Ryan Stanley

Staff AI/ML Engineer | Aerospace AI, simulation, inference, and secure ML infrastructure | VT Aerospace, GT Computer Science | Views mine

[Ryan Stanley @AeroMLRyan]

The variable cardinality Choice interface reminds me of the pointer/attention operations used in models like DeepMind's AlphaStar models for target/unit selection. I can see how this approach can be significantly parallelized and utilized in agent harnesses. Very clever and congrats to the team.

[Ryan Stanley @AeroMLRyan]

@thsottiaux The ability to select multiple conversations for deletion from my chat history instead of the current one-by-one approach