iiinova — where your digital self keeps evolving

punished felix @punishedfelix.com · 19/01/2025

alright after 2 months of testing this site definitely sucks and I will be probably moving to something else

punished felix @punishedfelix.com · 18/01/2025

I didn't make any assumptions about what you do or do not know, I just gave you some research provided by my friend who worked in computer vision research. What research do you have? It might be interesting.

010

punished felix @punishedfelix.com · 18/01/2025

really getting tired of this 300 char limit when trying to explain technical shit

030

punished felix @punishedfelix.com · 18/01/2025

www.youtube.com/watch?v=wawM... just throwing this one in because its cool. Warren McCulloch was a neurologist who got the ball rolling on neural networks and computer vision decades ago.

youtube.com

Interview with Inventor of Neural Nets Warren McCulloch, neurologist who helped start it way back.

YouTube video by trwappers

010

punished felix @punishedfelix.com · 18/01/2025

www.nature.com/articles/s41... - very technical, but goes into detail about this relationship homepages.math.uic.edu/~lreyzin/pap... - May not be accessible without OCR, old PDF of early research into CV

100

punished felix @punishedfelix.com · 18/01/2025

Also, my friend wanted to provide some resources in the next reply for you to learn more about computer vision, neural nets and its relationship to vision and neurology, but they are fairly academic. I still think you might benefit from the exposure though. Ignorance is not bliss!

210

punished felix @punishedfelix.com · 18/01/2025

In fact, "effective altruism" is directly connected to Peter Singer. If you know that name you know him as one of the most ableist philosophers in recent memory, who has argued against disability theorists in favor of a destructive utilitarian ethics that systematically eliminates the disabled.

140

punished felix @punishedfelix.com · 18/01/2025

I agree that there is a problem with the new AI, but in my opinion its not the technology itself, it's how its being used, marketed and advertised. Don't forget that people like Sam Altman ascribe to a Silicon Valley cult called "effective altruism" that is deeply rooted in ableism and body fascism.

120

punished felix @punishedfelix.com · 18/01/2025

Part of the problem with the discussion surrounding AI is that we are talking about a lot of different technologies. Hell even Pokemon Blue has AI. Don't get caught up in the marketing. In order to solve this problem we have to dodge the sales lies and avoid being reactionary.

010

punished felix @punishedfelix.com · 18/01/2025

In fact I tried fucking around with CV libraries over 10 years ago. Although I wasn't experienced enough to do much with it yet. CV technology is everywhere and not particularly costly. Actually, more concerning is how this technology is abused in surveillance and defence projects.

100

punished felix @punishedfelix.com · 18/01/2025

Actually, that's not true! I don't know their exact implementation, but computer vision technology has actually existed for a long time. Way before the OpenAI crap you're seeing today. It actually works much more like how your eyes and brain typically process vision.

200

punished felix @punishedfelix.com · 18/01/2025

It's good that they didn't remove human calls though because that allows people to have preference. And it encourages an interaction that engages sighted people with blind people.

000

punished felix @punishedfelix.com · 18/01/2025

That's true. I will say that out of all possible use cases CV AI is probably most valuable here in comparison to bullshit like summarizing paragraphs or making images. But yeah. A lot of blind people in OurBlind confirmed they do use this feature and sometimes waiting for a person can be a PITA.

210

punished felix @punishedfelix.com · 18/01/2025

Of course I've considered other disabled people. I'm disabled myself & vocal about issues with visual processing that gives me migraines & other symptoms that robbed me of my career. I'm telling you to actually interact with other kinds of disabled people. And to read people's replies.

000

punished felix @punishedfelix.com · 18/01/2025

I asked around today in a popular blind server cuz I couldn't tell if I brain farted but multiple people told me it was a really helpful tool.

010

punished felix @punishedfelix.com · 18/01/2025

I think your post is fine but it's crazy that sighted people are just assuming because you made a post about proper alt text usage that computer vision AI isn't used by blind people without even asking. I am curious though have you found personal value in BeMyEyes's AI feature?

200

punished felix @punishedfelix.com · 18/01/2025

If you actually read my replies you'll find that I am not saying that alt text is bad and I'm pointing out legitimate use cases for AI. seriously. Look up BeMyEyes and it's implementation of computer vision technology. TALK TO FUCKING BLIND PEOPLE FFS.

