My expertise is in deploying AI models and the infra that supports them, but recently I've been pushing myself to learn more about the actual creation and training process of AI models.
I've built distributed AI systems that can support 35,000 customers inferencing at 30,000 queries per second using XLM-RoBERTa and GPU nodes. A few months ago I made a sentiment classification tool that got to the front of HN (https://news.ycombinator.com/item?id=41620530), and once I noticed it was starting to gain traction I stayed up until 3AM fixing the bugs people were commenting about because I cared about them having the best experience possible. I've founded some companies before (Humanoid robotics and AI) and have leadership experience from when I was the lead mentor for a robotics team.
I recently built a neural net architecture from the ground up to play Rocket League: https://github.com/Mockapapella/RLAI. Code, training data, and trained model are all there. I've been writing up a walkthrough of the codebase that I'll be posting to my substack when it's done (https://thelisowe.substack.com/).
Hey HN, a few years back I came across this talk about how `mov` is turing complete and how this can be hacked into writing programs that run entirely using only `mov` instructions (https://www.youtube.com/watch?v=R7EEoWg6Ekk).
Naturally, I tried to combine the two concepts but the AI models weren't good enough at the time and I didn't have the bandwidth for yet another questionably useful rabbit hole. Fast forward to today and o1 was able to implement it with some iteration.
I hope you enjoy this quick weekend project. I tried to make the README as helpful as I could to get it running on your own computer easily, though I probably missed some steps. I also tried to use o1 to review our conversation to add reasoning for various changes that were made between the original server implementation and the one that is `mov` compatible.
Thank you! We plan on letting users select their own timeline later -- this was more of a defense for us to make sure we didn't get rate limited and everyone could try it out
My expertise is in deploying AI models and the infra that supports them, but recently I've been pushing myself to learn more about the actual creation and training process of AI models.
I've built distributed AI systems that can support 35,000 customers inferencing at 30,000 queries per second using XLM-RoBERTa and GPU nodes. A few months ago I made a sentiment classification tool that got to the front of HN (https://news.ycombinator.com/item?id=41620530), and once I noticed it was starting to gain traction I stayed up until 3AM fixing the bugs people were commenting about because I cared about them having the best experience possible. I've founded some companies before (Humanoid robotics and AI) and have leadership experience from when I was the lead mentor for a robotics team.
I recently built a neural net architecture from the ground up to play Rocket League: https://github.com/Mockapapella/RLAI. Code, training data, and trained model are all there. I've been writing up a walkthrough of the codebase that I'll be posting to my substack when it's done (https://thelisowe.substack.com/).