Skip to content
Tue, Oct 20: Microsoft says frontier models can't be trusted episode artwork

Oct 20, 2026 10 min

Tue, Oct 20: Microsoft says frontier models can't be trusted

from Today's AI News

Microsoft ships Guardian Agents and Defender agent detection while warning frontier models can't be trusted, OpenAI claims big GPT-6 gains on COBOL and Fortran with a new MultiLangBench, and OpenAI races to raise $30B ahead of Anthropic's IPO. Plus MiniMax M3.1 Flash, Inception's Mercury 3, a DHS DeepSeek ban, Claude Code 2.1.304, an AI-agent hacking campaign, a Canadian publishers' suit against Anthropic and Meta's Prometheus cluster.

Episode details come from the podcast's RSS feed.

Embed this episode

Ready to play

Tue, Oct 20: Microsoft says frontier models can't be trusted

0:00 10:07

No transcript yet.

Requests help us decide what to transcribe next.

Frequently asked questions

How long is this episode of Today's AI News?

This episode is 10 minutes long.

When was this Today's AI News episode published?

This episode was published on October 20, 2026.

Can I download this Today's AI News episode?

Yes. Use the download control on the episode player to save the publisher-provided media file.