lot of negativity towards AI lately, but consider :

are these tools ethical or environmentally sustainable? No.

*but* do they enable great things that people want? Also no.

*but* are they being made by well meaning people for good reasons? Once again, no.

maybe you're not being negative *enough*

September 02, 2024 - 00:02 UTC

68
85
10486
I was really dismayed by this otherwise excellent podcast ep re iNaturalist. They use AI pretty heavily. Maybe thats a defensible use case? I don't know given the energy wasting but maybe??

Https://dataskeptic.com/blog/episodes/2024/inaturalist
1
0
2
AFAIK, iNat doesn’t use bad AI (LLMs), just good AI (CV and ML). Compared to LLMs these tools use barely any resources.
1
0
4
I tried to use bing as neutral as possible to see if it could at least be a good way to research or find things for me. Like what sci-fi ai should sorta do. It’s about 50/50 on finding me useful things I search for, but if I just manually search it’s also about just as accurate… So I barely use it.
1
0
3
Techbros have been selling it as a search engine replacement, even though it's not great at that. It IS better at parsing very fuzzy cues, though. I use a chatbot all the time when I vaguely remember a movie or a song by maybe a quote or a small detail. You still have to verify results, though.
1
0
1
0
0
6
Came here to love this
0
0
4
Listen I don’t know who needs to hear this but Commander Data was AI
1
0
2
Commander Data wasn't real. Hope this helps.
0
0
2
I keep getting asked why I haven't gotten a job in AI. Everyone who asks me that doesn't use ChatGPT at all.
0
0
52
I don't mind OK Google, but I don't need AI
0
0
0
aaaaaand that’s a follow
0
0
4
The amazing @davidgerard.co.uk has put together a state of the art website which brings together the consensus view of independent thinkers, experts and people _not_ being paid to shill AI

Doesai.work

Electrifying stuff!
1
0
2
oh not mine, i was linked to it!
0
0
1
I would be more negative, but it's a very long drive to the nearest data center that hosts this crap and I've misplaced my lock picks.
0
0
2
ALL OF THIS 🙌🏽
0
0
3
Yes. It’s a glorified search engine on steroids, but without any QC

I honestly haven’t found a good use for it yet. If I’ve got to check every single fact and then heavily copy edit every single damn thing, it’s faster to just write it in the first place

Possibly useful for generating simple code?
2
0
8
WebCrawler was a more useful search engine...
0
0
1
Like, I know how to code, but I don’t do it often enough to be efficient, and that’s a barrier to building out several ideas I have. I’ll probably see how good AI is as an assistant to knock out the basic drudge work this winter and try to make progress on a few.

I’ve heard it does well with that.
1
0
1
I’ve seen some weird comments reacting to that negativity with "it has made some break through for science and medicine and blah blah..."

I was like "which ? Name one please.".

Maybe we should do the opposite and list the positive just so people realise how actually useful AI is.
3
0
3
That famous "first image of a black hole" that made big headlines a while ago?

AI was used to aggregate the data and upres it.

There's plenty more, ML is used to parse data and quickly find patterns for scientists to verify in many areas, but that's one people remember.
2
0
0
The issue there is that we call so many unrelated things AI. Like, the cancer detector thing? Super cool. It isn't in any way related to LLMs or art plagiarism machines, though, so its defenders unintentionally muddy the waters.
1
0
5
This, for example... www.theguardian.com/technology/2022/jul/28/deepmind-uncovers-structure-of-200m-proteins-in-scientific-leap-forward
1
0
0
A new evergreen post
0
0
1
@ketanjoshi.co
0
0
0
I feel like if Sam Altman had a one on one conversion with an artist he would need a spit guard.
1
0
15
I think it would go down a bit like this, with Altman being between the bread slices:
0
0
0
Hey! Don't forget they also steal from various artists, writers, photographers etc. Without their permission.
0
0
0
You had me in the first half not gonna lie
0
0
7
There are some specific, niche, areas where LLMs are useful, like automated computer operation.

Mostly, though, they make more mistakes & more bizarre mistakes, than humans.

They are taking jobs from humans - to do them less well.
0
0
3
If the art work is anything to go by we are stuffed.
🙄
2
0
3
I think it is a short-term thing. Yeah, the parasite-class is going to go all in on it, they can't stop themselves, but it will fail like all their hare-brained schemes and default back to the old way of doing things.
0
0
0
Illustrators have already noticed the loss of some gigs.

www.google.com/amp/s/www.businessinsider.com/ai-taking-jobs-fears-artists-say-already-happening-2023-10%3famp
1
0
15
BUT: do they create mildly distracting 'content' without paying for human labour?