100

punished felix @punishedfelix.com · 18/01/2025

Idk man accessibility is important to me

100

punished felix @punishedfelix.com · 18/01/2025

Also, it's easy to just talk to blind people instead of just quoting them. Check out r/blind. Twitter like interfaces are not ideal for discussions and I think it's essential to have two way discussions instead of treating blind users as accessibility PSAs

210

punished felix @punishedfelix.com · 18/01/2025

Only to have it ignored by someone who thinks they can make judgements on how something impacts people they literally cannot personally experience Like fuck dude just talk to blind people it's not that hard? And you'll discover so much more nuance than just alt text

000

punished felix @punishedfelix.com · 18/01/2025

If you read my OP you will observe that I am not talking about Bluesky's implementation and that I explicitly state that alt text is preferable I'm kinda irritated because a few months ago multiple blind people corrected me on this very issue and I felt like providing their nuance

200

punished felix @punishedfelix.com · 18/01/2025

GUYS. YOURE WITHOUTING THEM. YOURE NOT SUPPOSED TO DO THAT

020

punished felix @punishedfelix.com · 18/01/2025

It's crazy how much people here love alt text but won't actually talk to blind people or inform themselves on their issues beyond popular reposts

120

punished felix @punishedfelix.com · 18/01/2025

Why is it so hard to talk to blind people lol I wasn't even talking about Bluesky's implementation I was just pointing out that its not a one dimensional problem. And that's important if we are going to be serious about critiquing AI

000

punished felix @punishedfelix.com · 18/01/2025

Look bro I'm not saying it's super good and awesome I'm saying actually talk to blind people about the issue. Ive heard quite a few positive opinions from implementations like BeMyEyes. Obviously opinions are mixed because people have different experiences and obviously it doesn't replace alt text

100

punished felix @punishedfelix.com · 18/01/2025

I think i made it pretty clear multiple times that AI is not a substitute for alt text but I also think that it's a good idea to talk to MULTIPLE blind people regarding the issue of descriptive AI outside of just the context of one implementation to understand its impact

010

punished felix @punishedfelix.com · 18/01/2025

Im just asking you to actually involve yourself with the people that alt text impacts the most and ask their opinions on it I hate AI and Im curious about the impacts on accessibility but multiple blind people have corrected me on this point so I'd recommend you talk to them instead of without them

110

punished felix @punishedfelix.com · 18/01/2025

Like to be clear Alt text is preferable and frankly it should be part of image metadata But I've had multiple blind people discuss the usefulness of descriptive AI tech and it is accurate enough to be very usable

130

punished felix @punishedfelix.com · 18/01/2025

Have you asked blind people what they think or are you just basing this on your opinion

430

punished felix @punishedfelix.com · 18/01/2025

Okay, to be fair, describing images for blind people is like the ONE use case I've seen with this AI shit that actually does improve people's lives, even though it's rarely implemented in usable cases. But obviously, alt text is preferable.

110

punished felix @punishedfelix.com · 18/01/2025

Next thing you're gonna tell me is that white English Deleuzoids have a fascism problem! Wow, can't imagine why centering so much discussion around a guy who unironically spent years going on racist rants about brain drain could lead to a bunch of white boys being fucking racists!

060

punished felix @punishedfelix.com · 18/01/2025

Yeah simp for the guy who literally argues there's no escape and not the guy who tells you to try something new. My God isn't this what Guattari literally complained about with his criticism of postmodern art and architecture. Like come the fuck on

130

punished felix @punishedfelix.com · 18/01/2025

I was just thinking about how The Three Ecologies describes an extinction event that is directly related to the issues being produced by AI all the way back in 89 but y'all would rather talk about a fake AI that doesn't mean anything but it's cool cuz it used words like "capitalist" and "future"

120

punished felix @punishedfelix.com · 18/01/2025

It genuinely pisses me off. Like how the fuck can you read through fanged noumena and actually take it seriously but look at three ecologies and be like. Nah. I get it Land is more "fun" but Guattari is actually talking about the real world okay

120

punished felix @punishedfelix.com · 18/01/2025

the worst thing about the deleuze-guattari fandom is they would rather read esoteric larp work by an annoying British meth head racist who likes to throw around tech words because he just discovered UNIX at the university rather than the guy whose name is literally half of "deleuze-guattari"

150

punished felix @punishedfelix.com · 18/01/2025

It's so fucking cool that we are going to have climate fueled famine for hundreds of years just so I can shove minion purple drank up my ass

