Justin Milligan

[Justin Milligan @Justin_m34]

TL;DR I wrote a paper with @DaveBlundin about ethical compute usage: (https://t.co/FpQgkfqs3G)

I'm excited to post for the first time about Lightwave, the startup I've recently been working on after graduating Princeton and working at JPMorgan and Link Ventures. I care deeply about steering AI development in a direction that is good for humanity. In this vein, we're going to be posting a series of articles about our beliefs and our technical perspective in hopes that A) they'll spread and B) that people will provide valuable feedback

Since starting Lightwave I've been thinking about how compute should benefit humanity. Now and for the next couple years, we are living in a compute constrained world, meaning we don't have the infrastructure to support the increasing demand for AI. I've already seen this firsthand in our search for compute internally.

This has a really important philosophical implication: every use of compute has an opportunity cost. Using AI for one purpose necessitates less compute for another. Super simple idea, but also super meaningful. This led Dave and I to the question: if we could allocate the world's compute perfectly, how would we do it? We landed on efficiency, flourishing, and alignment, and we wrote a paper explaining exactly what those mean.

I believe anybody wielding significant compute should think deeply about any short-term use case that does not fall into one of these categories - and I hope that by putting this out into the world I might help spread these principles.

https://t.co/FpQgkfqs3G