Yarvin is BACK.
Curtis Yarvin Wants to Make the Internet Uncensorable
@curtis_yarvin has one mission: make Casey @Rodarmor try Urbit again.
Curtis joins Casey and Erin Redwing @realizingerin for a live demo of the new Urbit experience. We also get into Curtis's nearly 25-year quest to reinvent computing, what went wrong with the original Urbit, and his new vision for a "self-driving computer" where AI agents write and maintain their own applications.
Maybe the world just needed to get weird enough for Urbit to work.
TIMESTAMPS
0:00 Why Curtis Yarvin Came to Hell Money
5:05 Usenet, Haikus & Early Internet Culture
7:00 How the Internet Lost Personal Autonomy
11:35 Designing Urbit: Address Space, Funding Model, Lisp-Like Architecture
15:10 Berkeley, Grad School & Early Mobile Computing
17:10 Funding Urbit & Founding Tlon
18:06 What Is Urbit?
22:10 Casey's Thoughts on Urbit
25:50 Live Demo: Curtis Shows Casey the New Urbit
53:54 Curtis Yarvin's Self-Driving Computer Manifesto
1:02:00 Why AI Could Finally Make Urbit Work
1:18:00 When Will Urbit Get AI Agents?
1:28:00 Uncensorable Computing & Internet Privacy
1:39:00 Curtis Yarvin on AI Doomerism & Cybersecurity
1:42:00 Could AI Be King?
1:49:00 Final Thoughts: The Future of Urbit & AI
Found a fun out-of-bounds teleport in Order of the Sinking Star. @Jonathan_Blow Great game, highly recommended!
I don't know what it will look like, but some amount of computer use will be completely reshaped by the ability of ordinary users to produce custom software at will.
We interviewed Curtis for @hellmoneypod about his latest endeavor, dropping next week.
As the barest courtesy, the AIs could at least bestow upon us the sweet release of death.
Is AI Actually Going to Kill Us?
On the Halloween special episode of Hell Money, Casey @Rodarmor and Erin Redwing @realizingerin dress up as @AnthropicAI's Dario Amodei and @OpenAI's Sam Altman to discuss AI doomerism and the possibility that superintelligent AI is going to kill us all.
Nothing ever happens. Everything always happens.
TIMESTAMPS
0:00 @sama vs @DarioAmodei
6:34 P(doom), AI Doomerism, Rationalism, & Super Intelligence (ASI)
10:04 AI Doomsday Scenarios
13:30 How Superintelligent AI Could Escape Containment
18:23 The Vulnerable World Hypothesis: Bioweapons & Nuclear Weapons
22:30 Agentic AI Worms & Hacking Everything
25:30 What If AI Takes Over the Internet?
28:55 Will AI Keep Humans Around?
33:30 Recursive Self-Improvement
38:55 The Case Against AI Doom
50:20 Are AI Doomers Afraid of Becoming Obsolete?
52:35 Building the Machine That Replaces You
59:35 AI Safetyism & Silicon Valley Doomer Culture
1:05:40 What If AI Just Plateaus?
1:09:50 The AI Bubble & Anthropic's IPO
1:14:00 Sacrificing Capital to the AI God
Ord 0.29.1 is now available! A couple of notable changes:
- Correctly deserialize Bitcoin Core RPC responses containing anchor outputs
- New recursive endpoint `/r/block/<QUERY>` that returns the entire block
Share and enjoy!
github.com/ordinals/ord/rele…
Big blockers and communists both have good criticisms of bitcoin and capitalism. The problem is that they then go out and write moronic sci-fi fanfics about their completely unworkable alternatives with no grounding in reality.
youtube.com/shorts/NGrEUxgps…
Two very good scripts, spool and despool.
Easily move files between computers using a shared git repository.
# source
$ spool FILE
# destination
$ despool FILE
github.com/casey/dotfiles/bl…
github.com/casey/dotfiles/bl…
…is telling the agent to never touch git, and using Fork[0] in a separate window to stage changes as I review them. This lets review be incremental and continuous, even as the agent continues to work.
[0] git-fork.com/





