10 ms·

If it works well, it could be actually very useful for longer pieces. I have read so many books with some actually good ideas hammered for +200 pages, (just to

If it works well, it could be actually very useful for longer pieces. I have read so many books with some actually good ideas hammered for +200 pages, (just to justify the cost of printing, satisfy the industry standards, or whatever) A half decent summary of all those would be of actual value. Get 80% of the value in 20% (or less) of the time.

your_challenger 4y ago
I could expand the product to handle such large content. Can you link me to a large content like that?
- jamager 4y ago
  All copyrighted material, I don't have links to the actual works, but some books that come to mind that I would have enjoyed much more if their were 20% of pages: - The courage to be disliked (Ichiro Kishimi) - The simple path to wealth (J.L. Collins) - Peak (K Anders Ericsson) - Happiness (Matthieu Ricard) - Clean Code (Robert C Martin)
- geoelectric 4y ago
  Not that they'd be the particular books OP wants, but if you're looking to summarize large content, perhaps grab it from Project Gutenberg? https://www.gutenberg.org/ https://www.gutenberg.org/
onlyrealcuzzo 4y ago
Isn't this available for most books already? I think the problem is - there's examples and anecdotes and whatever scattered throughout the book that make those ideas connect for you. And this is different for everyone. Maybe an ML you train yourself on highlights would be able to find the stuff that will connect for you - but I'm skeptical enough people read & highlight enough to train ML models to do this (or if it would even work).
- jamager 4y ago
  Yes there are a number of services, and I would happily pay for them. But either they have a very small catalog, or their summaries are too short, or both.
jerrygoyal 4y ago
doesn't blinkist do that
petters 4y ago
Would be difficult since GPT-3 uses a low number of tokens compared to a book. Around 8000 I think? Could possibly be done by iteratively summarizing section by section but that would give suboptimal results.

your_challenger 4y ago

I could expand the product to handle such large content. Can you link me to a large content like that?

jamager 4y ago
All copyrighted material, I don't have links to the actual works, but some books that come to mind that I would have enjoyed much more if their were 20% of pages: - The courage to be disliked (Ichiro Kishimi) - The simple path to wealth (J.L. Collins) - Peak (K Anders Ericsson) - Happiness (Matthieu Ricard) - Clean Code (Robert C Martin)
geoelectric 4y ago
Not that they'd be the particular books OP wants, but if you're looking to summarize large content, perhaps grab it from Project Gutenberg? https://www.gutenberg.org/ https://www.gutenberg.org/

onlyrealcuzzo 4y ago

Isn't this available for most books already? I think the problem is - there's examples and anecdotes and whatever scattered throughout the book that make those ideas connect for you. And this is different for everyone. Maybe an ML you train yourself on highlights would be able to find the stuff that will connect for you - but I'm skeptical enough people read & highlight enough to train ML models to do this (or if it would even work).

jamager 4y ago
Yes there are a number of services, and I would happily pay for them. But either they have a very small catalog, or their summaries are too short, or both.

jerrygoyal 4y ago

doesn't blinkist do that

petters 4y ago

Would be difficult since GPT-3 uses a low number of tokens compared to a book. Around 8000 I think? Could possibly be done by iteratively summarizing section by section but that would give suboptimal results.