News for hackers. Stuff that matters.
Re:Daniel Kahneman has died by voisin on Wednesday March 27, 2024 @12:25PM (#39841269)
Because most nonfiction books are one, relatively small set of ideas (profound or not, novel or not) that could be concisely written as a few blog posts or a single long form article, but in order to monetize and build the author’s brand, get exaggerated into a full book. It is really painful and something I hope GPT will help the rest of us cut through in the future (“summarize the points. Dig into that point. What evidence is given? Etc etc etc” as a conversation rather than wasting 30 hours reading noise for a book)
Re:Daniel Kahneman has died by geodel on Wednesday March 27, 2024 @01:05PM (#39841733)
Well, life of most people on the earth is just same boring repetitions with few novel events. So I wonder what would people do with their ample amount of saved time thanks to ChatGPT. Perhaps writing another Javascript framework, launch new food delivery apps besides raging on social media.
Re:Daniel Kahneman has died by olvy0 on Wednesday March 27, 2024 @01:02PM (#39841690)
That's exacerbating the original environmental problem, in addition to thick paper books, filled with filler material just to promote the author's brand, you now want to waste electricity on running an LLM that will give you a the short version? That's.... short sighted.

This should be dealt with by pressuring the publishing industry not to inflate books and fill them with fluff. This could be done by not buying these kind of books, and publicly shaming publishers who engage in this behavior. It's easier in non fiction books since the amount of fluff in fiction books is a more subjective matter.

Re:Daniel Kahneman has died by mattmaroon on Wednesday March 27, 2024 @04:32PM (#39844248)
They do it because a short book looks like a pamphelet and nobody will buy it. Most Gladwell books could easily be 30 pages, but nobody will pay $14.99 for that.

You can't shame them into buying books they can't sell.

How much electricity does it take an LLM to summarize a book? I'm sure the carbon emissions involved are trivial, and if they aren't, I've always been of the belief that (like eating meat) people are going to do what they want to do regardless of the environmental cost, so it's better to focus your ire on reducing the environmental cost. The problem here isn't using an LLM to summarize a book, it's that we've got a power grid fueled mainly by fossil fuels. (That is a problem that will fix itself in no time anyway now that renewables are cheaper and the gap is widening.)

Re:Daniel Kahneman has died by BeFlatXIII on Wednesday March 27, 2024 @05:11PM (#39844648)
> They do it because a short book looks like a pamphelet and nobody will buy it. Most Gladwell books could easily be 30 pages, but nobody will pay $14.99 for that.

This applies equally to Sci-Fi and fantasy doorstopper novels. At least those have interesting filler—sometimes even better than the main story.

Re:Daniel Kahneman has died by flawn on Wednesday March 27, 2024 @04:24PM (#39844158)
The latter will happen already by more and more people doing ways to summarize books
Re:Daniel Kahneman has died by CuriouslyC on Wednesday March 27, 2024 @01:32PM (#39842136)
Except that a lot of people read these books for the entertainment value of the anecdotes, and a lot of people enjoy feeling self-important for having read long books.
Re:Daniel Kahneman has died by smcin on Wednesday March 27, 2024 @09:18PM (#39846710)
Not just self-important, but the anecdotes are designed to be bite-sized to be easily retellable, like the 2016 NYT Abraham Wald/ beware survivorship bias on graphing bulletholes on WWII airplanes viz which locations were more damaging.

So you get a decade of "business nonfiction" bookstand fodder which is claims with high shock quality but often not relicable, mixed with anecdotes, to signal the reader's self-declared learnedness.

Combine this with the long-tailed nature of the statistic "the average American only reads one book a year" and it's frightening. I think Scott Galloway said sales of his e-book are >10x more than his paper book.

And during Covid lots of people got exposed for having fake bookshelf backdrops.

Re:Daniel Kahneman has died by mikhailfranco on Thursday March 28, 2024 @11:14AM (#39852605)
There was one Covid Zoom podcast (about economics?) with a middle-aged man sitting in front of a large set of bookshelves, housing row-upon-row of gnarly old books. The only recently-used book was lying flat on top of the other books.

Can you read the title?

https://pasteboard.co/imwZw8fcSpLy.png

Re:Daniel Kahneman has died by owisd on Wednesday March 27, 2024 @12:36PM (#39841398)
You can just leverage the “second brain” crowd — for every vaguely well-known non-fiction book someone has written up a summary for themselves and posted it on their blog.
Re:Daniel Kahneman has died by smcin on Wednesday March 27, 2024 @09:19PM (#39846721)
Sure, but be sure to check they're vaguely objective, not just shills, compensated reviewers or affiliates.
Re:Daniel Kahneman has died by ratg13 on Wednesday March 27, 2024 @12:39PM (#39841444)
Most people don’t absorb concepts immediately with only a simple explanation.

Unless you are reading a topic you are already familiar with, reinforcement of an idea helps you to examine a concept from different angles and to solidify what is being discussed.

If everyone fully absorbed and understood everything they read, schooling could be completed years in advance.

Re:Daniel Kahneman has died by 1letterunixname on Wednesday March 27, 2024 @04:22PM (#39844128)
Apart from CYOA, I think a hypertext "book" would be the most effective way to dynamically scale detail preferences without resorting to skimming or skipping.

Some people want more or less evidence based on their level of skepticism or critical thinking, while others want more evidence to reinforce the soundness of their inferred position on a topic, especially if it's a topic unfamiliar to them. Other people are under time constraints and just want the key points and a brief presentation.