@karpathy Wait, is it correct to say that there’s an “intrinsic reward function” in the LLM? Or is this specific to when the llm performs chain-of-thought at a token level?
@jeremyphoward@cursor_ai@AnthropicAI Hi Jeremy, I wanted to get your opinion on something. Right now, who provides the best "coding buddy" service among the openAI, Anthropic and Google (i.e. Gemini)?
I am looking for something that can help me understand a whole GitHub project, not just auto complete code.
@karpathy Please do something similar with papers you liked or you think are landmarks in the LLM/deep learning space! It would be cool to have such a playlist of episodes curated by yourself 😊
Not to be too much of an AI Doomer but I just had a FaceTime call with a near realtime ChatGPT-powered AI avatar and it was…surprisingly human.
Anyway, I recorded it.
This is the actual response speed, it’s very fast.
(You can also hear me kind of lose it towards the end)
New paper! 🚨
Our PhD alumni, Dr @damiFass developed a pipeline for in silico study of engineered heart tissue in infarcted ventricle models, published in Computers in Biology and Medicine (1/2):
Such a nice day in the park we got today: weather was nice(-ish), pizza was provided and dogs were welcome!
It was the perfect occasion to update the group photo (although not everyone could make it today ☹️)
Find more about us at https://t.co/8qNzyvIG1J
We end the week with great news: @stelong_ just passed his #PhDviva with minor corrections!🎉
Congratulations Dr Longobardi and best of luck on your new role in @GSK!
Paper alert! 🚨
Our #PhD student @damiFass just got his first first-author paper out in @PLOSCompBiol !
In it, they modelled heart tissue patches on different mammals and explore the effect of different patch designs
Acess it 🔓at:
https://t.co/XKEPkfIncW