Thoughts on Kyle Drake's stance on AI

[malice]

Welp, Kyle Drake, the creator of Neocities, made a blog post in defense of AI/machine learning. I suggest reading the full post of course, but he basically says we should embrace AI, including the use of AI "art" and using AI for therapy. While AI/ML has its uses (ex: the cancer screening tool he mentions in the post), stuff like Midjourney and therapy chat bots are just crossing the line in my opinion.

I know that some people are migrating to Nekoweb as a result of this post, which I am considering partially because of the post, but mostly because I found out they have anti AI/ML scraping measures in place, but due to the scope of one my websites, that will not be for awhile. Besides, Kyle is not getting any of my money so I don't feel too bad about staying at this moment. This situation sucks, but I don't feel like it's worth moving literally thousands of images over to a new host and getting everyone to change their links to my new host.

What do y'all think about all of this?

EDIT: I posted this like 1 beat before the forum closed and since then I found out that
1. That blog post was over a year old! Sorry for unintentionally misleading everyone
2. Nekoweb rocks actually, I switched mostly because of the features they have (I can upload BMP files!! The file editor is a dream to use!!!). Even if this whole situation hadn't come out, I would've switched anyway the moment I found out about NW's features.
3. I found out I could just download my entire site and NW has an upload zip file button for porting sites. Only took 2 clicks!

[xixxii]

i feel like it's worth noting that this blog post is over a year old - the way you phrased your post makes it sound like it's something that just happened haha.

~corrected a confusing typo! -melon

[DiffydaDude]

I honestly dont even care. I remember back when that stuff started showing up 3 years back, people had all sorts of fun making tf2 animations with it and messing around with the fun stuff, but when the creators started trying to market it in everything, people started avoiding it like the plague, and now they dont even touch the stuff and will dawg on anyone who even says anything about AI.
The main point of his article wasnt about how he thinks AI will make the world better. Most of his article is primarily about how people treat it like its anthrax and will turn the whole world upside down when thats not exactly set in stone. He has a point and i agree with him on that.
I still hate it when they use ai art in stuff but i also hate how people like Rick Beato are like "AI IS DESTROYING MUSIC ITS A PRIVILEGE NOW", like why do people act like theyre prophets whenever literally anything concerning happens. Like shut the fuck up.

[malice]

i feel like it's worth nothing that this blog post is over a year old

Ah so sorry! I posted this right before the forum closed and I found out about that after. I made a post edit clearing it up

[larvapuppy]

Some dude with two first names instead of a first and last name strikes again!

In all seriousness, I don't love his takes but I don't think I'll leave neocities over this. I still think it's a good service and a nonprofit with an admirable mission. Plus, I don't have a very good impression of Nekoweb so I wouldn't want to move over there. I'd have to do more research on other services before deciding to move, and I don't have time to look into anything that deeply at the moment.

I personally don't think AI is completely evil or useless either, but I do think that it's a net negative at the moment, and nowhere near as impressive as its supporters make it out to be. If people would stop using it to suck all of the joy and humanity out of life that would be great. We really don't need anyone playing devil's advocate for it right now!

[candycanearter07]

Nekoweb is definitely an interesting site but
1. I don't know if it supports the same api I'm using to make my site with Jekyll + Git
2. I'm already paying for supporter on Neocities?
3. Porting everything over would be a nightmare considering I'm already procastinating on multiple projects on my site currently (old comic index -> new comic index, comics to the new comicfury archive, and refactoring everything) and everything is so spagetti also changing all my links..

[Kallistero]

This seems centered around AI as a creative tool. Every framing around AI seems to have some issue or other with it, but I think some of the effects of AI can wind up being helpful to the artist or creative who feels threatened by it.

One thing I started to see very early on with AI is that when AI coding entered the public eye, people in different levels of software development, instead of wholly panicking that it would replace them, started to leverage it in order to get their work done. It became their assistant. Some of those who were more forward-thinking would also leverage it as a learning tool, to make themselves better at the craft. I've never personally used AI-generated code, but it wasn't that I didn't get curious! I think it was more that when I create, I like to deviate from norms, and AI lives in a world made of norms!

