Software Development at 800 Words per Minute

His comments on VS Code reminds me of the quote "Good accessibility design is mostly just good design":

> Consistent keyboard shortcuts across all features, and the ability to jump to any part of the interface by keyboard

This is something I notice and appreciate about VS Code as a fully sighted person. Just like I appreciate slopped sidewalk cutouts when I'm walking with luggage.

A11y is a big commitment and cost, and of course not all a11y features benefit everyone equally, but it has a larger impact than most people realize.

MS has to make accessibility a priority because it's mandated by government (its customer).

Smaller companies would benefit from better libraries and design systems that make it easier to incorporate accessibility. Make accessible the default.

I was referring to how synthetic speech always talk in exactly the same way down to inflection and pauses whenever it encounters the same phrase, which isn't how people talk. So this helps a lot with comprehension. Structure of the content does help as well.

I've been using Optimal Recognition Point (ORP) style apps (I use Balto Speed Reader) to comfortably read at 800wpm or more - for fact books and materials I can keep this up for quite some time - though for fiction I find it too fast to parse all the characters and tones of voice (https://www.researchgate.net/figure/Spritz-display-The-Optim... )

The fact that the speed of information is consistent across languages makes it unlikely to have a speedup as described by the article

> Casually dropping that everyone is speaking so slow. That you must use the time between sentences for something meaningful, pretty funny. :-)

Didn't mean it that way . But that is truly the only way that I can use the computer while in a meeting.

Great post. When you're reading a Mermaid diagram, do you just happen to have memorised that "dash dash greater than" means "arrow"? I assume the screen reader doesn't understand ASCII art.

And how painful is reading emails? HTML email is notoriously limited compared to HTML (and CSS) in the browser, but it's pretty hard to add structure to a plain text email too. How annoying is it when I do so using e.g. a "line" made out of repeated dashes?

Great article. I was of course surprised to learn that it's possible to learn to understand the super-fast TTS, since videos and podcasts start to get very tough to follow around 2.5x and higher. I've been wondering: surely better algorithms for generating high-speed speech are possible, especially as we have more and more compute around to throw at it. It's not easy to search for, since "speed" for most tools is about speed of generation rather than wpm. As normal-speed neural net TTS models get incredibly good, I am hoping to see more attention paid to the high-speed use case.

Thanks for the blog post!

I was wondering what TTS voices you use? I've heard from other blind people that they tend to prefer the classic, robotic voices rather than modern ML-enhanced voices. Is that true in your experience, too?

Really liked the article.

The interesting part for me was that you can recognize synthetic voice much faster than human speech. Is there a specific voice you are using for 800wpm or it can be any TTS? Also, I think older voices sound more robotic that the newer ones (I mean pre AI, like the default on android is newer for me). Is there a difference for how fast you can listen to the newer more nicely sounding ones or the older more robotic ones?

Yes, that is how I usually consume my content. Cognitive load is actually lower for unstructured prose compared to code, think about fiction for example. Code is much denser.

When I read to relax, it is for enjoyment, so I don't aim to read as fast as possible. This is why I still listen to human narrated audiobooks, since a good narrator adds to the experience.

FYI in case you didn't read the article - it's not about LLM coding but about a blind software developer who uses a screenreader at 800wpm to read code. It's really astounding to hear how fast that is I recommend checking it out!

I can imagine it's practice and self-disciplilne; pick and know a language/tool instead of reinventing the wheel or doing research first, write like you've always written, do new projects in that same toolkit multiple times, and (self discipline), just get on with it.

I struggle with that tbh, every time I'm on a new project I get into a "beginner's mindset" and look up the basics for a tool again instead of trusting myself that I know enough and what I write will be good enough.

Thanks for reading :)

It was quite likely this one: https://www.vincit.com/blog/software-development-450-words-p...

Sadly the audio samples are gone. I'll need to pester someone to fix that.

It looks like it's not the article you were looking for, but there was also this similar article about one of the original authors of NVDA, who hits 900 WPM, also including an audio sample: https://www.9news.com.au/national/the-secret-language-only-b...

It's a matter of practice and progression. I listen to podcasts and audiobooks in English at 3x speed, but it took me maybe 2-3 years to get to this level.

I understand all casual and technical content just fine. The only thing tripping me is fiction where I struggle with character names that I don't know how to spell (my visual memory needs it!). That's an English-only problem though. I don't have any of those issues with content in my native language (Bulgarian).

Uh, the author is visually impaired. So they really have no other choice than "really comprehending what's being said."

I can read at speeds higher than 800wpm and comprehend everything. It's about practice. Although, yes, simpler material is easier to read at higher speeds.

I don't know exactly what top speeds I used, above 2.5x, under 5x, but I listened to some lectures on youtube as a background process to doing some work. I would be sure I have missed something, as I worked so I had this practice of just setting the video time back and checking if it is stuff that I already know. I was amazed to see that mostly all data was retained already and I had the feeling of listening to the same stuff twice.

These speeds were already non-comprehensive to my colleagues, but I thought there is no point in trying to get used to higher speeds, and to see the author processing these kinds of flows of data is just inspiring and amazing, Ill try to get better. But I am not planing on listening to code :D

I'm sure being blind stimulates developing your remaining senses a lot.

Tbh even closing your eyes should.

From the article:

    I adjust its speed based on cognitive load. For routine tasks like reading emails, documentation, or familiar code patterns, 800 WPM works perfectly and allows me to process information far faster than one can usually read. I’m not working to understand what the screen reader is saying, so I can focus entirely on processing the meaning of the content. However, I slow down a little when debugging complex logic or working through denser material. At that point, the limiting factor isn’t how fast I can hear the words but how quickly I can understand their meaning.

I think it’s about different modes of operation. If I’m skimming an academic text, it’s not that I’m reading faster, I’m just jumping between text to find interesting keywords and sentences. A way to pick out raisins from the cookie.

When listening to a podcast or reading a book the pleasure is important, so I almost never speed up. If the pacing (or information density) is too low, I just don’t listen. On the contrary I’ve read books that are so dense I have to slow down and repeat. Those can be great works.

One mode is for navigation, one is for embedding the brain in the story, the knowledge, or whatever it is. It’s like looking out the window on a train, I’m not thinking ”wow I have seen these cows for 1.3 seconds, what a waste of time, I could have processed them in less than a second”.

Not to speak for the blind, but I assume an enormous utility need of navigating and structuring information from a linear medium into whatever is the representation in our brains.

Edit: typo

I didn't post this onto HN, so I only just found out about this thread.

Thanks for mentioning the margin issue, I've tried fixing it now. Let me know if its still an issue.

> I wonder how quickly ChatGPT replaces much of the customized tools here?

Probably not many. It is prone to Hallucinations, and the latency involved for getting a response means that I only use it when I have to.