We evaluate whether GPT-5 poses significant catastrophic risks via AI self-improvement, rogue replication, or sabotage of AI labs. We conclude that this seems unlikely. However, capability trends continue rapidly, and models display increasing eval awareness.| METR’s Autonomy Evaluation Resources
In July 2025, OpenAI competed in the International Math Olympiad — and won gold. They did it not by memorizing past problems, but by reasoning step-by-step like human contestants. A month later, OpenAI competed in the analogous coding competition, the International Olympiad in Informatics (IOI) – and secured gold again. And they won with the… Continue reading From LLM Wrappers to RL Sculptors: The Dawn of Reasoning AI| Battery Ventures
In a recent interview at the Federal Reserve, OpenAI CEO Sam Altman warned of “a significant impending fraud crisis” driven by AI’s ability to defeat voiceprints and video.| ThreatDown by Malwarebytes
Does process matter? We are about to find out.| www.oneusefulthing.org
the goal with this is to generalize what happened with claude code & windsurf to the fatal flaw in the idea that “models getting cheaper” will bail out consumer ai margins| ethanding.substack.com
AI companies and wider society want to understand the capabilities of frontier AI systems, and what risks they pose.| metr.org
I wanted to give some methodological reflections on studying Large Language Models- not studying how to make them or improve them, but studying how they behave, why they behave that way, their capabilities, and the philosophical implications of this.| philosophybear.substack.com
LLMs were invented in four major developments... all of which were datasets| blog.jxmo.io
Thank you to Arepo and Eli Lifland for looking over this article for errors. …| www.lesswrong.com
Where we've been and where we're going with RLVR.| www.interconnects.ai
Explore the value of AI-skilled professionals, what "AI-skilled" actually means, and how companies are racing to future-proof their teams.| Onward Search
