I thought it was very good for the purpose, which is to give the high level overview of relevant techniques and approaches. The right level of detail in a book that aims to cover multiple hot areas in GenAI. If it skipped more,it would have been too shallow. If it included much more, it would have been twice the size. The references in each chapter help the interested reader go deeper.
It could have been somewhat improved by reducing some redundant content between chapters and using that space to go a bit deeper.
Forgot to add the only technical book I read this year lol. Read this for interview prep in July this year. I used the exact same framework in an interview and got a good review, so it's a testimony that this book works, but I feel like it could have been a booklet instead of being a book (could have cost less lol) because the case studies were very similar to each other.
I love the ML system design interview book - it has a greater variety of case studies and topics that are asked in interviews, so both books in combination are super helpful.
The authors took the following approach: 1. Take a paper about some Gen AI features, like text to image 2. Frame it as an interview question 3. Summarize a paper. 4. Rinse and repeat. There were a couple of chapters, e.g. on RAG, which were more or less relevant for a system design interview question, but majority were absolutely generic and high level.