Every model that's in the contemporary lexicon of AI, it's more or less just a massive interpolator that takes in lots of data points and then can output information that falls somewhere between those data points. If your vision or style or context is outside of its data set & can't be constructed through something that's between its data points, you are beyond the AI in that domain. Also, because the input format is linguistic, the limitations of spoken language will also apply. If you cannot fully describe the genre of a piece of music, you cannot instruct an AI to generate it. 

Of course, the above applies to any other form of art, too. AI seems unable to generate the likeness of work from artists such as LunaticBinko or Trick Or Trick, whose artistic styles fall outside of norms that you might find in an AI training set. The same might be said of a lot of the sites that we've made as webmasters. AI lives in a box, and as a creative, you have the ability to move outside of that box. You can also leverage AI to learn what it knows and then use that knowledge to move beyond it. It's a forward way of thinking about AI that I believe Kyle was getting at in his article, specifically when he said that he sees AI "as a tool of personal creativity and intellectual empowerment." You can shun the tool for diminishing the value of more standard forms of art, or you can use the tool as a springboard to make things that the tool could never create on its own. No tool can replace an artist who can use that tool as one of many in their toolbox.

[TheFrugalGamer]

You can shun the tool for diminishing the value of more standard forms of art, or you can use the tool as a springboard to make things that the tool could never create on its own. No tool can replace an artist who can use that tool as one of many in their toolbox.

Agreed, I think this is a much more practical way of looking at AI that many have a lot of disdain for. But the fact of the matter is that AI isn't going to just go away no matter how much we want that, so we ought to at least benefit from it. Otherwise the only ones getting a leg up will be venture capitalists and techbros, and those aren't the people we need leading the world.

I think the urge to put it into everything at the moment, though, is mirroring the dot com bubble--there are some things that just don't need it. There's that viral tweet that went around that I thought put it really nicely:

I don't want AI to create art and write novels for me. I want AI to do my dishes and my laundry, so that I can create art and write novels.

or something like that. I want AI to make our lives better, not meaningless.

I'm also really glad he brought up self driving cars. While I agree they are not safe right now, the potential for saving lives is something I've been really hopeful about since they were introduced. AIs can't get drunk before they take over the wheel. They can't get angry and make stupid decisions as a result, and they don't have anything to prove, or the ability to feel a rush when going over safe speed limits. Preventable deaths due to cars are a tragedy and it's about time we start to work towards making them a thing of the past.

[Melooon]

I always took his article to be more about not being excessively dramatic about the dangers of machine learning, rather than actively supporting its use. I don't appreciate all the "AI" corporate hype, but I also don't appreciate all the anti-AI hype that seems just as toxic.

I also deeply don't think it makes sense to use this blog post as some sort of Neocities v Nekoweb contest; both platforms have their pros and cons, and none of those have anything to do with AI or blog posts  (Not that anyone's doing this here, but I feel like its bubbling away in some chat somewhere)

As for machine learning tools in general, and the risk to artists; well I'm an artist, and many of my family are artists (I also studied AI design in college and have a pretty good grasp of how it works even though its developed a lot since then). I have lived and worked within arts spaces all my life, and within my circle, I don't know anyone who appears threatened or has expressed worry about diffusion image generation ~ that's a fear Iv mainly seen expressed online, and it doesn't really connect to the reality I live in ~ generated images, and the work of most artists, are simply different things!

On a practical note, I, and most of people reading this topic, have used machine learning based tools for years. Many image resizers use machine learning to correctly scale images, spam filtering has used machine learning for ages, none of this is new ~ more recently Iv used ChatGPT to help me figure out complex bits of code (particularly when there is hard maths involved such as 3D movement calculations); and as a family of dyslexics, machine learning based spellcheckers have helped me and many others in my family to improve our writing; my mum actually just sent me a draft of her first book! 

When it comes to generated content, such as images or text or even whole websites; Iv played around with it and I have not found it to be good enough to include in any of my projects. An LLM cant write with the style and originality that I want from my work, nor can it generate images or audio that have the little sparks of uniqueness that I look for. That's because, by its nature an LLM generated work is a lowest common denominator, its a generic re-combination of things that already exist (again, I studied it, I know how it functionally works), so it really has nothing new to offer ~ sometimes that's fine; its great for things that have traditionally needed AI, such as NPC movement in games; but for most creative applications, like making a whole image or a whole game, its just not good enough.