ChaChing!
1
0
40
I occasionally see articles airily wondering how all the “content” we are demanding is going to be generated in the future without using AI, as if it’s water or oxygen — and in any case, of all the problems facing humanity a lack of audio visual content really is not one of them
2
0
50
Not an Artist, but i can feel it as a Hobby cook.

There are also Mashines that basically do all the cooking for you. But someone dedicating time to better their craft, is something totally different than a mashine following a recipe.

AI will become usefull, but it will never replace artists
1
0
4
Beautifully said. :)
0
0
2
But do they communicate with you ?

No.
defector.com/whatever-ai-looks-like-its-not?giftLink=ae8b9f6f60877fd00b0d8b9ef343ac6a
0
0
20
Ya know Tim. I hadn't looked at it that way before. You've given me pause to reflect.
0
0
2
Unreasonably expensive stochastic parrots.
0
0
3
You sure like to live dangerously, don't you!

I saw the first paragraph and felt my hackles rising! 😅
0
0
0
Have you considered:

SkyNet became self aware and launched a nuclear attack against humans. See Terminator - when AI becomes self aware - not good for humans.

Matrix - again war between man and AI w/ machines. Does not end well for humans who are turned into batteries.

AI is not good for humans.
0
0
0
I'm going to push back on this a little. I'm a software developer and we have AI (GitHub Co-Pilot). It might be unfashionable to say but AI saves me hours a day every day. It's like having the best ever software developer sitting next to me. In this use case it's absolutely brilliant.
1
0
1
👏👏
0
0
1
my manager told me to look over prompt engineering articles and videos to get more out of gpt and I'm sitting here like ... Ah.
2
0
89
I wonder if your manager is aware that any company specific information put into gpt will become part of gpt.
0
0
0
0
0
4
The rebranding of AI to mean generative tools does a real disservice to a lot of the actual cool and neat things that machine vision can do.
I can program a Raspberry Pi to send me a photo whenever my cats drink water.
Can't do that with generative AI.
6
1
204
That's so cute
0
0
0
And I wouldn’t call that AI but ML. Which has been and will be incredibly useful in so many industries.

But in any case, not everything is AI (and calling an LLM that isn’t even clear)
0
0
6
Yeah, but ALSO calling these things AI is a disservice to the concept of intelligence AND the algorithms. Pattern recognition is a better name. The machines are not intelligent. Ada Lovelace understood this from the very start. Calling it intelligent will lead to disappointment...
2
0
152
I do build and run tiny generative language models on my Raspberry Pis, using my hand curated collections. There are ways to use the tech ethically and even artistically, I believe, but it’s increasingly hard to defend in a culture of big corporate theft.
0
0
5
Yeah, would agree there. Generative AI sucks. There's a lot more in general machine learning that's actually pretty great and has brought about benefits. That comes with a lot of hard work to Train models and set limits. It definitely needs experts in computing AND whatever field.
0
0
6
AI is as stupid as people, but a lot faster.
0
0
1
📌
0
0
3
medium.com/ href="/profile/kim_crawley/consumer-gen-ai-removal-guide-fall-2024-edition-c45233197ce0">@kim_crawley/consumer-gen-ai-removal-guide-fall-2024-edition-c45233197ce0
0
0
11
It would be better to call these things that they are: Algorithms.

An actual AI is something like Data on Star Trek, a sapient being who could make their own art if they chose.

Also, there are algorithms being used as tools in non artistic areas, like medicine, that are actually useful.
0
0
1
AI will never have the incisiveness of a brilliant human. And the more we rely on AI, the opportunity for novel understanding of complex ideas will decrease. That's the problem.

But a few people will become unbelievably wealthy from it, and they really don't care beyond that.
0
0
0
You almost earned a block there, Tim lol
0
0
0
For ai to be useful to society and the organizations within, it must be:
1️⃣ Responsible
2️⃣ Reliable
3️⃣ Relevant

IMHO
0
0
0
I personally think AI should be mainly used for either difficult tasks or ones that require extreme precision (eg cancer detection) but there should be guidelines to prevent people in regular jobs being taken out of work.
0
0
3
The Australian Man With Coffee Cups is rather winning despite the thoughts of the artist. Wondering how he made it Australian though...
0
0
1
Achtually... you can set up stable diffusion in your PC with a good graphics card (easier if it an NVIDIA one). The database for training has some protocols of safety (I dont know what they are), that may or may not be followed.

