Post by @mntmn@mastodon.social
i wrote down my current thoughts on AI, first draft, titled "AI is Fake People": https://minute.social/index.html
what do you think?
i also should say that sections from Jordan S. Carroll's book "Speculative Whiteness—Science Fiction and the Alt-Right", which @holo_memory showed me a while ago, influenced my perspective on tech moguls' too literal embraces of sci-fi ideas https://www.upress.umn.edu/9781517917081/speculative-whiteness/
funny, after reading the draft, @dngrs sent me this video by Another Roof, which is concerned with AI in mathematics but hits on very similar fundamental themes, which is that we enjoy art made by humans, and this also applies to mathematics https://www.youtube.com/watch?v=o0MjPe7wXs8
@mntmn
> In the same way in which I want to experience art made by other humans as a form of emotional and mental connection, I enjoy using software, hardware, and other products that are made by caring, considerate human beings, and not by unfeeling robots. All of these artifacts are essentially fragments of human communication
So much this.
@mntmn What do I think? I think that I wholeheartedly agree. I think that you write with a beautiful clarity. I think that I'm more happy than ever that I decided to order my next laptop from you. :-) Please keep creating--with words, with code, and with circuit diagrams.
@mntmn I would reconsider the assertion that genAI boosters lack empathy. that may be true, but even people with ASPD — who truly feel no empathy — can choose to be kind.
the lacking quality, in my opinion, is compassion.
@mntmn i'm 100% with you on the second half (discussing programming as an art form and as an empathetic endeavour), but not really on the first one (transhumanism being exclusively a billionaire's fantasy). to the extent i was given a human mind i'm not happy with it and i won't be until i shape it to my own ends. do i get to be put into the same category as the likes of Musk and Bezos, now?
@mntmn the other, not really immediately relevant but still important thing is... i can conceive of a machine mind being a peer (though not an equal), in theory; in the same way i can conceive of a cat's mind or a rabbit's mind being a peer (but not an equal)
LLMs are obviously not it, and it's obvious that the people arguing for this being the case right now are doing this in bad faith. but i don't subscribe to what i will (perhaps uncharitably) describe as "human chauvinism". haven't for a long time before the current discourse has been relevant, if anything, the tipping point for me was how people treat pets
this sort of thing is why i haven't, personally, affiliated myself with any of the number of "by humans, for humans" movements in computing despite my work being de facto part of a shared effort. i just don't feel good about accepting that kind of worldview. i don't really talk about it because it would be needlessly divisive in almost any context
@mntmn I was speaking at a conference recently and many of the others there would do this charade about their business not heing possible without AI. I'd ask how and they'd say "I couldn't possibly do all the work myself" with no hint of irony when I talk about nobody hiring when workloads get tough and instead throw money at AI... which they agree hurts the community.
@mntmn "what is their art trying to tell you?" Less art than "artifice" I think.
Not much I'd change. Some light restructuring here and there where some thoughts get ahead of themselves, at least in terms of where they appear (sometimes parenthetically) in their respective sentences. I don't think I need to point them out as I'm sure you'll work them out from rereading it.
I like the bit about schmerz/suffering. A nice distinction. Reminds me of another German word I hadn't thought about in a while... weltschmerz.
Anyway, nice read. Thanks for sharing it with us.
@mntmn this is without a doubt one of the best pieces of writing on the topic. Thank you for sharing your thoughts <3
I like your point about Star Trek's "Data" really being a human actor playing human-written stories!
@mntmn Lucie, that article is spot on.
The kind of people who push this are power hungry, their empathy is damaged.
We can tell the care and empathy you have for the mnt community.
@mntmn I'd love to boost this, very well thought out and written, but is this a permalink? Seems like if you blog more in the future, this would shift out from the main index.html, but correct me.
@mntmn I think you express yourself very well, and your thoughts on the subject mirror my own. Art is communication, and communication is human connection.
@chainq Here's an archive link: https://web.archive.org/web/20261026074728/https://minute.social/index.html
@mntmn I'm reading Magnifica humanitas by the Pope. It's vey similar yet it falls short of consequences. Doesn't makle it less valuable because any reader should come to the right conclusions.
