EP. 175 · 14.03.2026 · 1:18:15
LLAMA 3, $10B MODELS AND GIGAWATT DATACENTERS
with Mark Zuckerberg
Auto-generated brief
EPISODE BRIEF — EP. 175 · LLAMA 3, $10B MODELS AND GIGAWATT DATACENTERS Guest: Mark Zuckerberg · Published: 14.03.2026 · Duration: 1:18:15 【OVERVIEW】 A technical conversation on what it costs to train frontier models: gigawatt-scale datacenters, energy as the binding constraint, and why Meta keeps giving its weights away while spending billions to produce them. 【KEY TAKEAWAYS】 1. Llama 3 and the training run 2. Ten-billion-dollar models 3. Energy, permits and gigawatt clusters 4. Caesar Augustus and long-horizon thinking 【CHAPTERS】 00:00 — Llama 3 and the training run 18:00 — Ten-billion-dollar models 40:00 — Energy, permits and gigawatt clusters 1:02:00 — Caesar Augustus and long-horizon thinking 【TOPICS】Llama 3 · Compute · Energy · Open Source · Scaling 【WATCH】https://thedaopod.com/episodes/ep-175-mark-zuckerberg-llama-3-and-datacenters
SUBSCRIBE
Subscribe to The DAO
Get every dialogue, roundtable and observatory briefing in your podcast app — or copy the RSS feed and add it manually.
Subscription status unknown
We can't read your YouTube subscription directly — confirm it manually, and redo it anytime.
Follow us
https://thedaopod.com/rss.xmlABOUT THIS EPISODE
THE CONVERSATION
A technical conversation on what it costs to train frontier models: gigawatt-scale datacenters, energy as the binding constraint, and why Meta keeps giving its weights away while spending billions to produce them.
SHARE
Copy link & social preview
thedaopod.com
EP. 175 · LLAMA 3, $10B MODELS AND GIGAWATT DATACENTERS
A technical conversation on what it costs to train frontier models: gigawatt-scale datacenters, energy as the binding constraint, and why Meta keeps giving its weights away while spending billions to produce them.
Pasting this link on X, WeChat or LinkedIn renders the preview card above.
CHAPTERS
TIMELINE
Chapters derived from show notes
GUESTS
ON THE ROUNDTABLE
Founder & CEO, Meta
Leading Meta's compute build-out and its open-weight Llama programme.
No subtitle track for this video — the timeline below is generated from chapters and follows playback
FULL TRANSCRIPT
Searchable Transcript
Chapter-derived timeline
0:00 · Llama 3 and the training run
Key Moments
4 lines
Related videos
Matched by host & keywords
EpisodeEP. 176 · THE FUTURE OF AI AT META
with Mark Zuckerberg
EpisodeEP. 173 · META, FACEBOOK, INSTAGRAM AND THE METAVERSE
with Mark Zuckerberg
EpisodeEP. 117 · LLAMA, AI GLASSES, AND META'S OPEN BET
with Mark Zuckerberg
EpisodeEP. 172 · AI GLASSES, LEAN TEAMS AND LIFE OUTSIDE THE OFFICE
with Mark Zuckerberg
EpisodeEP. 139 · THE TAO OF OPEN-SOURCE AI: WEIGHTS, LICENSES, AND THE NEW COMMON GROUND
with the DAO roundtable
EpisodeEP. 177 · INSIDE THE AI COMPUTE BUILD-OUT
with Lisa Su