devinprater

Holy crap does the visitors in realtime thing have to be a live region? Every sentence or two: "visitors in realtime party popper 8 weary cat face."

I'm in the weights! I've successfully been immortalized! Except... I don't podcast; I hate my voice. Guess the models are trying to tell me to podcast. And other people say I should podcast. But that takes so much darn time!

https://www.intheweights.com/p/devin-prater

Can confirm.

github.com/devinprater

Meh, I wonder how many Claude commits iOS, MacOS, or Windows has that we just don't see?

Mine works on a Super Dee Duper AI!

There's my dopamine hit for the year.

I'm trying to get out of this. I'm a blind person. No one in tech goes to bed thinking about us, as it were. So, as a non-programmer, vibe coding accessibility fixes was an outlet to the daily million papercuts of using operating systems built by people who cannot understand me.

Well, I have barely anything to show for months of this. I made Termux more accessible on Android, made an MUD client for Emacs, fixed up some Emacspeak stuff because it's been abandonned going on 3 years now, and Emacs packages wait for no one, and tried added Grade 2 Braille entry support to BRLTTY. That failed because depression sucks and who would even use this vibe coded junk anyway.

The more open nature of Android made it rather easier. How far behind in features TalkBack is compared to VoiceOver, besides AI image description, made it feel like trying to heal a broken arm with pain pills. So I'm trying to tell myself that I can't fix everything, and that it's not my fault if other people, and companies, choose to not consider accessibility. I mean I can't help Google if they choose to not be helped.

Ah well, Global Accessibility Awareness Day is this Thursday. Maybe Apple will finally announce LLM image descriptions, and hopefully my iPhone 16 will be good enough for them because I can't afford to upgrade in this economy.

Oh my. I may not want to know what selecting all and then pressing Delete would do to you.

Wait, you mean there are senior software engineers who don't know how to use a *nix terminal? I've been using them since I was like 16 or so.

Ugh such overreaction. ADB is still a thing. Apple doesn't even have an official command like tool where you can just push an IPA to your phone. Goodness.

Integrating AI where it’s most meaningful, with craft and focus.

Spoken like a true AI.

A lot of these models struggle with small text strings, like "next button" that screen readers are going to speak a lot.

Lol last night, on a forked and accessible version of Termux I vibecoded into existence, on an Emacs and Emacspeak vibejiggered to work on Termux, I vibecoded, with gptel-agent, an Emacspeak package to make it speak when tool calls are being asked for by the model, and automatically speak any explanatory text after all the tools are called and edits are made. All on my phone with a Bluetooth keyboard. It's so easy, even a blind man can do it! :)

And because it's all controlled by me, I can tell it how to have the package speak, what it should ignore, and I'm not stuck with whatever some sighted person at some big company thinks a blind person wants. Everything should at most be open source, and at least be hackable.

All that to say, AI has helped me out a ton. Now I can be as productive as Emacs, and a Linux terminal, and maybe one day a Linux GUI with real Firefox and such, allows. And it would have *never* happened without AI.

So let's please do continue bringing on the AI. Make it smart and local, so I can have continuous AI descriptions right on my phone, with the ability to screen share or even agent-control my phone to get around inaccessible apps. Oh and fix AI app accessibility so the app sends output to screen readers when I type to it cause I hate talking to my phone and not every blind person wants to speak all the time. Ugh I hate that stereotype.

I'm glad I have chatGPT to turn that image with benchmarks into an accessible table lol. I like claude Code, but their accessibility in anything other than accidental CLI accessibility is frustrating. Try it. Load a screen reader like VoiceOver for Mac (cause I know most programmers use Macs) and go to claude.ai. In the "write your prompt to Claude" box, type something like "What will the weather be like tomorrow?" and press Enter/Return. Try closing your eyes for a good 30 seconds and within those 30 seconds, tell me how you'd know if a reply has been given by the model. Then try the same thing with ChatGPT. I would /love/ to be proven wrong.

Over the past month, with vibe-coding, I've:

* Made Termux accessible enough for me to use.

* Made an MUD client for Emacs.

Gotten Emacs and Emacspeak working on Termux.

Gotten XFCE to run with Orca and AT-SPI communicating to make the desktop environment accessible.

None of this would have happened without AI. Of course, it's only useful for a few people that are blind, use Android, and love Linux and Emacs and such. But it's improved my life a ton. I can do actual work on my phone. I've got Org-mode, calendar, Org-journal, desktop chromium, ETC. all on my phone. And if AI dies tomorrow, I'll still have it. The code is all there for me to learn from, tweak, and update.