I actually have a phrase I use; the "value of potentiality", e.g. what is the value of all the potential outcomes you can get from an interaction, what new things can you learn? what relationships can you build? what can you offer to the person or thing you are interacting with?

In almost all cases the value of potentiality from AI generated work is almost 0, whereas its quite high for any human made work. For example; If a human writes the dialogue for a game character; I can find out who they are, and contact them, saying how much I like their work (or hate it!), and who knows where that interaction could lead. We could becomes friends, they could marry my cousin, we could end up stranded on a desert island, or we could just have a short interesting conversation ~ all of those are are potential outcomes that add some value to our lives, and they still exist, even if none of them ever happen and we never interact! (That's why existing is fun!)

However if an LLM generates the dialogue of a game character, then there is no further interaction I can have; the potential outcome is 0, nothing, no further possibilities. This potential interaction is part of what makes engaging with art so fascinating and powerful to people, and its why content generated by an LLM will never really be a threat; even if the content is of a high quality, it cannot satisfy the human need for possibilities, and therefor it cannot satisfy the purpose of art It might seem flashy for a while, but what you'll quickly find is that people will gravitate towards art that has a higher potentiality (For example; Thats why YouTube beats TV, because your potential of having an interaction with a YouTuber feels more real than having an interaction with a TV show)

Ooph, that was a bit of a tangent! However its all to say that I agree with not being overly dramatic about the dangers of machine learning based tools (or any new technology); they have their uses and they have their limits. I may not agree with everything he says, but I support Kyle's stance, simply because he approached an issue with a level head, particularly at a time like last year when most people were busy loosing their heads in the AI v AntiAI buzzword tech war of the week (If there's one thing I truly can't stand, its people jumping on a bandwagon!)

[arcus]

I disagree with him on a few points. However, as long as he does not turn around and start selling data for training purposes, I have few qualms with the site.

Nekoweb uses Cloudflare. If ethics are a concern, please keep that in mind. Cloudflare have knowingly and willingly let hate sites keep using their services, up until they were pressured to drop them by news outlets. In addition, their terms of service is incredibly bad from a privacy perspective.

I am not calling Nekoweb bad here. Only that their use of Cloudflare is unnecessary when there's less questionable alternatives, and that it goes against their claims of privacy.

[malice]

Wow, I'm seeing lot's of interesting answers to this! I do agree with those who are saying that people jumping ship on Neocities solely because of this post are overreacting a bit. I admit, I kind of did too when the news first broke out, it is after all why I looked into moving hosts in the first place, and this conversation brought me back down to earth.

I know that AI does have a lot of great uses, but I'm seeing people mention things I haven't even really considered before:omg: it's quite interesting.

I also deeply don't think it makes sense to use this blog post as some sort of Neocities v Nekoweb contest; both platforms have their pros and cons, and none of those have anything to do with AI or blog posts  (Not that anyone's doing this here, but I feel like its bubbling away in some chat somewhere)

Yeah, I don't know why Nekoweb was the one getting pushed the most, there's a ton of hosting sites out there. What helped me make my decision to switch was mainly due to me finding a lot of personal benefits with NW's features, but it does have a couple of drawbacks for me. Everyone is doing different things with their websites, so whatever hosting site works for one might not work for the other.

[cmdr_nova]

I have my own feelings about AI, in that scraping the entire web for art done by real people in order to generate something for people too lazy to pick up a pencil is really ridiculous, but one guy's opinion isn't going to stop me from using one of the best website hosts online. And to top that all off, I actually tried out Nekoweb, and was stopped at account creation because I have a VPN on at all times (I pay for Proton mail and it comes with it).

When I asked in their Discord why they're blocking VPN use and why this isn't even in their FAQ, I got scolded for using a VPN, which was really weird and not all that privacy conscious. Not a good look for people who have a service storing people's data

[di]

I'm very anti-AI as conceptualized by "tech bros" and soulless corporations looking to pay less human beings, and I am a regular reader of Pivot to AI, so that tells you where I'm at..

That being said I kind of appreciated hearing Drake's perspective. I don't know if I'd call AI art "beautiful," and I also think the therapy thing is a bit much (but I guess if nobody is being exploited that is fine?), but I was surprised by his mention of how AI could help protect the mental health of content moderators. I appreciated that a lot. Though to be fair I believe we've had "image recognition" for a long time at this point.

