Hard Fork
Meta Shifts the Blame + Do Data Center Bans Work? + The Final HatGPT
2026-08-28 · 1:06:06 · 9 chapters · 8 clips · 3 ad reads · analysis.json
Ranked clips
8 in this episode, in the order they happenEpisode 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 rankingEpisode analysis
Every item carries confidence and evidence · timestamps play in the playerChapters
- NYT Games promo
Juliet introduces New York Times Games and encourages listeners to try the games. - Wine spill banter
Kevin and Casey joke about spilling wine and set a light‑hearted tone. - Meta settlement overview
Kevin and Casey introduce the episode and outline Meta’s $17 billion settlement. - Teen‑safety product changes
Discussion of the specific platform changes Meta must implement for child safety. - Data‑center debate intro
Kevin and Casey pivot to the data‑center backlash and introduce guest Arvind Narayanan. - Data‑center impact deep dive
Arvind explains why banning new data centers will not meaningfully slow AI progress. - Hat GPT finale
The final Hat GPT segment where the hosts draw the last hat and discuss the ending. - Tech news roundup
A rapid series of tech news items: Vision Pro surgery, GTA VI leaks, carrier‑pigeon messaging, AI‑agent fatigue, meat‑proxy term, Tiangong robot sprint, and AGI estimates. - Credits & subscription ads
Closing credits followed by a subscription promotion for the NYT app.
Speakers
- 41%Casey Newton
host · self introduction · 0.98
Evidence - 39%
Kevin Roose
host · self introduction · 0.98
Evidence - 16%
Arvind Narayanan
guest · episode notes · 0.97
Evidence - 2%Jonathan Knight
ad voice · self introduction · 0.96
Evidence - 1%Not identified
ad voice · unknown · 0.94
Evidence
Topics
Data‑center impact on AI progress16%
Meta settlement details8%
AGI progress estimates6%
Teen‑safety product changes5%
Hat GPT final segment5%
GTA VI leak coverage3%
Carrier‑pigeon messaging app3%
Tiangong robot sprint2%
AI agents and founder insomnia2%
Meat‑proxy concept2%
Style
- Format
- Conversation
- Tone
- informal, humorous, analytical
- Pacing
- Moderate · 5.6 turns/min · 187 wpm
- Recurring
- Hat GPT, Tech news roundup, Data‑center segment, Ads
- Confidence
- 0.96
Quality
LoudnessOK-16.8 LUFSPodcast targets sit around −16 to −14
True peakCheck0.4 dBTPAbove 0 dBTP: may clip after encoding
Noise floor-54.3 dBFS
Ad reads3 host read · New York Times Games; host read · New York Times Games; host read · The New York Times
file_22bb6d8d139a · sha256 22bb…17cf · openai/gpt-oss-120b · timestamps belong to this download; ads are inserted per download