Ask HN: Is there a great book for understanding LLMs?
8 comments
No, LLMs move fast and the people who write good technical books
1) are rare
2) take time to write them
I also don't know if there is an incentive to spend a lot of time on an elaborate piece of work given how likely it is to be obsolete within 1-2 years. There's certainly incentive to look like you have written "the book" on LLMs which why the space is flooded with rushed, low quality books.
The good thing is that you can learn everything you need to know without a book on the topic - papers, tutorials, videos, code repositories etc.
1) are rare
2) take time to write them
I also don't know if there is an incentive to spend a lot of time on an elaborate piece of work given how likely it is to be obsolete within 1-2 years. There's certainly incentive to look like you have written "the book" on LLMs which why the space is flooded with rushed, low quality books.
The good thing is that you can learn everything you need to know without a book on the topic - papers, tutorials, videos, code repositories etc.
Andrej Karpathy: Let's build GPT: from scratch, in code, spelled out.
We build a Generatively Pretrained Transformer (GPT), following the paper "Attention is All You Need" and OpenAI's GPT-2 / GPT-3
https://www.youtube.com/watch?v=kCc8FmEb1nY
Look at this book preview and see if it fits your needs: https://www.google.com/books/edition/Introduction_to_Large_L...
Read Stephen Wolfram's book on ChatGPT
I guess for folks who don't want to pay, this is the material to read?
https://writings.stephenwolfram.com/2023/02/what-is-chatgpt-...
https://writings.stephenwolfram.com/2023/02/what-is-chatgpt-...
How AI Works: From Sorcery to Science
Book by Ronald T Kneusel
Might give you some answers
Might give you some answers
meanwhile, books are getting outdated thanks to LLMs. the irony...
Any suggestions?