AI & ML

Subtopics

Subscribe to News

Get the Weekly Future Brief delivered to your inbox.

Latest Updates

An AI Agent Team Improved a Math Bound Built on 1979 Work. A Proof Checker Confirmed It.

A team of OpenAI agents raised a lower bound in combinatorics, and a computer checked the proof step by step. The result is small, but the way it was verified is what makes it worth reading.

Oct 06, 2026 • 5 mins read

AI & ML LLMs & Foundations

Can AI Spot a Cyberattack It Has Never Seen Before? Research on Detecting Zero-Day Attacks

How security teams spot novel zero-day attacks using behavioral provenance graphs and navigate the base-rate fallacy in enterprise operations.

Oct 05, 2026 • 7 mins read

AI & ML LLMs & Foundations

Can AI-Generated Code Be Trusted? Research on Security Vulnerabilities in Code Written by AI Models

Empirical audits reveal that up to 40% of LLM-generated code contains critical vulnerabilities, while developer overconfidence and package hallucinations expose enterprise software supply chains.

Oct 05, 2026 • 7 mins read

AI & ML LLMs & Foundations

The Harness Is Doing the Work: How Small Open Models Climbed ARC-AGI-3

Top scores in a Kaggle contest on AI reasoning rose from about 1% to nearly 56%. Much of the gain appears to come from the code around small open models, but the numbers need a careful read.

Oct 05, 2026 • 5 mins read

AI & ML Enterprise AI

FDA-Cleared AI Medical Devices: Cleared, Not Proven. What the FDA's List Doesn't Tell You

Of 1,357 FDA-cleared AI medical devices, only 3 were evaluated for patient outcomes such as death or readmission. Here is what that does and does not mean.

Oct 05, 2026 • 6 mins read

AI & ML LLMs & Foundations

Model Distillation: How Do You Steal an AI's Reasoning? Inside OpenAI's Accusation Against Moonshot

Model distillation is normal in AI. OpenAI says this campaign crossed the line by extracting hidden reasoning. Here is what is claimed and what is confirmed.

Oct 04, 2026 • 5 mins read

Page 3 of 15 (Total 85 articles)