@jeremykauffman Or better yet, why should I care about monkeys? monkeys don't care about me

@OpenAI Couldn't help noticing that sometimes the model is guided to refrain from giving too much detail to not overload the user. Sometimes that results in lazy and over broad answers. I think the model should be able to decide when detailed answers are suitable

@sama the chatgpt ux has become bloated with too many models, features that are sometimes available in a seemingly arbitrary way, and ui elements that don't interact well with each other. we need to go back to the simple text box, and make it smart enough to understand what to do

@adonis_singh honestly, couldn't make a single thing work like this with o1. its super lazy and write script with placeholders instead of actual logic

@arankomatsuzaki "Overall, these results indicate that scaling general-purpose reinforcement learning, rather than relying on domain-specific techniques, offers a robust path toward state-of-the-art AI in reasoning domains" Frame it on the wall

@SunoMusic @Timbaland Feature suggestion - give me the option to upload a song and get a detail musical style description for it, for getting inspiration

@DimitrisPapail Spiky intelligence

@Josh9817 @aidan_mclau None of those things are ooen source

@aidan_mclau the real concern is the fact we're going to have an authoritarian superpower developing the first ASI. and it's not going to be china.

@Yuchenj_UW If we could only find better ways to name models...

@tszzl the obvious solution is stop eating fast food. in fact, you should do this regardless of being on ozempic.

@zephyr_z9 why not 155 trillion while we're making up numbers?

@MumbaichaDon @narendramodi the real safety issue having the USA develop the first ASI while being an authoritarian state

@labsdotgoogle when do we get veo2?

@TheHumanoidHub why are all humanoid robots converging to this faceless boring design? sci fi robots looked cooler

@daniel_mac8 twitter is much better since I did that

@annmarie @elonmusk @TomMackenzieTV honestly I have never seen nothing other but honesty and genuineness from Sam. every bad thing that was said about him comes from business opponents, and always without much proof. musk is a thousandfold less trustworthy

@MKBHD youtube IS my tv

@emollick though its just based off Claude usage, which is tiny compared to chatgpt. and also really popular among certain niches. not sure this data is representative.

@karpathy man, this post was really prophetic