An experiment to test if Bionic Reading helps you read faster(blog.readwise.io) |
An experiment to test if Bionic Reading helps you read faster(blog.readwise.io) |
You read 14% faster with Literata
But I also found I am a verrrry slow reader. My reading speed came to 147 wpm. Any recommendation to improve this?
read with the opendyslexic font.
Was there an ad or something?
One of the many issues I have with contemporary literate culture, another being those who insist on reading a large volume of books, on being “well read”, rather than focusing on quality.
But the comment made me think - can we even improve the comprehension, as in processing what we read? (not as in understanding the vocabulary/structure) I've never heard of ways to purposefully do that.
The idea of reading at a given speed seems wonky to me as well. Its all about information density and how deeply you need to understand something imo. I read light fiction pretty quick, non fiction slower, technical stuff much slower etc
An interesting experiment is to try reading a word at a time, but without moving the eyes. Here's [1] a command line program that takes text on standard input and displays it one word at a time in a fixed position on the screen, holding each word for N milliseconds (3N milliseconds if there is punctuation) where N defaults to 250 but can be set on the command line.
At the default 250 msec per word I found it very easy to read the material I tested with (random extracts from a Project Gutenberg edition of"The Valley of Fear", by Arthur Conan Doyle). That works out to around 160-180 wpm. (Not the 240 wpm you would expect from 250 msec/word because of the delays for punctuation).
At 200 msec/word, it still feels like a very slow read. Rate was around 270 wpm.
150 msec/word gave around 320 wpm. Still not a problem keeping up.
120 msec/word pushed it up to around 380 wpm and it starts to get hard for me. If I don't quite catch a word and have to think a little to figure out what it was I can get distracted enough to miss more words unless some punctuation comes up soon to give me a little break.
100 msec/word, around 480 wpm, is still reasonably comprehensible but at that point requires a lot of focus and feels tiring even though my eyes don't have to move. Sometimes not moving can be as tiring as moving when you are trying to not move for a long time.
I would not want to read a lot this way, but there are some places I wish it were offered. Many music players for example if the title does not fit in the space available autoscroll it back and forth. It can be very hard to read it while it is scrolling. A word flash display might work better there.
It. Was. Horrible. For. Me.
My eyes are used to processing much bigger chunks than a single word, and know how to move to the chunk size that they used. Therefore I topped out at a fraction of my usual reading speed.
The last company that tried to popularize this was Spritz, and I believe they are now owned/operated by their erstwhile VC. It probably didn't help that the proprietary bits they added were not compelling, so other companies just implemented vanilla versions and therefore didn't have to pay them anything.
Note that my natural reading speed is ~950 wpm, so I tend to ignore these fads. I read fast enough already. The main barrier to reading faster is that it gets exhausting having my brain constantly trying to catch up to my eyes.
Were you just born with 4-5x the average reading speed? Or did you employ some other 'fads'? This seems pretty astonishing. Particularly in light of the following sentence on the "Speed Reading" entry on Wikipedia:
>Cognitive neuroscientist Stanislas Dehaene says that claims of reading up to 1,000 words per minute "must be viewed with skepticism"
I am autistic and I have two unusual informational input traits.
I would consider myself a good reader. A sizeable chunk of my life has been dedicated to reading (and absorbing other forms of media). I can't read at a 950 wpm (that figure is immediately raising suspicions for me) but I'm still fairly fast.
The first weird thing:
I taught myself how to read and I've been reading since I was 3 years old. Apparently this freaked out my aunt when I was in her car reading street signs aloud. I still have memories of being far ahead of my peers in early childhood. School was unable to challenge me and this led to me having a lax study attitude and I became lazy. As an adult I'm still lazy, but I've been able to turn this into a strength as a programmer. (See Bill Gate's quote on "a lazy person").
For anyone else who has or knows someone who is experiencing this: HealthyGamerGG's video: "Why Gifted Kids Are Actually Special Needs" can give some great information to help understand this.
The second weird thing:
I regularly watch informational/tutorial/conference/etc... videos on youtube between 2-4 times their standard speed. I do this in the browser's console with the following command:
$('video').playbackRate = x;
Where x is a number. (e.g. 1, 3, 2.75, etc...)
After you've already typed it once, a simple press of the Up key will bring it back as if you had just typed it.
I've been told there are extensions that do this while avoiding the terminal, but this is already ingrained in my muscle memory. (F12 -> Up -> Delete -> type number -> Enter -> F12.)
Understanding sped-up talking is a skill I've built up over time. To other people around me who have tried to follow along it sounds like gibberish. I've heard of deaf developers who commonly develop this skill so that they're listening 600-800%+ standard speed but I don't think the upper range is possible on videos with different voices and accents using a wider vocabulary.
