My understanding is that the AI companies push their servers so hard that the components are basically consumables. Consumers don’t really press their machines to the point of physical exhaustion.

In WarGames the computer plays tic tac toe against itself until it realizes it’s a solved game and there is no way to win.

They forgot to make their LLMs play thousands of games of tic-tac-toe first.

They did gove some advice. They said to go with a vendor that is transparent about problems and reveals the results of their third party security audits. I’m sure if you read between the lines it means they likely reviewed several vendors and chose to spend their time attacking ones that are opaque about their security stance and used outdated encryption or bad implementations of E2E encryption. So all three are likely suspect. Like if 1Password were developed similarly to LastPass wouldn’t they have spent time attacking it?

Edit: https://support.1password.com/security-assessments/

1Password are posting the results of their external pen testing now.