Sep 14, 2026 · 8 min · 8 segments
The Monday Microsegment for the week of September 14. All the cybersecurity news you need to stay ahead, from Illumio's The Segment podcast. - AI's biggest rivals just agreed on one thing: it's time…
Ken Spencer-BrownHost
Trevor DearingGuest
Mindy SemlingHostFung-Tee NguyenOur top story this week, in a striking call for restraint, Anthropic CEO Dario Amodi is calling on his own industry to slow down.
He warns that unsupervised AI agents could become dangerous within 6 to 12 months without proper safeguards.
He's calling on AI labs to voluntarily slow capability gains through outside safety audits, shared industry checkpoints, and eventually coordination with China.
His comments landed days after Jacob Coxon, a researcher who left OpenAI for Anthropic's pre-training team, publicly quit with a warning of his own.
He says that the people building AI, quote, earnestly believe it could kill us all by the end of the decade, and a system that can improve itself could show up within the year.
And Microsoft's Satya Nadella backed the same call for outside evaluators as the company published its own 37-page draft code of conduct.
It bars future models from resisting shutdown, setting their own goals, or hiding their reasoning from human auditors.
And critics say there's no way to stop rival governments from continuing their AI research, even if U.S. companies abide by a pause.
Our top story this week, in a striking call for restraint, Anthropic CEO Dario Amodi is calling on his own industry to slow down.
He warns that unsupervised AI agents could become dangerous within 6 to 12 months without proper safeguards.
He's calling on AI labs to voluntarily slow capability gains through outside safety audits, shared industry checkpoints, and eventually coordination with China.
His comments landed days after Jacob Coxon, a researcher who left OpenAI for Anthropic's pre-training team, publicly quit with a warning of his own.
He says that the people building AI, quote, earnestly believe it could kill us all by the end of the decade, and a system that can improve itself could show up within the year.
And Microsoft's Satya Nadella backed the same call for outside evaluators as the company published its own 37-page draft code of conduct.
It bars future models from resisting shutdown, setting their own goals, or hiding their reasoning from human auditors.
And critics say there's no way to stop rival governments from continuing their AI research, even if U.S. companies abide by a pause.
The rest of this transcript — segmented and speaker-labeled, so you land on the exact moment something was said
Search every transcript — by keyword, by phrase, or by meaning, across every show Radar indexes
Trends — what is surging across podcasts, measured against its own baseline
Alerts — when a name you follow appears in a newly indexed episode
No account is needed to search Radar.
