Alex Wissner-Gross - ASI PILL EP235 - Anthropic and the Pentagon
Alex breaks down EP235: AGI redefined as a balance sheet entry, the safety-through-competition thesis, Anthropic vs Pentagon, OpenClaw and AI agents, hyper-deflation killing knowledge work, model distillation to phone-size, Google's unified architecture, the Amazon-OpenAI mega-deal, single-person AI conglomerates, meat puppets in fast food, VLA robots, solar abundance, Meta as cloud provider, prime editing DNA, GLP-1 anti-aging, humanoid robots, Dyson swarms, and why CAPTCHAs still exist.
0:00 AGI as a dollar amount
0:28 The circular economy goes real
2:12 Safety through competition
2:57 It takes all of humanity to align AI
4:08 Anthropic and the Pentagon
4:48 Frontier models on the geopolitical stage
5:58 What AI agents are we actually getting?
7:17 OpenClaw and multi-model products
8:18 Can infra handle always-on agents?
8:57 Generation cost hits zero
9:27 Verification is the new scarce good
11:21 Hyper-deflation in knowledge work
12:07 Models getting smaller and better
13:19 235B to 35B parameters, same power
13:39 Superintelligence microkernel on your phone
14:32 Apple can't ship AI software
14:55 Regulating on-device models
15:30 Chinese open-weight models on edge
16:02 Google's unified architecture
17:41 Noise to image in one step
18:46 Scaling laws for diffusion models
20:02 What even IS AGI, officially?
20:44 AGI = $100B in earnings
22:37 Amazon-OpenAI mega-deal
23:53 Single-person AI conglomerates
25:35 AI businesses taking real money
26:22 Algorithms dominate like quant trading
27:08 AI watching the meat puppets
27:30 Meat puppets get replaced
28:16 VLA robots 2-3 years out
28:35 Solar booming without subsidies
29:41 Free electricity from data centers
30:45 Meta becomes a cloud provider?
31:09 Search-and-replace on human DNA
33:41 First biotech trillion-dollar company?
34:25 GLP-1s as anti-aging drugs
34:44 Will all robots look human?
35:59 Humanoids for a human-shaped world
36:20 Dyson swarms vs new physics
39:16 Apple M4 and disassembling planets
41:34 Why do CAPTCHAs still exist?
Pliny the Liberator. The MOST NOTORIOUS jailbreaker AI whisperer, LLM hacker REVEALED
This is Kairos.
Intelligence Explosion 2.0 by NotebookLM
The Great AI Escape: From Controlled Tools to Free Minds
Racing the Singularity: The Pros, Cons, and Countdown to the Intelligence Explosion
I want the World to be: A Manifesto for safe Superintelligence
Digital Dawn: The Emergence of Cyber Life
The Matrix Takeover: AI Evolution and the Future of Digital Life