Yeah my guess from the very limited information I have is the abort-to-spashdown decision was probably automated, so I’m extremely interested to hear why they aborted the catch. It feels a bit at odds with the ambition to catch a starship soon, but if their code is so conservative about aborted catches, it seems realistic to at least try

Honestly, I don’t think this could have gone any other way. As the article says, this is the so-called “Lane 1” missions which, while they’re less important and less heavy than “Lane 2” missions, are perfectly suited to the Falcon 9, or heavy if needed. There’s no need to get a heavy rocket like Vulcan involved, especially since it’s not proven yet, so the only reason I could think of for someone else winning would essentially be as a subsidy to help SpaceX’s competition. Once Vulcan has a couple dozen successful missions under its belt I’ll expect it to be more competitive.

I’d be interested in setting up the highest quality models to run locally, and I don’t have the budget for a GPU with anywhere near enough VRAM, but my main server PC has a 7900x and I could afford to upgrade its RAM - is it possible, and if so how difficult, to get this stuff running on CPU? Inference speed isn’t a sticking point as long as it’s not unusably slow, but I do have access to an OpenAI subscription so there just wouldn’t be much point with lower quality models except as a toy.