Book

• Contact On The Record
• Contact Midday
• Contact the Pratt – done

SBIRs

• 9:15 Standup
• Talked to Rukan about the idea of using a mixed NN/Markov sim. He’s going to see if he can find a preexisting model (e.g. a multi-link pendulum) to try training the model with. Also set up an overleaf doc for the methods and results
• 1:00 BMD status
• Need to get slides submitted for approval (and maybe reformatting?)

GPT Agents

• Made good progress on the paper and got a lot of good results over the long weekend. I’d like to get the methods section done before the Wednesday meeting
• Starting the LangChain tutorial, which looks good. Got my Huggingface token so I can try models on the HF hub without downloading. Should make it nice to try context prompting with smaller, more servable models