scripod.com
The $10B Satellite Empire Putting AI in Orbit, Why Chips Beat Rockets & China's #1 Open Model | EP #266

Overview

Shownote

Highlights

Transcript

Chapters

Pins

The $10B Satellite Empire Putting AI in Orbit, Why Chips Beat Rockets & China's #1 Open Model | EP #266

Moonshots with Peter Diamandis

23 HOURS AGO
Moonshots with Peter Diamandis

Moonshots with Peter Diamandis

23 HOURS AGO
OverviewShownoteHighlightsTranscriptChaptersPins

Shownote

This episode is about the collision of Earth intelligence, orbital compute, Chinese open-weight AI, and a new wave of space infrastructure. The big idea is that “large earth models” and “project Suncatcher” are turning space data and space compute into cor...

Highlights

This podcast explores the convergence of Earth observation, orbital computing, and advanced AI, featuring insights from Planet Labs CEO Will Marshall. The discussion centers on how vast satellite data streams and on-orbit processing are creating new 'large Earth models' that can answer physical questions about our planet, while also touching on the competitive landscape of AI models and the future of space infrastructure.
00:00
Large Earth models for indexing the planet
05:30
We have 150 petabytes of data covering every land point daily for 10 years
14:08
Space provides vast real-world data for AI models
25:08
Transparency deters war by reducing misinformation.
33:32
Edge computing reduces data delivery from hours to minutes
48:26
The biggest space revolution is satellite miniaturization.
1:03:11
45% of dementia cases are preventable.
1:11:45
The key to lowering launch costs is rapid reusability and mass manufacturing.
1:23:11
Earth is the best planet.
1:36:50
Current LLMs are brains in a vat.
2:08:59
Technology outpacing social systems poses a cosmic risk.

Chapters

Building a Searchable Index of the Entire Planet with AI
00:00
The Satellite Fleet: 200 Eyes in the Sky Covering Every Land Point Daily
05:30
The Mutual Benefits of Space Data and Artificial Intelligence
14:08
Tokenizing Earth: Envisioning a Planetary-Scale AI to Predict GDP and Land Use
22:35
Real-Time AI in Orbit: How On-Board GPUs Speed Up Disaster Response
33:32
Project Suncatcher: Why Google's TPUs Might Be Better Off in Space
48:26
Cleaning Up Space: How Low Orbits and Ground Lasers Solve the Debris Problem
57:09
The Strategic Pivot: From 3D Printing Rockets to Combining Launch and Data Centers
1:06:00
The Battle for Efficiency: Why Google's TPUs Beat NVIDIA's GPUs for Space AI
1:17:12
The Great AI Talent War: Can Google's Incumbency Hold Off Anthropic's Singularity Believers?
1:28:50
Beyond 'Brains in a Vat': Why Space Data is the Missing Ingredient for True AI Consciousness
1:39:39

Transcript

Peter H. Diamandis: Today, Planet's a $10 billion company. You've coined the term large Earth models. What's that mean? Will Marshall: They, like Google, indexed the internet to make it searchable. We're indexing the Earth to make it searchable. It will f...