I did the test from the article, and got about 500 wpm with the standard font (about 400 wpm on the Bionic Reading thing). Even that speed is at the limit; I'm already skipping parts of the text that don't seem to contribute much (I could still correctly answer the reading comprehension quizzes at the end though).
I ignore all those fads too. This Bionic Reading font in particular feels completely backwards: it tries to make me read word by word which is needlessly slow.
Your results
Typeface Bionic Reading Literata
Article 1 591 WPM 352 WPM
Article 2 657 WPM 457 WPM
Average Speed 624 WPM 405 WPM
You read 43% faster with Bionic Reading
maybe some enterprising and bored HN nerd can make a Calibre plugin that converts regular epubs to bionic reading enabled epub files
----
but the truth is, for the aforementioned storybooks, often I just LOOK at a whole paragraph or even the whole page and just pick out the relevant word of two in a story.
Unlike actual educational content, where the exact text matters, in fiction, after having developed a hobby to read for entertainment for so long and having gotten used to so many tropes, I just often just breeze through and look for the word that confirms which direction the author is taking that paragraph in, and often just glide over the paragraph and go to the next one.
It's like learning how to drive, I guess. At some point you are not supposed to look at every thing, you just take the overall picture and just go.
Not sure what is achieved by making every road sign and billboard flashing neon is supposed to achieve in such a scenario. Not all words are worthy of equal attention, most are meant to be glossed over.
____
but for actual educational purposes, the best test IMHO is to use it for the boring but important texts, like training manuals. See if it actually helps people learn and retain more import information about the new tool or procedure they are learning about.
"Here's how we've designed the pilot study:
...
4. After finishing each document, we asked three multiple choice questions to control for and confirm comprehension."
Literata font with a peach background color to optimize reading speed and no selective bolding:
javascript:void function(){javascript:(function(){var a=Math.floor,b=document.querySelectorAll("p, title, a, ul"),c=[],e="",f="",g="",h=0,k=0,l="",m="",n=window.open("","_blank");for(var d in b){var i=b[d].textContent;i%26%26(c=c+"\n"+i)}for(f=c,e=f.replace(/\n/g," <br></br> "),g=e.split(" "),h=0;h<g.length;h++)k=a(g[h].length/3)+1,l="<span style='font-weight:lighter'>"+g[h].substring(0,k)+"</span><span style='font-weight:lighter'>"+g[h].substring(k,g[h].length)+"</span> ","."==g[h].substring(g[h].length-1,g[h].length)%26%26(l+="<span style='color:red'> </span>"),m+=l;n.document.write("<html><p style='background-color:#EDD1B0;font-size:40;line-height:200%25;font-family:Literata'>"+m+"</p></html>")})()}();
Selective bolding with a peach background for those who find a benefit:
javascript:void function(){javascript:(function(){var a=Math.floor,b=document.querySelectorAll("p, title, a, ul"),c=[],e="",f="",g="",h=0,k=0,l="",m="",n=window.open("","_blank");for(var d in b){var i=b[d].textContent;i%26%26(c=c+"\n"+i)}for(f=c,e=f.replace(/\n/g," <br></br> "),g=e.split(" "),h=0;h<g.length;h++)k=a(g[h].length/3)+1,l="<span style='font-weight:bolder'>"+g[h].substring(0,k)+"</span><span style='font-weight:lighter'>"+g[h].substring(k,g[h].length)+"</span> ","."==g[h].substring(g[h].length-1,g[h].length)%26%26(l+="<span style='color:red'> * </span>"),m+=l;n.document.write("<html><p style='background-color:#EDD1B0;font-size:40;line-height:200%25;font-family:Arial'>"+m+"</p></html>")})()}();
Curious what results people here got. I got 23% faster with BR but I couldn't tell if I was just trying harder.
I can't tell whether I was intentionally going faster because I had preconceived negative opinions about BR. It's a possibility.
My assumption is that I'm able to scan the words better with BR, but my eye gets pulled along too fast and I'm reading without comprehending.
(my results were 2% faster with BR)
i know a significant number of people here are probably already familiar with what bionic reading is, but still, could be an interesting control.
As the article says, a better comparison may be to have a fake version of BR as the control.
results are generally in the ballpark of insignificant, and there seems to be a very weak inverse relationship between reading speed and comprehension.
which is to say, on preliminary it looks like most "speed reading" claims (I.e. quackery).
what test do you use to compare variances though?
