=> More informations about this toot | View the thread
blogged about deepseek: i am not sure that small models on everyone's computer is a good thing, but if LLMs are gonna be forced down our throats either way this seems a lot better than giving sam altman a monopoly on them
https://blog.bront.rodeo/deepseek-r1-initial-notes/
=> More informations about this toot | View the thread
=> More informations about this toot | View the thread
doing my therapy homework
=> More informations about this toot | View the thread
i really do think more attention should be paid to the absolutely insane rationalist community as the rosetta stone of the last 20 years of (american) tech
=> More informations about this toot | View the thread
i have started another hundred days project but this time i'm blogging. i wrote about why roko's basilisk is stupid, a project i did back in august, and how "agentic" ai isn't really much of anything right now https://blog.bront.rodeo/making-a-reverse-rokos-basilisk/
=> More informations about this toot | View the thread
now more than ever i am unsure how to reconcile the part of me that gets excited about the potential of a tool with the part of me that knows how it will inevitably be used
=> More informations about this toot | View the thread
i have been trying to be more open-minded about AI stuff since, despite my best efforts, people are finding genuine utility out of these tools and i don't think that should be discounted. i am just not convinced that "i got claude to make a buggy app for me" is going to be good for anyone in the long run
=> More informations about this toot | View the thread
just in time for openai announcing they've achieved AGI (lol, lmao, fuck off) i wrote a little bit about my problems with LLMs and "frictionless design" https://blog.bront.rodeo/the-trap-of-frictionless-design/
=> More informations about this toot | View the thread
mastodon vs. bluesky debate largely appears to be based on faulty prior that websites can be good
=> More informations about this toot | View the thread
pretty good breakdown from @cwebber if you, like me, have no plans to read the ATProto whitepaper. tl;dr: bluesky's not really decentralized, but also that's not really the problem it's solving for.
https://dustycloud.org/blog/how-decentralized-is-bluesky/
=> More informations about this toot | View the thread
=> More informations about this toot | View the thread
still working out a couple build kinks, but the game i've been making on and off for the past year is available to the public! it's a short visual novel about data labelers, AI, labor, and space aliens. would love for you to play it. https://hellagrapes.itch.io/dataville
=> More informations about this toot | View the thread
=> More informations about this toot | View the thread
i like to think that i would make the right choice in most analogues of the stanford marshmallow test but 90% of the technical decisions i make are strong evidence to the contrary
=> More informations about this toot | View the thread
=> More informations about this toot | View the thread
=> More informations about this toot | View the thread
=> More informations about this toot | View the thread
i guess it's a sign of growth to go from "i feel bad because i don't understand this stackoverflow answer" to "i feel bad because i don't have opinions about python enhancement proposals"
=> More informations about this toot | View the thread
one cool thing about being a programmer is that the act of learning something inevitably introduces you to a new, weirder thing to have impostor syndrome about
=> More informations about this toot | View the thread
=> This profile without reblog | Go to brent@post.lurk.org account This content has been proxied by September (ba2dc).Proxy Information
text/gemini