sean watters

@swatters.myatproto.social

sean.ordinary.host | software engineer. profile alt: a much younger me, sitting in front of an 11yr old, mac se/30. header alt: in focus; a lego bonsai tree, in front of—out of focus—guitar head/neck, back of a heavily stickered 16" macbook.

Now is a good reminder that machine learning is a rebranding of statistical learning. Reinforcement learning is something I first came across in control theory. Many ML formulations are from information theory, a much more theoretically grounded discipline that's been around for decades.

Casey Newton@caseynewton.bsky.social · last mo.

There's a reason why "these are just statistical predictions" hasn't caught on in the discourse, and it's not the tech press. It's that people see them reading and writing and reasoning and scheming and they realize that the stochastic-parrots framing is outdated and inadequate to the moment

I really wish folks would update the language around "AI" to be "generative AI". There's other forms of it (which is really just machine learning) that y'all lean on periodically (like ducking autocorrect, OCR, spam detection, [good] search engines, etc, etc). Fuck Anthropic and OpenAI for that.

I'm coming to the conclusion that community-owned and operated small clouds (co-ops) with easy onramps for self-hosting open source services like mail, storage, and VPN are the only way forward. Every corpo service is eventually going to make you ashamed to use it.

It's been a journey, but all the conversations have finally shifted from "we need to build a better Twitter on *my* protocol" to "we need to build coalitions *across* protocols" Technology isn't the goal, it's the tool. Community is the goal. From my "But First, Interoperability" essay last year:

There will come a day when we can debate which protocol is the greatest of them all, what the definition of decentralization is, and which company, organization, or developer did it best. We should have that conversation when the open social web is what users have chosen as a default option.

But first, interoperability.

It's easy to see how selecting for willingness to support white supremacy, weeds out competent Black and brown people. It's harder to see that it also weeds out competent white people that don't tolerate open racism. That's half of your white talent pool! And the half that's better at science! 🤷🏿‍♂️