User avatar
🇨🇦Samuel Proulx🇨🇦 @fastfinge@interfree.ca
Admin
completely blind computer geek, lover of science fiction and fantasy (especially LitRPG). I work in accessibility, but my opinions are my own, not that of my employer. Fandoms: Harry Potter, Discworld, My Little Pony: Friendship is Magic, Buffy, Dead Like Me, Glee, and I'll read fanfic of pretty much anything that crosses over with one of those.
keyoxide: aspe:keyoxide.org:PFAQDLXSBNO7MZRNPUMWWKQ7TQ
Location Ottawa
Birthday 1987-12-20
Pronouns he/him (EN)
xmpp fastfinge@im.interfree.ca
keyoxide aspe:keyoxide.org:PFAQDLXSBNO7MZRNPUMWWKQ7TQ
User avatar
🇨🇦Samuel Proulx🇨🇦 @fastfinge@interfree.ca
1mo
@Tamasg I'm not actually having any of those issues. I'm having other ones, though. Like for me "notifications" is said "norifications". But "g" is said fine. As is ctrl and alt.
User avatar
🇨🇦Samuel Proulx🇨🇦 @fastfinge@interfree.ca
1mo
@Tamasg "Well they should just learn English. Idiots. They should stop being blind while they're at it." -- probably some big tech CEO, somewhere
User avatar
🇨🇦Samuel Proulx🇨🇦 @fastfinge@interfree.ca
1mo
@Tamasg Right, but they just used codex and probably included all of torch. There are onnxruntime versions available, and compiled versions in rust.
User avatar
🇨🇦Samuel Proulx🇨🇦 @fastfinge@interfree.ca
1mo
@Tamasg I've also been meaning to look into this. They advertise streaming TTS lag at faster than realtime on CPU, with 100 MS lag or less: github.com/kyutai-labs/pocket-tts?tab=readme-ov-file
User avatar
🇨🇦Samuel Proulx🇨🇦 @fastfinge@interfree.ca
1mo
@Tamasg It's possible to get Onnxruntime to be snappy with a correctly optimized model, though. Blastbay TTS uses it.
User avatar
🇨🇦Samuel Proulx🇨🇦 @fastfinge@interfree.ca
1mo
@Tamasg I suspect you're going to wind up training your own g2p model. Eloquence can already output its phonemes, so you'd just have to write a script to convert from eloquence to IPA, and then you could just make a bunch of training data.
User avatar
🇨🇦Samuel Proulx🇨🇦 @fastfinge@interfree.ca
1mo
@Tamasg Also some phoneme tuning: thread and threat sound almost identical.
User avatar
🇨🇦Samuel Proulx🇨🇦 @fastfinge@interfree.ca
1mo
@Tamasg The lack of prosody markers is a blocker, though.
User avatar
🇨🇦Samuel Proulx🇨🇦 @fastfinge@interfree.ca
1mo
@Tamasg Yeah, not quite what we need. But a step in the right direction...
User avatar
🇨🇦Samuel Proulx🇨🇦 @fastfinge@interfree.ca
1mo
Tagging @Tamasg odorediamanka600-source/FYLs-G2P: A lightweight hybrid G2P engine with less than 1.8M parameters and can be deployed on any devices (almost) github.com/odorediamanka600-source/FYLs-G2P
User avatar
🇨🇦Samuel Proulx🇨🇦 @fastfinge@interfree.ca
1mo
Openclaw Is More Fun to Fiddle With Than To Use: Of course I have to write about , the latest viral sensation: stuff.interfree.ca/2026/02/06/openclaw.html
User avatar
🇨🇦Samuel Proulx🇨🇦 @fastfinge@interfree.ca
1mo
@kaveinthran @Tamasg @nick It might be because the newer versions of the driver had settings that the older one didn't, and it's causing a crash. I'm really not sure; everything is different about the driver since I wrote that.
User avatar
🇨🇦Samuel Proulx🇨🇦 @fastfinge@interfree.ca
1mo
@kaveinthran @Tamasg @nick It only includes support for 64-bit NVDA, because gruut needs 64-bit.
User avatar
🇨🇦Samuel Proulx🇨🇦 @fastfinge@interfree.ca
1mo
@Tamasg @kaveinthran @nick Here, the version that used an ancient version of speech player was still on my file host. I won't be updating it, but if you want to have a play, here you go: share.interfree.ca/app/open/AX2qBbXiJ8D-2wZ6znjs4bv-nZAAAyQ1zkB-9FHwonCxUDQ?view=1
User avatar
🇨🇦Samuel Proulx🇨🇦 @fastfinge@interfree.ca
1mo
@kaveinthran @Tamasg @nick Unlikely. It needed, like, 700 megs of just Python dependencies. And loading all that into NVDA slowed the entire thing down hugely.
User avatar
🇨🇦Samuel Proulx🇨🇦 @fastfinge@interfree.ca
1mo
@sclower I suspect it would work better if I wrote the project requirements document myself. Then fed it into the AI for implementation. Maybe it would shift me from thinking "What do I want coded" and towards "What are my requirements". I tend to start thinking too low level, about what libraries I want to use and what language, when I still haven't one hundred percent laid out what I want to do. The example that prompted me to post was reformatting a set of files. If I'd have thought just a bit harder about what I actually needed, I would have ended up with three regular expressions and a batch script, not a C# DotNet app.
User avatar
🇨🇦Samuel Proulx🇨🇦 @fastfinge@interfree.ca
1mo
@x0 I haven't looked into it, yet. But maybe?
User avatar
🇨🇦Samuel Proulx🇨🇦 @fastfinge@interfree.ca
1mo
Honestly the biggest problem with coding is that it lets me build things without having to think about what I want to build. It's happened twice now where the AI coded something perfectly, only for me to realize when I had exactly what I asked for, it wasn't what I needed or really wanted. I usually realize this during the act of building, and the act of building often helps me clarify what it is that I actually want.
User avatar
🇨🇦Samuel Proulx🇨🇦 @fastfinge@interfree.ca
1mo
Today I learned that Google Gemini can understand ASCII maps like in nethacks or other roguelikes: Route to the Stairs
From your current position [@], follow these steps:
1.
Move East (Right): Move one room over to a cleared room [#].
2.
Move South (Down): Move one room down to another cleared room [#].
3.
Move South (Down) again: This will take you directly to the room containing the stairs [>].
User avatar
🇨🇦Samuel Proulx🇨🇦 @fastfinge@interfree.ca
1mo
@mcourcel @jaybird110127 They are! And I went with a bank of interconnected EcoFlow's.