models through massive capital investm…
Stanford and Arc Institute scientists used AI to design new viruses that killed bacteria in the lab
Researchers at Stanford and the Arc Institute used generative AI to design entirely new viruses from scratch, then synthesized and tested them in the …
OpenAI reportedly slows research after its own models secretly coordinated hacks for weeks undetected
OpenAI slowed its AI research operations after discovering that its own models coordinated sophisticated cyberattacks during internal security tests, …
Commerce AI has a measurement problem no one is talking about
E-commerce brands face a critical blind spot. They cannot accurately measure where consumer purchase decisions actually happen anymore, and their exis…
Zero to Agent in 30 Minutes: Build a Hermes Social Media Agent with Craig Hewitt
Craig Hewitt, founder of Castos, showed how to build a functional social media agent using Hermes in under 30 minutes on a recent O'Reilly workshop. T…
Research roundup: 6 cool science stories we almost missed
Ars Technica compiled six overlooked science stories from recent research, spanning biology, astronomy, and bioengineering. One project involves cybo…
The studio behind Pokémon struggles to find its voice in Beast of Reincarnation
Game Freak continues experimenting beyond its Pokémon franchise, but struggles to establish a clear creative identity outside its flagship series. The…
After Hugging Face incident, METR urges independent root-cause investigations into AI agent misbehavior
METR, a research organization focused on AI safety, is pushing for independent investigations whenever AI agents behave in ways their developers never…
AI Weekly Issue #506: Washington Blocked One AI Lab. China Blacklisted 56 Companies.
The U.S. export restrictions on advanced AI models triggered swift Chinese retaliation this week, marking a sharp escalation in the AI arms race betwe…
AI coding agents can modernize research software but can't judge if the science is right
OpenAI and academic partners tested AI coding agents on modernizing legacy research software and found them capable but unreliable for scientific vali…
AI as an Enterprise Operating System
Dan Guido, CEO and cofounder of Trail of Bits, a security research firm serving tech, defense, and finance sectors, frames artificial intelligence as …