Discussion
Twitter is TESCREAL social media (for obvious reasons), which helps me see their talking points
1. Anyone pushing "stochastic parrots framing" shouldn't be taken seriously.
2. People "denying" the machine god narrative, like me, are akin to anti-vaxxers & climate deniers.
These are things they're saying to the left politicians they've captured, like Bernie Sanders who's being advised on "AI" by eugenicist doomsday cult leaders.
I wish we really were in a simulation like this people say 🙄
@timnitGebru Hot take amidst the governing class in the USA, but literate people can listen to cult leaders without buying into it.
@timnitGebru I'm not worried about ai becoming smarter than humans - I'm worried about humans becoming dumber than ai.
@timnitGebru "Captured" with money, right? I just want to make sure before I start doing my "talk to the hand" to all of the uncritical stans who have been making me ill since 2016.
@timnitGebru
I am ashamed of my own field, journalism, for its many failures. Atop them is how they fail us in this time of political -- and technological -- fascism.
I was slightly heartened to see Michelle Goldberg's column today. But I'm writing a post now lamenting that she -- and Cal Newport in his columns -- do not call on your and 'Emille Torres' work on the eugenics of TESCREAL.
Where is bothsidesing when we need it?
@timnitGebru I wonder how hard it might be for you to set up a meeting or get a call with Sanders. Politicians get lobbied by industry all the time, but less so by reasonable voices. Is he really completely surrounded by AI-bros?
Then again I don't know how influential he still is and if it's a good strategy or just a waste of time.
@timnitGebru what the hey happened to Bernie?
@timnitGebru clearly you're not doing real "ai safety" because you aren't considering the important question, "what if our product is too good?"
@timnitGebru while we are at it, can someone ban breaking points and their ilk from making commentary pieces on anything "ai" adjacent?
They're worse than Bernie in certain regards and reach.
@timnitGebru when i was a kid i read a book called The Blue Nowhere
it was a fine thriller, it had a lot of real historical tech but just as much made up hacksploitation fantasy tech
if somebody thought that book was somehow real, how could you convince them that it was all make up and ridiculous?
when like so much of it is obviously disprovable and ridiculous already
Why are so many people on the left falling for these billionaire con artists? 😞
It's what you would expect from the right who are seemingly blinded by net worth, but why would someone on the left be so gullible to accept their statements uncritically? What does AI offer them?
@FediThing @timnitGebru I would like to answer on "what does AI offer", taking the chance of Timnit Gebru involved here - I have read long stories with her involved, talking about so-called "AI companions", people seeking relationships with bots, with the poor man Michael who worked in that area, responding with romantic and dirty things to any stranger.
That's why I gladly and openly want to enter this debate with her involved and it is not just for AI, it's for everyday tech in general. I'm politically a progressive, and am very sad of where tech is going but no one offers me valid alternatives.
I won't put it on "left" and "right" political matter, as however, it's much more complex than this: I am blind from birth, and unfortunately am forced to use commercial/closed technologies (especially apple and microsoft) because there are a lot of concrete supports for people living my same condition. Maybe not created by apple/microsoft themselves, but assistive tech is created around those systems, mostly. I stopped using Linux in 2005 when I changed my Braille display and the software supporting command line then, BRLTTY installed on an old Debian, didn't have the drivers for my poor display, Alva Traveller. And I spent thousand euro of guys coming to my home trying to help me figure it out, with no success.
Then tried the GUI with its screen reader Orca, but it's still much, much behind comparing it to modern Windows software. Non-visual desktop access, NVDA, is open source and for Windows, it's a screen reader, too bad it's not portable to Linux.
That said, I found somehow hostility by part of those folks claiming that "linux is for everyone", "open source cares for inclusion", etc; the approach has always been, "it's open source, you can do it yourself, if you need something specific, develop it on your own". Like wanting to keep a restricted niche, a cult, rather than an open community. Not discussing things they give for granted on their own.
Now comes AI: there are a lot of blind-related assistive tech which have integrated AI for very precious resources: images descriptions, physical environment description especially on glasses, translate into simple text lots of user manuals which would otherwise be available dependant on screenshots only.
Not to talk about assisted coding, what marketing calls "vibe coding" with that meaningless word. It has allowed me to unlock many software features I couldn't reach myself, and overall, features that original apps' developers did not want ever to explore, with the same excuse: "we are a small company, I work alone, blind people are too few". All that bullshit that pushes you dangerously into far-right's hands.
Despite using it, I'm the first person who wants to keep guard high (and be able to defend myself and blind community) because this mechanism is always the same: they let you do things easily, in a quite affordable price - in fact most AI features designed for blind, are free or affordable. Then, I fear that within some years, they'll start to restrict.
Currently, the closest danger for us blind, comes from biases and moral judgments.
You cannot get description of sex-explicit material because AI policies do not allow them to be processed. Because someone in power has decided that sex is a taboo.
If there's a song title mentioning suicide, and that title is on-screen on a video, you can't get the exact title because it's obscured. Assuming that just a word will make you cut your own veins. Wtf.
I'm carefully following DAIR Institute, hoping to become one day or another, as an AI user and accessibility professional, part of a more ethical technical usage.
What scares me most, however, are those brain implants such as Neuralink and stuff. I can't trust something made by the same man who as first action when acquiring former-twitter, posted "I'm Elon Musket" with weapons next to him, and then fired all Twitter's accessibility team. You could spend money to have something implanted, then he can turn it off any moment if he thinks you're too progressive.
Thank you so much for this Elena, this is a really useful perspective which isn't represented enough.
@FediThing @timnitGebru It is not represented at all, and this debate about blindness, tech, AI and open source is also kept by other folks: @WeirdWriter who has created a content on his site about the so-called "disability dongles", those hype and inspiration-porn based tools which promise accessibility then aren't tested in real life at all. On his site - sightlessscribbles.com - he has very often expressed his opposition against AI monopoly. I don't share every opinion, but in a serious debate EVERY opinion should be considered.
Then there's @menelion (André) who thinks like me, about hostile environment in many open source fields. Many so-called "inclusive" thinkers who don't think about disability. And we also have Jonathan Mosen, @JonathanMosen who wrote a very long and interesting article on AI. He's like me, much more experienced than me, and talks about blind, tech, history and AI - https://mosen.org/ai
You know what scares me most from "lowest positions", I mean community members like me and us?
Extremism in both parties:
those being excessively enthusiastic for AI, who assume they get super-power because they can create some working code with natural language then distribute it with a narcissist instinct, with no long-term tests. It's from today a post, I shared on my profile, where someone discussed of a component making something work, but breaking something else already working.
And, the second scare, comes from the other extremists: those who compare AI users to murderers and perverts. I've been directly involved in such a discussion, the usual "AI ruins water, consumes energy, enslaves people" etc. Then when I explained my reasons, she said "it's like a murder, I can't tolerate it for any reason".
Ethics must cover everyone, these extremists are manipulators, they are not needed into our communities IMHO.
Install bonfire.cafe
Get the full app experience
@timnitGebru
Being in a simulation is only helpful if we can find a buffer overflow or other exploit/sandbox escape!