Every idea, kept. Take one.

A visual interpretation of what connects across episodes

pull what was said on tape (claims, numbers, people, positions) and show how it links across five, ten, fifty episodes: who said what first, what came back, what changed. The vault graph is the seed; this is the version built from the tape itself

Run the NASA–IBM Lunar Foundation Model ourselves

nasa-ibm-ai4science/NASA-IBM-Lunar-Foundation-Model

INDUS-SDE as a research engine

NASA's science-discovery language model over 600K+ documents (nasa-impact/indus-sde-v0.2)

A LEIA entry, scoped small

one of the five areas under a US-led team

The agentic curriculum

model-based systems engineering applied to the Krav syllabus @1:07:37. "Why isn't my curriculum agentic in the sense that everyone can consult it, can talk with it, can view it? … The same way you'd introduce a change to a car — where are the parts that it touches?"

Libraries as AI centres

@1:22:31 — refurbish public libraries into AI education places. He proposed it independently; the library is already in CNN's list of what Quincy's data-centre tax revenue built

Reverse BOM from SEC filings

@0:00 — build the bill of materials, then run it backwards off mandatory filings "because filings that are due to scrutiny" beat social media

Biomechanically correct video

@24:25 — "a digital twin of your ability to move"; centre of mass, feet, arms

Open-weight models so the data stays yours

"Do I continue using the tools that I have, or do I go and search for open-source, open-weight models that do not take my data?"

Client-matching engine for AJC

built from the dating-matchmaking segment @1:10:16 — same market-awareness approach, pointed at helping clients find the right person to connect with ⚠️(see README)

Agent swarm for a one-person content pipeline

Emad runs 18 Grokbots in parallel @22:58

Route cheap Chinese open models for the bulk work

GLM flash at ~10× cheaper for the same performance @36:28

"The humans became the bottleneck"

the WSJ 10-day-work-week finding @1:20:02 is his own review-debt gate described from the abundance side

Buy-the-cheap-stack vs build-your-own

from Waymo OEMing Chinese hardware @1:38:36

Track the administrator, not the news

@NASAAdmin states policy weeks before it becomes a solicitation with a deadline

The orbital data centre demand chain, drawn

NASA prizes as the product's front door

"What the taxpayer already bought"

the Perseverance spare at JPL, ~$0.5B, idle

Embodied AI for martial arts

what sensors does a model need to understand movement

Taxonomy of fighting

Planet's ~3,000 Earth layers prompted "what's the organisation of fighting?"

Local vs frontier model router

a written rule for which model does which job

Fire-spread app

for departments and municipalities

Commodity / container tracker

from satellite imagery

Data-centre location prediction

as a service

3D printing for space, and in space

Dyson swarm explainer

"what does a Dyson swarm mean for a day-to-day person?"

Operator pages

Eric Schmidt, Elon: what they know about organisations, for operators

Build your own ORAN

an index on how AI labs spend capital

Alternative launch architectures

elevators, spin launch, air-launch

Image-provenance ledger

"how do we know the image is true?"

Satellite architecture comparison

Doves and Pelicans vs Starlink

The three space KPIs

launch cost/kg, bytes-per-kg, flops-per-watt, charted over time

Property valuation from satellite history

the "EV toll", where to buy

Crop / agriculture 10×

via his contact at the Portuguese Agricultural Association

A contest on Planet Labs data

"what can we do with this?"

Interview Will Marshall from the small-business chair

"how can Planet Labs help people building small businesses?"

Project Suncatcher deep-dive

"this deserves a whole episode, nobody explains it in detail"

Predictive "where to live" model

schools, streets, weather, population, forward-looking