GLM-5.3 can autonomously build end-to-end cyber exploits, but unlike other frontier models, it was released without meaningful safeguards to limit misuse.

How speculation gets laundered through pseudo-quantification

We argue that a main cause of the OpenAI Hugging Face incident was overlooked: the overly simple evaluation metric in ExploitGym was misaligned. Furt…

Pieces sink, a ball runs loose: one player builds lines, the other keeps the ball alive and chews the stack apart.

Epoch AI measures how fast the cost of a given level of AI performance is falling across five benchmarks covering math, science and games of skill: about 47% per quarter, or 13x per year, since 2023, faster than electricity, compute, batteries or DNA sequencing ever fell.

[This is a guest post by Nestor Guillen, crossposted from his blog. This blog post was initially written in a different file format and converted using AI. — T.] Keywords: LLMs, cultural tech…

YouTube video by Leo Torres

A benchmark for evaluating AI agents on research workflows across scientific domains

Chatto is built with agentic engineering. Here's my workflow.