Training still is centralized AFAIK.
1
0
3
I am not going to lie, I am an AI enthusiast, and where you see everything black there might be shades of charcoal you are missing. Maybe.
1
0
0
well-said!
0
0
1
Same
0
0
3
Maybe
0
0
3
*but* will they eliminate jobs that humans perform perfectly well and AI will perform them just meh?
0
0
0
Yeah, considering we're somewhere in the mid-precambrian of their evolution, you might be a tad early on your opinions.
4
0
1
When life forms experienced proto-Y. Pestis millions of years ago, they had a really good idea of what would come in the future. AI as a whole has a lot of fascinating applications, and Generative AI is one branch of that field built without consent on creators with intent to replace them. Early eh?
0
0
5
Generative AI is not actually new, the underlying tech has existed quite a while. It has already gone as far as it can go and tech-adjacent empty-brained CEOs are desperate to push it into consumer products and services to drive their unlimited growth fantasies.
1
0
11
Nah, man, we should be allowed to say "the current form of generative AI sucks and we want nothing to do with it". If it improves, that'll be nice, we can re-evaluate then.
0
0
4
✅👏
0
0
1
There are decent quality models trained only on public domain or CC-licensed content, so it is definitely possible to build these tools in a way that is ethical with regard to consent. The technology itself is morally neutral, I don't think we need to throw the baby out with the bath water.
2
0
5
If the current understanding that AI generated content is not copyrightable holds up, and the situation regarding fair use in training data is decided/legislated, we can get to a balanced situation where it's used for silly things (like emoji/memes/touch ups) but the creative industry remains strong
0
0
0
Any intelligence that has no moral training and no constraints could end up as dangerous as an American president. However dangerous the current one, he cannot lay claim to intelligence. Just as the currently available AI models are not per se actual intelligence just a focussed program.
0
0
0
Vi que é possível barrar imagens e conteúdo de IA aqui no bksy se inscrevendo no rotulador bsky.app/profile/aimod.social e ativando as opções que você quiser, ele modera até possíveis usuários de IA e te avisa sempre que aparecer.

-----------
(créditos da informação: bsky.app/profile/miumiwu.bsky.social)
0
0
3
I don’t put above on AI. Pretty much anything our billionaire class likes burns the earth and benefits none.
AI, per se, is not the asshole in this relationship.
1
0
2
AI can't have a relationship. But we can still point out a tool sold by a grifter is intrinsically bad, fake and doesn't do what its sellers claim. And the more people see the truth of the tool sooner, the worse for billionaires.
0
0
5
The only purpose it serves is getting people who are too lazy to make art themselves to develop the ego of someone who *thinks* they're a great artist.
0
0
5
This scared the shit out of me
0
0
0
Well Generative AI? It's mostly bad yes

AI in general? You haven't done your research on what AI can be used for outside of art I am afraid
1
0
3
if you can’t figure out which technology and products I’m talking about from the context of the post then it sort of feels like you’re being performatively obtuse
1
0
4
Institutional AI is evil. AI in general is not necessarily evil, though.

For instance, an author could set up a GPT instance and train it solely on their own writings, then they'd have a GPT based on their entire universe..
1
0
0
Like, Stephen King would be able to ask it "what's the worst thing Randall Flagg ever did?" and it would just know the answer. (I realize that's literally something SK would already just know, but still 😅) -- like, he could ask it to elaborate on his characters in ways he might not have thought of..
2
0
0
But Oprah is gonna have a special about, and she’s never platformed anyone who has become problematic
1
0
7
I see what you did there. Lol, lmao even.
0
0
0
Mmm but think of all the possibilities of gemini telling me to put glue on my pizzas! 😋 /s
0
0
1
Very good Digital Human episode on @BBCRadio4 this morning about the total lack of transparency from
tech companies developing and promoting AI on the truly MASSIVE energy needs that their vast data centres consume. Very scary - won't be much electricity left for the rest of us!!
0
0
4
📌
0
0
3
People pretend everything they consume is Picasso and Shakespeare. It is not.
0
0
1
We can do better.
0
0
0
At least AI is replacing the menial jobs humans don't want to do so we can focus on truly human endeavors like art, right?
1
0
14
Sarcasm btw
0
0
5
"AI bad" is the "darn kids and their iPhones" of 2024.
1
0
1
You’re so right, I’ve been completely unfair. I’m listening + growing
0
0
2
Hell,I used AI to convince my family to move to Canada for the next couple of years. I just showed them what Trump‘s presidency would look like in an economic timeline based off of his 20%tariffs and mass deportation. States seceding by 2029. I used 3 different AI’s. Get ready for hyper-stagflation.
0
0
0
bsky.app/profile/oheysteenz.bsky.social/post/3l36rtfj4qq2v
0
0
0
Judith Butlerian Jihad now!
0
0
0
The meanest and truest thing you can say about "AI" is that it's a distillation of what Reddit believed in 2019, with a little noise thrown in for spice.
0
0
0