1220

punished felix @punishedfelix.com · 18/01/2025

It's crazy watching tech bros trying to get "ahead of the curve" with AI when every other discipline in computer science is demonstrating how full of shit they are Like seriously just admit you suck at programming, writing and art and don't set the planet on fire

040

punished felix @punishedfelix.com · 18/01/2025

And it's so true that none of these guys want to admit they have a bad egg and have to rethink things. They're more concerned about profits and fucking over their clients and their data than they are about taking the time to do things right. It's like AI automated the worst tendencies in tech bros

000

punished felix @punishedfelix.com · 18/01/2025

"Fundamental problems don't get solved by adding layers of ad hoc mitigations." This isn't unique to AI. This is just how every startup outside of security deals with security 😆 Last tech job seriously thought multiple iterations of LZW compression counted as "security" I shit you not

110

Reposted by punished felix

Meredith Whittaker @meredithmeredith.bsky.social · 14/01/2025

This is an excellent primer on some of the privacy dangers posed by large scale AI, from a cybersecurity perspective. Written in clear language, it's the most accessible rundown I've seen yet on these topics! desfontain.es/blog/privacy...

desfontain.es

Five things privacy experts know about AI - Ted is writing things

… and that AI salespeople don't want you to know!

32425184

punished felix @punishedfelix.com · 18/01/2025

That's because the spec for email was written by actual engineers and not some fucking assholes with money who builds shitty JavaScript in his spare time

030

Reposted by punished felix

Carolyn Bennett Glauda @notoriouscbg.bsky.social · 17/01/2025

This article has solid, actionable tips for minimizing the damage caused by social media collecting our personal data - without having to delete profiles or burn it all down.

5739632360

punished felix @punishedfelix.com · 18/01/2025

What's even scarier though is stuff where the sources are not as easily traced. Sure I can help organize Chimeres in English but what about other writers who don't happen to have family and friends still sitting on his works. I'd imagine stuff older than the 1900s is particularly at risk

020

punished felix @punishedfelix.com · 18/01/2025

Just something about that output it provides though feels wrong. There is a certain flow to him that is impossible to replicate. So when an AI does schizo analysis it feels like a sort of simulation of it. He already had a hard enough time trying to explain things so this is just... It worries me

100

punished felix @punishedfelix.com · 18/01/2025

It's extra challenging with him. Unlike something like Aristotle it's not easy to pin down Guattari so there's a real serious risk of his work being diffused. At least we may be able to pull together old writings. But this is why I made my archival wiki prototype editable only by restricted parties.

120

punished felix @punishedfelix.com · 18/01/2025

This kinda shit concerns me. I'm really into Guattari who is pretty obscure in English and hard to understand, and want to work with other researchers to archive and interpret his work. But ChatGPT gives me a lot of anxiety. Its hard to place why the output doesn't seem right even if it's not wrong?

110

Reposted by punished felix

Dr. Damien P. Williams, dread portent down from a mountain cave @wolven.blacksky.app · 18/01/2025

So deeply not kidding when I say that the current paradigm of "Gen AI" is doing deep, lasting harm to our ability to make knowledge, meaning, & truth w/ each other. Like… it's really fucking us up & if we don't stop it soon we're gonna see a kind of negative chaos frankly unimaginable to most ppl 👍🏾

screenshot of a Tumblr post from user ecofear yelling "what is HAPPENING" in response to a screenshot of a quotetweet from user @Ovichknox reading "this is so scary to me. i recommended a book to a friend recently and she came back to me a day later saying it was really good and i asked if she finished it already and she said "yeah i got chatgpt to summarise each chapter into 3 paragraphs to it was easy to read"
which is replying to a tweet from @alz_zyd reading "Personally, I read a bunch of classics in high register in college. I guess it was fun. I recently ChatGPT'd Aristotle into readable English, finished it around 5x as fast as a translation, and felt l got the main gist of things. idk does the pain incurred…"

Followed by a a reblog from tumblr use moonglade-poetess reading "Hey guys, with so much love, and as someone with an actual English degree:

Please just use Sparknotes if you're going to do this. I get it. I do. But chatgpt or other genAl shit doesn't actually know what's important for you to know, and in some cases it might fully make shit up. Use sparknotes. Failing that, talk to someone who did read it. I'm begging.

14150911705