← The Archive The Wavelength

Issue 252 · August 19, 2026 · 5 min read

OpenAI slows the rollout of its most capable model, and calls it a cyber-risk first

OpenAI says its next model is the first to rate as a critical cybersecurity risk under its own safety rules, and it is taking extra time before wide release. For business owners, the takeaway is that these tools now cut both ways, helping defenders and attackers alike.

Listen to this issue

Inside this issue

The Big Story

OpenAI delays wide release of its next model over cyber risk, and ships a security model aimed at defenders. What happened and why it matters to your business, in plain English.

The Research

3 new research findings, translated into plain terms:

  • AI helper tools can quietly pile up and hurt each other's work
  • A new scheduling method lets AI training systems handle up to 53 percent more work
  • A new signal spots when an AI is confident but wrong

The Discourse

What Ethan Mollick, a Wharton professor who studies AI at work, Gary Marcus, a longtime AI skeptic and psychology professor, Gary Marcus, again on the industry's finances, and Simon Willison, a well-known software developer who writes about AI said on X this week, and where the consensus is forming.

The Rules

3 government moves on AI worth knowing: federal, state, and global.

Read the full issue

The takeaways live in the email: what each story means for your business, what the experts actually said, and where the rules land. One email, every Wednesday. Subscribe and never miss an issue.

Free. One email a week. Unsubscribe anytime.

Already a reader? Join the room on Alani Connect: every full issue searchable, and the community reading alongside you.