Discussion about this post

User's avatar
Defender's avatar

I am still eagerly awaiting part 2!!! I've been checking every few days (I've shared this with a bunch of people already, I think it's great and I can't wait to see this line of thought fully fleshed out)

I've been thinking a lot about this lightning animation (https://x.com/i/status/2036461057699688511) as a metaphor for the correct resource allocation. Once a "loop" is done, looking backwards and retroactively rewarding all the parts along the path that helped that loop close, to the degree that they helped. This is the "ultimate" / highest correctness way to do it, but it can only be done retroactively. Systems that offer reward ahead of time try their best to approximate/predict this future performance, but do need to receive feedback "when all is said and done" and update.

Ready for more?