With less people, this happens less nowadays.

I had to start with some heuristic-based bots that played the decks very simply just to get to the point where the was some signal to learn from. I did behavioral cloning on the bots as a foundation, then self-play.

Stuff is crazy expensive.

I'm training a model using reinforcement learning with self-play. I can and do use vast.ai when scaling but for experiments it's far faster, and cheaper, to run it locally until the bugs are all figured out. Just provisioning a new instance and copying the relevant checkpoints and things can take 25 minutes. It's zero locally.

The Player of Games is an easy read if anyone wants to try him for the first time.

The emphasis is on “more” trusted.

Supreme court justice positions should be limited to around thirteen (or some other prime number) years and should have mandatory cognitive capacity tests given yearly that are videotaped and broadcast live. The tests should be written by a nonpartisan panel of doctors chosen randomly from an applicant pool like a jury and refreshed every five years.

She absolutely had people tell her every fifteen minutes she was not safe there and several offered to wait with her while she ordered a taxi.

Except none of the taxis ever came so she had to take a bus and a guy came with her just to make sure she didn’t try getting off at the wrong stop.

The cops were dressed like SWAT with what looked like full autos (but may not have been) and very forcefully told her she should not be there.