It’s the cleanest model I’ve ever seen. I sat on it for a year before going back to it and studying further and it holds up.
I try to find criticism with it and I can’t. It works.
I think OpenAI is the direction things will continue on as it has hooks to industry and scholarship with actual chips being developed, community support, etc. So, whatever work comes out of industrial / commercial // consumer value, will go through concepts linked to OpenAI’s models most likely.
It’s not the worst thing in the world, although I think it’ll make progress much slower than it needs to be.
Still, with a Practopoeisis model, I think it’s possible to utilize the products of OpenAI research to create some amazing things.