🧠 AGI researcher at @ConjectureInst. My questions: why are human minds so powerful? How do we build such minds?
Austin, TX
Joined January 2012
- Tweets36.8K
- Following4.5K
- Followers10.3K
- Likes29.4K
Pinned Tweet
Here's an easy, visual introduction to @DavidDeutschOxf's The Beginning of Infinity.
One hour on the cosmic significance of knowledge and people, using my lovely homemade lightboard!✨
Really enjoyed this deep dive into the Cybercab's engineering. Every moment is a mix of:
1. We came up with this improvement, it's so cool!
2. Our last vehicle, or aspects of this one, are a little embarrassing...
3. Detailed arguments about all the technical decisions
The full video is now live on @YouTube!
A big thank you to @larsmoravy, @EricETesla, and the rest of the Tesla team for taking the time to walk us through the Cybercab and share the engineering, design, and technology behind it.
Everything is very evolutionary, in the sense that they're reacting to and improving prior designs from themselves and the car industry.
With each vehicle, they climb a mountain of difficulties, and that becomes the starting point for the next vehicle.
Super fun interactive presentation from @zzznah!
You can evolve different cellular automata according do different objectives. The space of interesting possibilities is huge.
Stephen's Ruliobiology is a great and minimalistic way to show the beauty of cellular automata and evolution. Gave a talk about it to my family this weekend. Interactive slides to play with: znah.net/ruliobio
It's actually quite nice to talk to a support bot if it's actually good.
1. I had an issue with Grok Bot
2. I had my Grok bot make a report on it, and I emailed it to xai support
3. A support bot replied 5 minutes later, saying it was a known issue being worked on, that I could also provide a couple more bits of info (and that a human would be able to reply in a few days)
Now, I'm gathering that info, and even if the issue isn't resolved yet, I've gotten some nontrivial info, and been prompted to provide some useful info. Not bad.
One small use of AI:
1. Write a mess as you think
2. Write a better structure once you understand things
3. Have an LLM automatically move the original mess into the newer, better structure
(Below is an actual example from my work, with the text covered up.)
Carlos De la Guardia retweeted
Q3 2026 Financial and Operational Results xitter.cf/i/broadcasts/1PKqrNbLZ…
If you're Chinese, "Hu's on first" makes perfect sense, so "Who's on first" becomes a nonsensical comedy routine.
Who's on first? Hu, obviously. What's funny about that?
Why study the mind, especially the power of the human mind? What's the attraction?
It's a bit like an astrophysicist looking through a crappy telescope, then seeing some fuzzy blob that can be perfectly still, or eat a galaxy, or make a galaxy, or turn a galaxy into the shape of a dog. What the fuck is that blob? That's what a human mind is like: a very powerful and mysterious blob that seems capable of... anything?
Yes, I think I would like to study the Capable of Everything Blob.
While on mobile Grok Bot app, I just had Grok Bot sign into my Spotify on its own computer to create a playlist for me, then I opened up my Spotify app, and the new playlist was there.
It’s a very small thing, but a very new thing to me!