My first thought was '20 books on starting a profitable SaaS business that were all actually written by ChatGPT'.
I'm guessing if the 20 books are actually written by people in all seriousness and the belief that they understood and cared about the subject, they're probably still going to be better than what we'll have in ten years, which is 200 books all of which are echoing voids of ChatGPT.
It'll be a lot cheaper to produce such books and put them into the marketplace if you don't actually have to write, learn, or know anything: so much cheaper that it'll become untenable to write 'real' books.
Which mostly means 'marketplace' will stop meaning anything. Knowledge will still mean something but we'll have to arrive at it and disseminate it in different ways that don't resemble the commercial marketplace, as that will not stand up against the cheapness of AI labor.
I read a lot of startup books. The value is spending time in these books to really internalize them.
Summarization tools have existed for a while. They don’t take off because they can solve the core problem of deep learning. It takes time to reshape how you think. Reading some summaries won’t do that.
The one book I’ve “quizzed” GPT4 on, it gave sufficiently good answers to convince someone they understood the book but actually missed almost everything salient about it.
Not super relevant but this was Henry George’s Progress & Poverty, so a pretty well-known book with fairly detailed arguments.
Let's say I want to know better about X. You can submit these prompt to chat GPT.
- Can you explain me X in simple terms?
- Explain me different between the items 1 and 2 in your response and provide example?
- Could you show this in a tabular form? or tree from using markdown
with chatGPT you have access to elaborate knowledge in the most suibitable way for you. You can learn faster. If you prefer an information in a tabular form, you can do it. You can ask differences between concepts and get examples.
In a book you can only learn something written by the author.
I'm guessing if the 20 books are actually written by people in all seriousness and the belief that they understood and cared about the subject, they're probably still going to be better than what we'll have in ten years, which is 200 books all of which are echoing voids of ChatGPT.
It'll be a lot cheaper to produce such books and put them into the marketplace if you don't actually have to write, learn, or know anything: so much cheaper that it'll become untenable to write 'real' books.
Which mostly means 'marketplace' will stop meaning anything. Knowledge will still mean something but we'll have to arrive at it and disseminate it in different ways that don't resemble the commercial marketplace, as that will not stand up against the cheapness of AI labor.