Do you think Italy banned ChatGPT because the Pope was annoyed at us putting him in so much drip? #Balenciaga twitter.com/EMostaque/stat…

@Altimor 2025 they outsourced my job to a Chinese AGI

@ElJayces @rishabh063 Whenever the devs are happy with it being done.

@lizardmech Check out the drip pope pictures

@Coolzippity 50 is good, will be optimised a lot during testing

@rishabh063 ...it literally says it in the above tweet

@Eleiber Two separate models two separate releases

@izzyz See my timeline. Now is the time for transparency on governance and standards even if industry agreed, directly in line with what those training these giant models themselves say.

@artificialguybr It's 2.5 times bigger than sd 2.1

@AI_Sensei_ Yes after model is finalized. Distilled models aren't really tunable amongst other things

Stable diffusion XL #SDXL is in testing with our partners & dreamstudio.ai, getting data to refine it before open source release. It is a 2.3 billion parameter variant (original 900m) with a range of improvements that will feed into Stable Diffusion 3 🚀 ✨

@followmarcos There is no evidence of the former and they are probably just waiting for the latter. All the North Korean hackers specialized in stealing crypto are likely repurposing to stealing model weights as we speak.

@followmarcos There is no evidence of the former and they are probably just waiting for the latter. All the North Korean hackers specialized in stealing crypto are likely repurposing to stealing model weights as we speak.

@itspoidaman Well yeah but that's better than the current status quo right

@ultracodemonkey Well it's a call for gigantic model training to take a breather until we put some policies in place. No open source gigantic runs.

@followmarcos The only thing that can stop a bad AI is a good AI eh? What if the bad actors just don't bother developing their own and espionage the weights away instead? Seems far simpler.

Pretty much echoes my stance tbh Experts mostly agree there is potential existential risk & we don't know how to mitigate it. Plenty of risks under existential level too. No transparent governance or insight into those building the models that could lead to this risk. twitter.com/JMannhart/stat…

The below from OpenAI agrees with the bulk of the core calls of the FLI letter and also notes potential existential threat etc I think v large language models will probably be fine but admit I could be wrong and they right Now is the time for proper transparency & governance twitter.com/EMostaque/stat…

Well this is interesting politico.eu/article/italia…

@PabloBernardoTW @Ugo_alves @RichardSocher I personally think even sci fi agentic AI will be fine but think now is the time to have the public discussion as I could be wrong and there are a whole host of real world impacts that will be hitting in the coming months from what we have today.

@PabloBernardoTW @Ugo_alves @RichardSocher We are not and do not want to train very large language models as our focus is on swarm not general intelligence Eve if we were, most of the letter is basically in line with what OpenAI themselves recently called for Good time for transparent governance twitter.com/EMostaque/stat…

@Plinz

@jasonprompts Tbh I signed because I saw Bill Gates did

@myndxero Fixed in #sdxl

@HoskinsAllen Yeah but it shows a willingness to have a chill break

@huemin_art laife

Honestly you should just all go and sign that FLI letter so we can all take a breather from the insane pace of progress. Maybe go on holiday. Read a good book. Summon a non-digital shoggoth. You know, relaxing stuff.

Another non-LLM project that's really cool! RoboPianist: training two hands to play the piano like humans. Very difficult but fun high-dimensional control challenge. - Simulator is fully open-source, built on Mujoco: github.com/google-researc… - Website: kzakka.com/robopianist/ro…… twitter.com/i/web/status/1…

Should we do open source robots 🤖