272 vs 272 wpm, English is not my native language and I just ran a 5k, not sure if that has any relevance
with Bionic Reading i had to re-read some sentences until i've processed them correctly
note that i'm not a native speaker
personal result:
Bionic Reading: 234 WPM
Literata: 262 WPM
"You read 11% faster with Literata"
My idea of hell. I'll wait for a better choice of source material.
Is it possible that people "calibrate" their brains towards a specific reading style, and that first style induces the calibration? Similar to how you can calibrate to quickly pick out a picture of a specific object, from a gallery of pictures, after you've prepared yourself to see that picture.
>This is likely because the first article was less challenging to read
I have thought about using Bionic or Dyslexic fonts but I'm worried that it'd cause a change over time that makes it even harder to read plain format text.
My reading speed overall is slow compared to others. I notice my reading speed is only fast and accurate when I'm extremely interested in a topic.
I guess I’m a pretty consistent slow reader.
I do read quickly but I have read that essay before, so perhaps I only skimmed it. I wasn't certain on one of the comprehension questions.
From their own metrics - speed went up by a lower percentage than comprehension dropped.
This sounds like a terrible idea...
Would it work with languages that have longer words and a case system?
I was horrified at how an attack piece had so many positive votes, and asked people why they were voting for it. It turns out that a lot (probably most) had just skimmed it, felt good, voted for it, and moved on. They literally hadn't seen the nasty things said about a variety of people.
But good point on hacking people's tendency to glide, I shall keep that in mind.
Also - It must be so painful for you to type, or heck even speak!
Also, I am not quite sure if I really understood the text... or used my knowledge of the views of the "tech" community, and just "guessed" which of the three mcq options paul graham probably meant?
It could be I just bullshitted my way into correct answers...Honestly, after a night's sleep, I don't recall much of what was said in those articles, so how much did I actually retain?
That said, I do find it painful to watch any content on less than 2x speed, but that maybe because I'm probably on the spectrum.... Things get complicated.
Edit: 2% faster with BR.
The jury is out as far as whether I'm also autistic, though. I have a great many traits in common with people who are, but I've also found a surprising amount of success reducing my more frustrating symptoms by a combination of working through early childhood trauma (which has helped my nervous system stop overreacting to stimuli) and extremely strict regulation of my diet (I have only eaten 10 foods, prepared in two ways, in the last year and a half) to reduce gut leakage and brain inflammation. (The diet is not expected to be this severe forever. Foods are reintroduced in stages. And it includes slow-cooked meat on bones and organ meats, so it's not as dire/risky a nutritional thing as it probably sounds.)
Because I had no language delay as a child and Asperger's was not a separate diagnosis until I was well into elementary school (and they certainly weren't looking for it in girls), I was never tested or diagnosed as a child. I also haven't identified enough advantages to pursuing a diagnosis as an adult to choose to pursue it now. But if healing my nervous system and my gut eliminates my symptoms (eventually), I'm inclined to think that would indicate I didn't have autism in the first place, just a cluster of overlapping symptoms.
> We recognize this is clearly not a gold standard, double-blind RCT. Researcher and participant alike knew what Bionic Reading claims to achieve. We simply did not have time to (1) develop a mock Bionic Reading placebo that, for example, might bold the middle of the word instead of the beginning and/or (2) find a set of test subjects who are blissfully offline and unaware of this viral sensation. That said, we controlled for what we could and we think we've already gathered some fairly compelling evidence.
As a kid I thought I was simply a slightly fast reader until I happened to take a copy of https://www.amazon.com/Clan-Cave-Bear-Jean-Auel/dp/060961097... to a bath. I started, got caught up with the story, finished it, and finished my bath. My mother was so astounded that she quizzed me to verify I had actually read the book, and then estimated how what my reading speed had to be.
The weirdest thing is that my brain really does play catchup. I told my mother to test me by picking random spots and reading a few sentences. I was then able to tell her what was going on, and when handed the book could find the passage. That I was able to do. But the book was still jumbled up in my brain - I couldn't have given a plot summary for a day or so.
I strongly suspect that this kind of "brain pipelining" with large amounts of buffering is critical for really fast reading. Get every slow step out of the loop, only do what's fast. Let the slow bits of your brain catch up.
So I like to read fast, and maybe there are some things I'm missing but I don't feel like that's a loss.
Mind you I'm not really speed reading (I think), I just read pretty fast. And skip uninteresting pieces of text (by skip I mean read some words or parts of sentences here and there, like the start and/or end of paragraphs, to know what the text is about so I know where to resume normal reading).
Now food is something else entirely :) I'm a pretty fast reader but a slow eater.
What I do isn't what most people call speed reading, though. I don't skip any words, and can recognize and pull out exact phrases. I don't know how much text I could hold like that. When I was 19, a long book. But I think my capacity has gone down with age.