Engineering AI Reading Group @ HomeBrew

​This week is on AutoResearch. It's been a slight update since the AutoML days. ;)

​"any metric you care about that is reasonably efficient to evaluate (or that has more efficient proxy metrics such as training a smaller network) can be autoresearched by an agent swarm": Karpathy.

​

​Karpathy's Agent got Nanochat Time-to-GPT2 down from 2 hours to 1.8 hours. https://x.com/karpathy/status/2031135152349524125

​Tobi Lutke's Agent got his retrieval score up by 19% for QMD

​https://x.com/tobi/status/2030771823151853938

​Patrick Collison's Agent self built a weather forecasting model

​https://x.com/patrickc/status/2030775339228742107

​

​Engineering AI RG comes together to read and code on tools, libraries and codebases that are used frequently in the modern AI stack.

​All are welcome.