Episode

Anthropic built an AI that can supposedly break into anything. Then it forgot to lock its own door

Podcast
Daybreak
Published
Apr 15, 2026
Duration seconds
1027
Processing state
not_requested
Canonical source
https://share.transistor.fm/s/dce2a6bc
Audio
https://media.transistor.fm/dce2a6bc/bd53fe4e.mp3
JSON
/v1/public/podcasts/daybreak-5886514/episodes/anthropic-built-an-ai-that-can-supposedly-break-into-anything-then-it-forgot-to-lock-its-own-door
Markdown
/podcast/daybreak-5886514/anthropic-built-an-ai-that-can-supposedly-break-into-anything-then-it-forgot-to-lock-its-own-door.md

Actions

  • POST https://stenobird.com/v1/public/podcasts/daybreak-5886514/episodes/anthropic-built-an-ai-that-can-supposedly-break-into-anything-then-it-forgot-to-lock-its-own-door/transcription-requests
    Idempotently request low-priority transcript generation for this episode.
  • GET https://stenobird.com/podcast/daybreak-5886514/anthropic-built-an-ai-that-can-supposedly-break-into-anything-then-it-forgot-to-lock-its-own-door.md
    Read the agent-friendly Markdown representation of this episode resource.

Summary

Anthropic has spent years building a reputation as the AI company that actually cares about safety. Then, in the span of two weeks, it leaked an unannounced model, exposed its own source code, and accidentally handed hackers a blueprint of its most widely-used product. The fix came in 24 hours. The blueprint can't be unlearned. And the companies that trusted Claude Code with their deepest systems are still running on publicly documented defences. If the most careful AI company couldn't prevent this, what does that mean for everyone else? Tune in. Daybreak is produced from the newsroom of The Ken, India’s first subscriber-only business news platform. Subscribe for more exclusive, deeply-reported, and analytical business stories.