20 November 2023

urbit is a zorp project

ZK is the next wave of pure CS

AJ LaMarc
AJ LaMarc @ajlamarc

The @ZorpZK presentation had incredible rizz and is very bullish for @urbit. Some thoughts that come to mind:

Innovating in CS is extremely difficult. Things happen in waves of increasing sophistication, and you can’t just copy the current trend, you have to aim far ahead (as @jlehman_ noted). This, of course, takes young and highly ambitious people with long time horizons, who also aren’t extremely profit-motivated (which Urbit has a respectable proportion of!)

Things that Urbit talks about that feel outdated: personal server, personal software, natively networked computer. These are solved problems, or at least, Urbit is not positioned to solve them.

We don’t act like it - but human-specific or “personal software” goes against the reality that software is hard to build, harder to scale, and costs a lot of money. The dynamic of this will change with strong enough AI - this will be constrained by compute, which is a blocker for Urbit.

The next frontier for “pure CS” is ZK. The key is that the ongoing LARP around software privacy, which the market has always rejected in favor of a better product, can become real. People will approach things a bit differently re: AGI and Brain-Machine Interfaces.

I can’t hope to appreciate 80% of the nuance but Logan’s description of the (my summary) “ZK, secure, private, performant, portable, God mesh computer network” sounds valuable in some capacity 30 years from now.

And they are willing to meaningfully tweak Urbit constraints to build it - the new programming language and fixed-width addresses come to mind. Excited to see what the full release looks like.