TheAlchemist•3m
Ask HN: AI Reading Listitem?id=44077040

In the thread about John Carmack presentation, somebody mentioned the reading list he got from Ilya which were crucial to understand what matters and the current state of the knowledge (at the time).

After some googling, it seems like this list is plausible, although not confirmed: https://github.com/dzyim/ilya-sutskever-recommended-reading?tab=readme-ov-file

What would an actualized list look today ?

I'm asking since I'm taking a similar direction to Carmack (with maybe 1% of his talent and skills, but hey it's all about the journey !) in that I don't intend to work on improvement existing models, which I also believe are a road to nowhere, but rather understand things from the ground up, and maybe figure out 'something different' at the end.

4 comments
  • nextos•3m

    Kevin Murphy's book, particularly volume 2, is really well organized and summarizes the state of the art: https://probml.github.io/pml-book/book2.html

    It's a really good entry point to current ML, written by an expert who has been in the field of generative models before they incorporated deep features.

    • tuyguntn•3m

      How do you find time to read 1300 page technical book? I admire people who do it for fun.

      By the time I finish this book, probably SOTA will look totally different

      • nextos•3m

        I don't think his book is meant to be read from cover to cover. It's useful in many ways, as an outline, as a reference, or as an entry point to your niche of interest.

        Let's say you are interested in diffusion models. You read the corresponding chapter (and play with the companion code).

        If you are missing some background, you can also backtrack and skim through the initial chapters.

  • MrCoffee7•3m

    I think a good approximation to getting a good current optimal list would be finding a syllabus online for a recent course offering in your AI area of interest. Most lists like that would combine some hands-on work with readings.