Notices by datarama (datarama@hachyderm.io), page 2

@mattly Satya Nadella *bragged* that with AI, Microsoft was moving from the bicycle paradigm to steam engines. He bragged about this in a world that is presently dying of fossil fuels.

@mattly There's one part of font design where I'd *love* for some ML to take over: Kerning tables. There is nothing in the world more tedious and uninspiring than kerning a font. I've tried to automate it myself in various ways (including neural nets), without success.

And here is the font design version of the dream of current AI trends: Make an AI take over all the fun parts of designing glyphs and making creative decisions, leaving only tedious soul-crushing kerning to humans.

I make fonts as a hobby. I've been using this one as my terminal daily driver for a while, and it's pretty much come to be just "what the terminal looks like" for me now.

Should I put it up in public somewhere?

(It will, of course, be scraped so someone can use it to make an AI font generator and extinguish a little bit more joy from life.)

@mattly @technomancy I don't really care about making money off them. My terminal font and book face I made just to have something nice to look at when using my terminal and setting various things in LaTeX, back when I was a teacher. And also to keep me sane during lockdown. My sans was a gift for my mother, for her 64th birthday. Economic value was never part of it.

I do care about not donating free labour to the AI barons.

@mattly @technomancy No - none of my fonts are publicly available. I keep meaning to do it, but there's always something I have to do before I'm satisfied.

I'm thinking about putting this up somewhere though. :)

(This is 2024, so it will be scraped and then some asshole will make an AI font generator to extinguish just a little bit more joy from life.)

@technomancy Mine is 4 years old and I made it myself. There's too much sunk cost in it now to ever change it, I think. :-)

@technomancy

Quick screenshot with a sample. My terminal apparently has janky rendering of the italics.

@mcc Short-term: I am personally lucky that I live in a country that has some fairly restrictive regulations on workplace surveillance. I'm sure that this will spawn even worse spyware, but there are upper limits to how much of it I can legally be forced to use.

Longer-term: The end-goal of all the AI training is of course to build the Mass Layoff Machine, and that's going to fuck people here over too, if they can just get enough data from people in countries without a history of strong unions.

@mcc I hate it so much. I don't even use Windows, but I know that the things they'll be building using this will be deployed against me nonetheless.

@mcc Remember when people were saying "you're not the customer, you're the product"?

You're not even that. You're the raw material for the product.

@mcc *Of course* it won't be kept local. Same with the "we'll listen in on your phone calls to check for scammers!" features Google is hawking.

"We're running out of high quality language data!" ... "It's a total coincidence that we've made this decision right now, but we've decided to listen in on every phone call / watch every software interaction anybody ever makes ever again".

They're after more training data.

@mcc I feel like an asshole when I say I enjoy (and used to make) "generative art" now.

@HauntedOwlbear @nota

I have a *very* hard time picking out a favourite species! There are so many to choose from. I also have a soft spot for Nile monitors, though Argentine black-and-white tegus occupy the same spot. And there's something humbling about the little, unimposing Eurasian viviparous lizards whose geographic distribution puts Genghis Khan's empire to shame.

My favourite individual lizard is an extremely sweet and friendly 16-year-old Northern blue-tongue named Igor.

@pinkdrunkenelephants @mcc In the EU, there actually is some legislation. Copyright explicitly *doesn't* protect works from being used in machine learning for academic research, but ML training for commercial products must respect a "machine-readable opt-out".

But that's easy enough to get around. That's why eg. Stability funded an "independent research lab" who did the actual data gathering for them.

@pinkdrunkenelephants @mcc That doesn't work if copyright *itself* doesn't apply to AI training, which is what all those court cases are about. Licenses start from the assumption that the copyright holder reserves all rights, and then the license explicitly waives some of those rights under a set of given conditions.

But with AI, it's up in the air whether a copyright holder has any rights at all.

@HauntedOwlbear @nota Also, given this: What *is* your favourite lizard? :)

@HauntedOwlbear @nota I don't know if I am currently outing myself as autistic, a herpetology nerd, an insufferable pedant or any combination thereof, but:

Tuatara are not lizards.

Where do people who self-host their websites, mastodon instances etc. actually host things these days?

(Hetzner are villains now, and so are DigitalOcean I gather?)

@drahardja @thomasfuchs Hanging out with its best buds, practical quantum computing and cold fusion.

@mattly Mine have been absolutely terrible. I've had three. The first I also thought was a heart attack; the second one had me crawling around the floor crying (my previous reference point for "terrible pain" was diverticulitis). But neither of them had complications beyond the pain.

The third one *did* cause complications (full-on bile duct infection), which is why I'm having it removed.