[greenmana]

"This tool won’t be appropriate for all audiobook listeners. For many, the silences matter, and removing them would degrade the quality of the book. For many, listening at 2.5x would also degrade the quality of the book. So use this tool with caution."

It's nice that this was also brought up. At least for me, even though I don't listen to audiobooks, it's too easy to become too "performance-oriented" when reading something. Meaning that ticking the box "achievement unlocked, read another book" becomes more important than what you really get out of the book. Did you now retain as much information as possible? Did you savor the work of art or just gulp it down? Ironically, if you don't use tools like these carefully, listening to books may end up being even a bigger waste of time.

[nine_k]

I would say that to have read a text is not the goal, it's a mere implementation detail. The goal is to understand ideas which the text tries to convey.

Some texts, and usual human speech, are low-density, high-redundancy. If you can easily understand the ideas they carry, you can listen at a really high speed and still don't miss anything, and later be able to adequately recall the ideas.

Some texts are either higher-density, or talk about concepts novel enough for the listener, and it takes some time to unpack the ideas and commit them to memory. Such texts cannot be read too fast, and sometimes require going back a sentence, or a paragraph, and re-reading.

An ideal (from the efficiency standpoint) audio book device would have a speed dial to control the level of adaptive compression, like described by the linked page. It would also have controls to jump back a sentence and replay it slower.

But such device is unlikely, because I suppose that the main audience of audio books is commuting drivers, who need to keep their hands on the wheel, and their eyes on the road, so they can't properly read.

And if course I suppose that the speedup and compression are barely applicable, or need a different approach, to books read with some artistic expression (that us, most fiction worth reading).

[benj111]

So. Assuming we are talking about non fiction. Why would you read a book that you don't really expect to learn anything new from?

Further, your supposed use case of audio books is when you're driving, when most of your attention should be spent... Driving.

As for fiction, efficiency totally misses the point. Efficiency isn't reading Harry potter as fast as possible. Efficiency is not reading it at all.

My most charitable interpretation is that we just have different aims when it comes to consuming literature, but that still leads back to the ops point that this is basically the gamification of book reading.

[throwanem]

> Why would you read a book that you don't really expect to learn anything new from?

Entertainment! A sufficiently wrongheaded idea advanced at length and with a measure of braggadocio can be funny as anything. Startup-culture biens-pensants are often good for this, I find.

That said, in the general case you're not wrong, and it should come as no surprise to anyone familiar with Goodhart's law that the gamification of reading produces perverse outcomes.

[clnq]

I wish we just didn't write low-density, high-redundancy non-fiction anymore. Especially books written in the "self-help" style where about 1% is the core message, and 99% are illustrations, examples, author's opinions, and similar.

That 99% is very easy to find online in the age of information deluge, it's probably where the author sourced it. Before the 00s, this content in a book would have been helpful. But these days, it makes sense that most people want to skip it or 2.5x it.

I used to carry a reporter notebook and take notes all the time I was listening to audio books. This made it very obvious that there's usually not a lot of information to extract from business books, leadership books, or any other kind of "coaching" books. Now I try to buy the summary variant of these books. Because ultimately, I was making my own summaries in a very time-wasteful way.

[duckmysick]

To make your own summaries you need to understand and think about the source material. I get more value from reading the book and analyzing it rather than glancing at the too-long-didn't-read version of notes somebody else wrote.

For example, the entire Checklist Manifesto book can be distilled to one sentence: break down crucial procedures into checklists and follow them. That's the core message. The 1%. On its own it means nothing to me. I can re-read it hundreds of times and it won't change my life. I'll recognize it but I won't understand it.

I need time to process my thoughts. Let the ideas marinate in my brain so to speak. I won't get much out of a book if I read it cover to cover in one sitting either.

My way to study a subject is to read not just one book, but 2 or 3 more from different authors. This way I can compare their opinions, find common things, see where they disagree, find the answers to questions I wrote down during my first read. Then I can take those ideas and make them my own.

Yes, it takes time. There's no shortcuts when I want to understand something.

As for the 99% filler content - I don't have to read it. When I read I ask myself "What's the goal of this section? What does the author try to achieve?". Often, I find a good reason. If not, I'll skip it. I can always return to it later. I have that option. I can't do that with summaries because they are lossy compression. I can't get more out of them because it's not there.

Sure, I can do my own research and look for it on the internet or interview relevant people. But at this point, I might as well write my own book.

[jasode]

>audiobooks, it's too easy to become too "performance-oriented" when reading something. Meaning that ticking the box "achievement unlocked, read another book" becomes more important than what you really get out of the book. Did you now retain as much information as possible?

You're expressing a common skepticism of 2x+ speedup but it misunderstands why people do it. It actually improves the presentation to 200+ wpm so the brain is more receptive to learning. The slower ~100 wpm (i.e. 1x speed) acted as a barrier to learning. I tried to explain the 2x+ advantage previously and how blind people had utilized this technique years before the general public did:

https://news.ycombinator.com/item?id=29661550

https://news.ycombinator.com/item?id=29660750

https://news.ycombinator.com/item?id=29665690

The motivations for speedup mostly applies to non-fiction audiobooks. For fiction, you may deliberately leave it at 1x because you want the deliberate pacing of the voice acting in addition to the info conveyed by the bare text.

Audio playback speed adjustment is just a tool. If you're a musician trying to learn a complicated guitar lick or drum fill, slow the music down to 0.5x or even 0.25x if that helps unlock it. If you're listening to Lord of the Rings, then playing it at 1x seems very reasonable. If you're listening to a speaker discussing inflation for an hour but keep tuning out because he talks too slow, speed him up to 4x if you have to so the information is presented at the higher speed your that brain prefers receiving it.

EDIT to replies to clarify "The slower ~100 wpm (i.e. 1x speed) acted as a barrier to learning."

This statement in isolation looks like an absolute science claim that 1x causes "zero learning" but of course that wasn't what I was saying. The context is "barrier to _optimal_ learning" for that particular reader. If 1x is too slow and makes some readers not stay mentally engaged, or abandon audiobooks, or leave podcasts/lectures in the queue and never listen to them, that's the "barrier to learning" that I'm trying to convey.

Likewise, if one needs to slow it down to 1/2 speed to comprehend difficult-to-parse text, that also meant 1x was too fast and also a barrier to optimal learning.

[spidersouris]

> The slower ~100 wpm (i.e. 1x speed) acted as a barrier to learning.

Any papers proving your point?

[RcouF1uZ4gsC]

> The slower ~100 wpm (i.e. 1x speed) acted as a barrier to learning.

Unless there is a lot of good evidence, I am skeptical of this claim

Human children are exposed verbally to 1x human speech. Do we really think that making teachers talk faster will improve learning and retention?

Also, according to linguistics, I believe pretty much human languages transmit close to the same bit rate (some languages have longer more descriptive words, some have shorter words, but by and large they average out).

Throughout our evolution, we have been exposed to 1x speech.

My guess would be that are brains don’t have a learning block to 1x speech.

[RobMurray]

It maybe doesn't come naturally to most people to be able to take in information from speech at high speeds, but it is definitely possible to learn. I am blind, and use a screen reader. My screen reader's voice is many times faster than normal human speech. (I don't know exactly how much faster, but most people can't understand a word of it.) I also listen to non-fiction podcasts and books at 1.5-2x speed, although I almost always listen to fiction at normal speed unless the reader is painfully slow.