supercluster.laniakea.tv What used to require racks of GPUs, massive cooling systems, and millions in infrastructure is now being compressed, optimized, and shipped into silicon that fits in your hand

https://video.wixstatic.com/video/8c5863_43e66ebf45084ce1b7f86b279ad1967c/720p/mp4/file.mp4

supercluster.laniakea.tv What used to require racks of GPUs, massive cooling systems, and millions in infrastructure is now being compressed, optimized, and shipped into silicon that fits in your hand. If you zoom out a bit, we are actually seeing extremely rapid take-off of AI, but there is enough movement of people/ideas that the leading AI companies, for now, appear to be in similar positions. I want really powerful local models, but I’m also bombarded with emails/messages from people complaining how even the top-tier models are not good enough, make mistakes, or don’t follow instructions well enough. Neural nets don’t care about your hardware labels. They don’t care if it’s a datacenter or a phone. They just map input to output. Managing open source is way harder than managing paid closed source projects. At work, you have authority. With OSS, you have GitHub Issues. Deterministic parts of the code, such as the operating system, are already addressed by various methods like triple-voting redundancy. It’s simulations all the way down, until you get to the 1-bit universe. Once you get to 4 bits, you have 16 possible outcomes. You can then do whatever you want with those outcomes.