I just use one agent, Codex. I don't do the agent swarms yet.

I'm completely blind. I like Linux. I've started to love Android since getting a Samsung and getting rid of OnePlus, cause accessibility. Termux is cool, but it's accessibility wasn't. So, I had Gemini rangle it up a bit into my fork of Termux [1]

Now it reads (usually) only newly incoming text, I can feel around the screen to read a line at a time, and cursor tracking works well enough. Then I got Emacs and Emacspeak working, having Gemini build DecTalk (TTS engine) for Termux and get the Emacspeak DecTalk speech server working with that. I'm still amazed that, with a Bluetooth keyboard, I have Linux, and Emacs, in my pocket. I can write Org and Markdown, read EPUB books in Emacs with Nov.el, look at an actual calendar not just a list of events, and even use Gemini CLI and Claude Code, all on my phone! This is proof that phones, with enough freedom, can be workstations. If I can get Orca working on a desktop environment in Termux-GUI. But even with just Emacs and the shell, I can do quite a bit.

Then I decided to go wild and make an MUD client for Emacs/Emacspeak, since accessible ones for Android are limited, and I didn't trust my hacks to Termux to handle Tintin++ very well. So, Emacs with Emacspeak it was, and Elmud [2] was born.

Elmud has a few cool features. First of all, since Emacspeak has voice-lock, like font-lock but for TTS, Ansi colors can be "heard", like red being a deeper voice. Also a few MUD clients have sound packs on Windows, which make them sound more like a modern video game, while still being text-based. I got a few of those working with Elmud. You just load one of the supported MUD's, and the sound pack is downloaded and installed for you. It's easy and simple. And honestly, that's what I want my tools to provide, something I, or anyone else who chooses to use them, that is easy to get the most out of.

None of this would have been possible without AI. None of it would have been done. It would have remained a dream. And yes, it was all vibe-coded, mostly with Codex 5.2 on high thinking. And yes, the code may look awful. But honestly, how many closed-source programs look just as bad or even worse under the covers of compilation?

[1] https://github.com/devinprater/Talking-termux-app

[2] https://github.com/devinprater/elmud

Nah, just Microsoft Copilot. No Os.

No way. There's just no way.

I hate that so much. When blind people are trying to start JAWS (the screen reader) by typing "jaws" into the start menu and pressing Enter, it will sometimes pull up a Bing page on Jaws the movie instead. And the blind person is just sitting there waiting for the screen reader to start. I tell people to use the run dialog for that reason. Sucks but that's what you have to do in the age of inshittisoft.

Linux is even getting more accessible. I'm thinking of Elementary OS which not only posted about their accessibility work, but linked to the articles which really fired things up. I'm a Fedora guy, mainly because I want the latest Orca, AT-SPI2 and such, so I don't feel like an Ubuntu dirivitive would work as well.

So I installed Fedora on my work machine and find that I can still get all of my work done. Well except the parts that require testing accessibility on Windows screen readers or helping with Windows-related issues.

The only thing I miss now are the many addons made for NVDA, especially the ones for image descriptions. But if I can get something to work with Wayland, I could probably vibe code some of them. Thank goodness for Claude Code.

A ton of studies colleges/universities/corporations do on blind people give gift cards as payment. Usually $20 or so for a good 40 minutes of time.

What are you trying to sell me again? :)

AI features you say? They could start by taking the alt-text generation model and making it browser-wide if blind people want it, not just for PDF's.

Wow, I never heard of that before today. Sega Genesis was my first console. I still remember the six button controller. It worked well for Mortal Kombat 3.

Can the tables have column headers so my screen reader can read the model name as I go across the benchmakrs? And the images should have alt-text.

Emacs and Emacspeak make me feel something. A lot of something. This kind of "playground" feeling where I can dive into a manual that's just sitting right there. The the entire Emacs is a manual. C-h m and boom, all keyboard commands for that mode are right, feaking, there. No hidden bullcrap, no patchwork HTML tables to drudge through, nothing. And if something doesn't work with Emacspeak, I can Codex it into working. Maybe. Enough to get what I want done, done.

Wow, just 32B? This could almost run on a good device with 64 GB RAM. Once it gets to Ollama I'll have to see just what I can get out of this.

Holy crap it even felt like the HN front page with my screen reader. I thought I'd clicked the wrong link until I read the LLAMA 12 and such.

I, myself, as a singular blind person, would absolutely love this. But we ain't there yet. On-device AI isn't finetuned for this, and neither Apple nor Google have shown indications of working on this in release software, so I'm sure we're a good 3 years away from the first version of this.