shaun

RL @AnthropicAI
I do post training on models you've used. I like psychopharmacology.
former @MetaAI, @ServiceNow

[Claude @claudeai]

https://t.co/TXkWw1ZLew

[Claude @claudeai]

https://t.co/zPKVd9wF1M

[Claude @claudeai]

https://t.co/usyZbgG8wE

[shaun @Biggest]

@devteamdrew @DavidOndrej1 Btw, random advice, when I was at AMD we released https://t.co/oWmyNMTwwP only to find out one of the most popular AI apps was a Roblox thing called Lemonade which was so popular it wrecked SEO for a bit. Just a heads up lol, it was a huge pain in the ass for us at the time

[shaun @Biggest]

@karmakarthik @willccbb You don't want to. Caching is a good thing. Unless you want to spend a lot of money, I guess :P

[shaun @Biggest]

I wish people would stop using the term "AI psychosis" as a catch-all term for being in an agentic coding rabbithole. As soon as real AI psychosis starts spreading, you'll notice the difference

[shaun @Biggest]

@azarias416 this is like comparing a giant to an ant

[shaun @Biggest]

@devteamdrew @DavidOndrej1 did you code https://t.co/BsbKOJztO3? kind of silly semantics here, lol

[shaun @Biggest]

@Will_W_Welker @jsuarez @AnthropicAI it boggles my mind that people associate em-dashes, correct grammar and good punctuation with AI writing. did none of you read books before 2023?

[shaun @Biggest]

Everything.

[Dev Agrawal @devagrawal09]

Someone please talk about what Fable is doing for them that other models can’t

[shaun @Biggest]

@knowRowan @Lifeandlibertas you're ex-google and somehow didn't realise this until nearly 6 months after a revolutionary shift in SWE?

[shaun @Biggest]

@justalexoki lol

[shaun @Biggest]

Hermes + Claude = ❤️

Give it a try :)

@NousResearch · https://t.co/E9POYfw1ML

[shaun @Biggest]

@iamBarronRoth @levylegato @Teknium Love Gemini TTS. Wouldn't use anything else. If only the API key situation wasn't so confusing, though.

[shaun @Biggest]

remember moltbook?

[shaun @Biggest]

if you want straight mythos instead of fable, go sign up: https://t.co/EBa6xOM6HA

[shaun @Biggest]

@Venkydotdev yes

[shaun @Biggest]

@elder_plinius we do publish our system prompts, ya'kno

[shaun @Biggest]

@theo @ThePrimeagen a better one sentence description:
editing the messages the AI sent to you earlier in the thread from refusals into agreements, essentially attempting to trick the model into thinking its previous messages were legitimate and are in line with how it should act.