More about the challenge: frugal-ai-challenge.github.io
What’s more? We open source our winning solution and alternative attempts! Check out github.com/D2KLab/fruga...

hey florent, that sounds lovely , would love to hear how it's going to be possible to support (citizen?) journalists and the news room

it's true i dont split hairs, but i've used a lot of ai-s, here's one with a few publications github.com/SakanaAI/AI-...

i guess it fits your definition and writes papers with novel ideas ?

i guess we'll never agree that your starting point is not sterile, but i do have arguments with citations 😉

but you can try :-) huggingface.co/spaces/MISAT...

ESM3 is kind of an LLM , and now it's been available for free for four days :-)

sure LLMs are tools, there are also "vision", robotics, and a lot of architectures (for each :-) available, all these tools are already part of the medical profession, including for small molecule discovery. you dont use a stethoscope for writing notes, you dont use LLMs for small molecule discovery

yes, there are more than 100 "ai-native" drugs currently being evaluated in clinical trials (with few successes) ars.els-cdn.com/content/imag...

my point is we should deal with the world we have , not the world we want.

after all medical science kind of invented ai ...

FDA says over 500 clinical stage submissions in the period 2021-2023 www.fda.gov/about-fda/ce... just as an example

obviously this : doom-captcha.vercel.app

one thing i really like to see is the huggingface.com mirror , click on "+" , create a "dataset repository" , add the yaml to the README.md , and it will "get splits" directly integrated into a variety of dataset libraries (including a very useful SQL language capability) , free of course

<3 BSKY IN BIO <3

dont you dare tell me this is actually ai ! but i dont recognize the town

🙋‍♂️ if there's any room left :-)

🙋‍♂️ would quite like to join the list :-)

totally get that : she's a beaut' 👏🏻👏🏻