Notas del episodio

What happens when an AI starts hacking itself? Today’s episode covers Claude’s jaw-dropping security flaws — and how it helped researchers break its own defenses. Also: OpenAI just released its first free, customizable AI models in 5 years. And they’re shockingly close to GPT-4 performance.

We’ll talk about:

  • How Claude accidentally revealed two major vulnerabilities in Claude Code
  • The wild competitions Claude is winning against human hackers
  • OpenAI’s launch of gpt-oss-120b and gpt-oss-20b — and what makes them different
  • Why Sam Altman says they were “on the wrong side of history” on open models

Keywords: Claude Code, OpenAI gpt-oss, AI security, CVE-2025-54794, gpt-oss-120b, hacking AI, agentic models, AI competition

Links:

 ...  Leer más
Palabras clave
Claude CodeAI securityOpenAI gpt-ossgpt-oss-120bAI competitionagentic modelshacking AI