Momentize Hear today’s moments
Hard Fork

OpenAI’s Two-Week Pause + Jill Lepore on the Threat of the “Artificial State” + Train of Thought

2026-08-21 · 1:06:17 · 5 chapters · 7 clips · 5 ad reads · analysis.json

Ranked clips

7 in this episode, in the order they happen

Episode timelineClick to play from there · a numbered clip plays from its start · pinch or ⌘-scroll to zoom, drag to move 1×
Loudness
Chapters
Speakers
Ads
Clips

Loading the episode timeline…

In this moment

    Clip detail

    In, out, and ranking

    Episode analysis

    Every item carries confidence and evidence · timestamps play in the player
    Chapters
    1. Games promo intro
      Juliet introduces New York Times Games and a short social‑media style chat.
    2. OpenAI pause & AI safety
      Kevin and Casey discuss OpenAI’s two‑week training pause, the Astra model, and new safety safeguards.
    3. Interview with Jill Lepore
      Kevin and Casey talk with historian Jill Lepore about her book on the artificial state and its political implications.
    4. Train of Thought: Spirit Airlines data
      The hosts explore Google’s purchase of Spirit Airlines data, data‑market dynamics, and reinforcement‑learning environments.
    5. Closing credits & sponsor messages
      Credits, production acknowledgments, and final New York Times Games sponsorship spots.
    Speakers
    1. 46%Kevin Roose
      host · self introduction · 0.99
      Evidence
    2. 32%Casey Newton
      co host · self introduction · 0.99
      Evidence
    3. 20%Jill Lepore
      guest · named by host · 0.98
      Evidence
    4. 2%Not identified
      ad voice · unknown · 0.85
      Evidence
    Topics
    Jill Lepore’s artificial state concept36%
    OpenAI training pause and Astra model safety15%
    Data markets, reinforcement‑learning environments, and AI training data11%
    Google acquisition of Spirit Airlines data9%
    New York Times Games sponsorship ads4%
    30‑minute rule and AI safety frameworks3%
    Style
    Format
    Interview
    Tone
    analytical and conversational with occasional humor
    Pacing
    Moderate · 5.4 turns/min · 187 wpm
    Recurring
    Train of Thought, New York Times Games ads, Credits
    Confidence
    0.96
    Quality
    LoudnessOK-16.7 LUFSPodcast targets sit around −16 to −14
    True peakCheck0.4 dBTPAbove 0 dBTP: may clip after encoding
    Noise floor-54.5 dBFS
    Ad reads5 host read · New York Times Games; host read · New York Times Games; host read · New York Times Games; host read · New York Times Games; host read · New York Times Games

    file_6a6e8a7b96fe · sha256 6a6e…d218 · openai/gpt-oss-120b · timestamps belong to this download; ads are inserted per download

    0:00