Demi Marie Obenour

153 Followers
139 Following
3.9K Posts
Software developer and security researcher. Currently working on Spectrum. Follows are not endorsements.
PronounsShe/her
GitHubhttps://github.com/DemiMarie
Matrix@alwayscurious:matrix.org

A problem someone should probably be working on: high quality high performance stepper motor drivers (ala TMCs) as open source firmware to run on dirt cheap commodity MCUs.

Both supply chain instability and fascist regimes wanting to restrict ordinary people's access to CNC machinery (including #3dprinting) risk making it hard to get commercial drivers, and they're overpriced anyway. But general purpose MCUs and mosfets are easy to get.

@whitequark you did not hear about this?

they trained on everyone's data and then started making suggestions in an "expert review" feature

so, for example, "Ariadne would say X about Y"

they did this to basically everyone who used grammarly who was notable

then they quickly backpedaled after they realized a lot of us are quite willing to sue

new blog dropped, grammarly has been excised from my workflow on this one and replaced with a libre alternative plus some homebrew tools I built to maintain tone and find sources using google news.

i will probably have to rework the latter because google news will probably become 100% AI slop soon. but whatever, smoke em while you've got em.

please let me know if there are errors.

I am a huge fan of machine learning for tasks that it is genuinely good at.

I am NOT a fan of the large LLM companies. At all.

Those two statements are not mutually contradictory. I want computationally efficient models trained on ethically sourced data that can be run locally.

Reminder for those who may not be aware that those "fancy/custom text" things using special unicode characters that bypass ASCII fonts to make your name look cool or fancy or whatever ruin accessibility, like hard.

They break screen readers hard, since most, if not all, don't know how to handle them properly and end up pronouncing something like "Special character S" or whatever. They're also significantly harder to read than a user's chosen font, or the default fonts on any reasonable operating system or website, especially for neurodivergent and in particular dyslexic people.

Please stop using them, and maybe nudge your friends to stop using them.

Boosts appreciated for awareness

#Accessibility #FancyFonts #Neurodivergent #Boostswelcome

RE: https://c.im/@cdarwin/116338321170772746

there is nothing like MAGA antiblack misogyny.

even their transphobia is rooted in their hatred of us Black women. lest we forget, fascists mainstreamed their transphobic hate with their campaign against Olympian runner Caster Semenya.

i have fuckwads sealioning about “it’s class not race” and i just don't have time for that ish. go elsewhere with your fuckery.

even with my ashy light-skinned ass, i have had to deal with this hate. ALL. MY. FUCKING. LIFE.

get rid of your racists.

Trans people aren't a debate. We're human beings. Take as long with that as you need.

How large should a standard library be?

I’m inclined to say “huge”, but I’m curious what others think. And my views are open to change here.

My biggest problem with the concept of LLMs, even if they weren’t a giant plagiarism laundering machine and disaster for the environment, is that they introduce so much unpredictability into computing. I became a professional computer toucher because they do exactly what you tell them to. Not always what you wanted, but exactly what you asked for.

LLMs turn that upside down. They turn a very autistic do-what-you-say, say-what-you-mean commmunication style with the machine into a neurotypical conversation talking around the issue, but never directly addressing the substance of problem.

In any conversation I have with a person, I’m modeling their understanding of the topic at hand, trying to tailor my communication style to their needs. The same applies to programming languages and frameworks. If you work with a language the way its author intended it goes a lot easier.

But LLMs don’t have an understanding of the conversation. There is no intent. It’s just a mostly-likely-next-word generator on steroids. You’re trying to give directions to a lossily compressed copy of the entire works of human writing. There is no mind to model, and no predictability to the output.

If I wanted to spend my time communicating in a superficial, neurotypical style my autistic ass certainly wouldn’t have gone into computering. LLMs are the final act of the finance bros and capitalists wrestling modern technology away from the technically literate proletariat who built it.

what we thought we knew about autism and... whatnot

'For decades, researchers had been measuring the wrong thing. Conflating communication style differences with empathy deficits produced dramatically inflated effect sizes and an illusion of empathy impairment'

#autism #actuallyAutistic #science #psychology

https://www.psychologytoday.com/us/blog/positively-different/202601/what-the-world-got-wrong-about-autistic-people

What the World Got Wrong About Autistic People

For decades, autism research compared autistic people to animals, denied them moral sensitivity, and assumed autistic traits made them miserable. All wrong.

Psychology Today