I realized that should I end up getting laid off soon I won't have an unlimited token budget and for the workflows I've settled into it would be quite expensive. So I was exploring what it would take to run open models at home.
Was quite disappointed to see that the PC side hasn't kept up. The unified architecture on Macs makes it very hard to justify spending money on a Linux machine for inference workloads.
It's more that Unix systems were timesharing systems, any user could run a daemon, but you didn't want users to have the ability to grab a port used by system services, not just because they could impersonate a system service on the network, but also because then you couldn't trust localhost services, either, as well as it just being a PITA. This is still true today; though vanishingly few Linux systems are multi-tenant, it's still common to implicitly trust a local service.
I'm enjoying it. It's wild to realize that I spent countless hours playing Theme Park when I was around 10 years old, and Demis had been a big contributor to the game when he wasn't much older.
Also I don't really care that it's a bit of a cheerleader for DeepMind and Hassabis. Substantive criticism is good, but too often with these kind of books it feels like an editor told the author that the book needs something negative and the author has to inflate an issue to meet the requirement.
The author did give him credit for the whole you-can-make-the-fries-super-salty-to-increase-demand-for-drinks thing in Theme Park, which I remember vividly. (I, too, dropped many hours on Theme Park as a kid.) Although I imagine there’s about half a dozen people who lay claim to that idea.
The first time we traveled domestically in China I kept thinking that my wife had to be mistaken, there has to be some kind of confirmation we need to show in order to board. But nope, it literally is just show up with your ID.