Drake sounds very optimistic, and that's not wrong. "AI" in a utopian world could perhaps do a lot of good, and on its own, the technology is neutral and fascinating. I appreciate his suggestion that we just call it machine learning - but unfortunately the "AI" buzz-ship has sailed, and aforementioned corporations are acting like AI is some esoteric, magical, omnipotent thing.

Still, I agree with the blogger Amod Lele when they described themselves as a reluctant techno pessimist. The greed, lack of critical thinking, and consent violations are enough to make me a luddite against AI.

[villain]

As for machine learning tools in general, and the risk to artists; well I'm an artist, and many of my family are artists (I also studied AI design in college and have a pretty good grasp of how it works even though its developed a lot since then). I have lived and worked within arts spaces all my life, and within my circle, I don't know anyone who appears threatened or has expressed worry about diffusion image generation ~ that's a fear Iv mainly seen expressed online, and it doesn't really connect to the reality I live in ~ generated images, and the work of most artists, are simply different things!

I'd just like to revisit this point - I absolutely agree with the sentiment that there's no way this could replace human creativity, but I think it's unfortunately just not true in the eyes of capitalism. It's great you don't know anyone who feels threatened or worried by genAI images but as someone working within the games industry it is already contributing to the massive amount of job loss happening at the moment. I am part of the UK Games Workers Union and every other week there's discussion in the Discord of folk's companies mass laying off employees.

The studio I work at started replacing early steps of the concept art pipeline with using AI generated images to replace an artist gathering reference for a moodboard. Even this early inclusion felt wildly insulting to our entire concept artist team (including myself) and was widely disliked as a choice across the company, including coders, designers and other staff outside of the art team. I managed to initiate an art team meeting about the inclusion of genAI, and all it really amounted to was a very circular discussion where all of our legitimate concerns raised about the copyright issues, ecological impact, and quality were met with 'yeah but its fast and cheap and all the other games companies are using it so we have to'.

I don't want this to sound like an attack but I think its worth bringing up as part of this discussion - livelihoods are materially being affected by this technology!

[ThunderPerfectWitchcraft]

@villain I agree: The results suck, but it still is harmful for artists especially at the "lower end" of the food chain (the lower, the worse). I recently had a talk with someone who professionally writes blog posts, and has to select a fitting image every time.
If he takes a stock photo, he spends time hours selecting it, than he needs to get it approved by the administration (licenses and so on), and - if it costs money - has to make a request for it; it takes hours. If he uses an online AI art generator, he can just take the first result. He said to me that being a stock photographer must be damn frustrating - every picture they would offer would be better but the shit the AI generates, but they have no chance.
In another thread, somebody described how a company chief just ditched the idea of getting a logo from a professional designer in favor of the AI art, cause its cheap.

And another thing: It doesn't stop at art. A colleague of my mother is training a "business AI" till she goes into pension. When she's gone, here place won't be assigned again - instead, this "AI" will send a selection drafts for the invoices and letters she wrote to others, who will then select the best one beneath their current workload.

The idea that the "better might succeed" in a capitalist environment is a daydream. Instead, techniques often cause social ruptures. Just as mass-produced sweatshop-clothing ousted artisan tailors, or the later luddites were right to smash the automated looms

---------------
Regarding the OP: As said elsewhere, there are at least strong indications that the admin of Neocities is a follower of a radical tech-positivist liberalism (Brand, who he cites here, is an icon of this movement) - e.g. the idea that social issues can be resolved by technological advance, and that any form of governmental regulation is potentially harmful to it.
Adepts of this ideology will either play the social ruptures of technical development down (for example, by stating that the positive factors and increase in wealth will be sufficient to weigh things out) or, in more extreme cases, welcome them as necessary collateral damages in the wake of the way to a glorious future.

The question if this is a reason to change the host is another one, though. If you set a piece of media online, you can safely assume that it will - sooner or later - be used to train AIs. Technical protection will be cracked, and I'm pretty sure that - regardless of any law - some corpo will always be there to mine it, if needed from some corner of the world were you can hardly take any juristic steps against it.