Trump won NH.

It’s helpful to understand that this is not something that Trump started. It’s been a long term effort on the part of the right wing to create and exploit “culture war” issues:

Also this, which seems legit but might not be. The account is now suspended so I can’t check, and I can run my tools on LLM error messages any more:

Sure seems to be a thing though:

SBIRs

• The LM meeting yesterday makes me think that “Simulation Acceleration” per se isn’t really what people want. There is a high level of confidence with Monte Carlo techniques, not the least of which is that they can be validated statically, not statistically, like a NN would have to be. But there is a lot of interest in using NNs to store and integrate the results of simulation across thousands of runs, so that new questions and answers can come from an existing “corpus” of training data. This kind of “simulation enhancement” or “simulation extending” I think is the key to effectively “accelerating” simulation.
• Travel expenses! Done!

GPT Agents

• Work on the RAG paper. Due tomorrow.
• Finished incorporating all the changes. Now anonymize and get ready to submit? Done! For now, one more pass tomorrow