Life is art, programming is art and letting a machine take over is cutting short on what's beautiful in life.
> The people who try this stuff are always of the same kind
I think this is wrong. It might feel that way, but it's weakening the argument. There is more to explore there about what drives people into all this.
> I had a realization that *pain* is the main anchor of reality
I think you're right on the money with this one. And that is also why the idea of AI alignment and trying to teach it ethics is not going to ever fully work.
Ethics are an emergent property of shared vulnerability. Something that can't die cannot truly feel empathy for something that can.
> because they only see the results
Again right on the money I'd say.
The obsession with "efficiency" and believing that it's about the destination, when in reality it is about the journey.
@mntmn
> they lack empathy and are, above all, power hungry. They see others as playthings, or "toys"
I guess it’s a common “just business, nothing personal” thing. Yes, everything/everyone is a tool in this POV
I think you put up a thesis and then wrote about 3-5 other things that are on your mind, but aren't actually related to the thesis.
Whether or not I value genAI "Art" or human "Art" has nothing to do with making fake people or causing intentional suffering.
I do agree that the remixed stuff isn't as helpful as the marketing says and that "not helpful" is on the bad side of a "feels good to use to feels bad to use" spectrum.
But I think that about human output too, so...
I've been carrying my Assumed Intelligence and Unsupervised Text Extruder dread around for a similar amount of time and today I started writing down my own concerns.
After sending them to a friend I read your thoughts and discovered a whole other angle that was hidden to me, even though as a software developer with over 40 years of experience, I've regularly drawn parallels between art and the process of software development in all its frustrating glory.
Thank you for bridging my gap!
@mntmn Every single paragraph of this dissects something that sends me into a bit of a spin on a daily basis, articulates it, and turns it into a seed of optimism, a nugget of hope.
Thank you.
@mntmn like it very much and <3 "please don't sacrifice human connection on the altar of productivity."
@mntmn this very much mirrors my own thoughts. thank you for writing it down and putting it out into the world!
@mntmn
Excellent piece!
Also quick note: You're linking to the homepage of your blog, i don't see a way to link to JUST this piece but that will be needed for sharing, as once you write something else it will get pushed off the front page.
@mntmn
Important perspective. Thanks for the write-up.
I just wonder: what should we do as a socitity? We will not get it back in the bottle. Do you have any ideas how to handle the situation?
I’ve never though about it as “fake people”, but it makes a lot of sense.
One of the questions that came to mind is “where is the line between a videogame NPC and a ‘fake person’ AI?”, but second thought is that that’s misleading as it completely misses the motivation aspect among others.
I’ve tried to write more here 3 times, but there is so much wrong with the current state of AI and the people who create it (very good question at the end!) that one could form a number of coherent arguments against it and they never even touch each other with a lot of space in between them.
@mntmn I think you really judge a bunch of people you don't know rather harshly.
I actually used to have interest in recreating actual human thinking in a computer and it had nothing to do with a lack of empathy or need for power but all about trying to understand how I think.
So I made it about 3 paragraphs in and got super offended and stopped.
Now everyone can call me some AI techbro or whatever.
genai
@mntmn it's great!
i've always stated that there is a flipside to the argument that all the data was taken without consent: that being that the output is chopped up human creations. this feels like another way to state the same thing. genai chatbots are trying to be fake humans, which is inherently devaluing of humanity itself
i think one of the main things the genai industry is doing is waging a blitzkrieg on humanity itself, and we should not let them devalue humanity
genai
@mntmn if using the term "blitzkrieg" here feels off, do tell me. i mean that they're trying to push it through as fast as possible so that we can't mobilize a proper resistance
It seems to me, a lot of awful things, get their names changed, in order to avoid recompense?
e.g. Windscale (renamed to Sellafield after nuclear accidents, e.g. https://en.wikipedia.org/wiki/Windscale_fire).
In the USA? Bribery is illegal; yet "lobbying"? Seems to be a different word, for the same sort of thing, thus apparently escaping whatever primitive regex pattern match was in place prohibiting the practice. To me? I perceive "cloud" computing as more or less synonymous with timesharing mainframe systems which the personal computer revolution in the last quarter of the 20th century, sought to displace. Facebook/Instagram/WhatsApp? Now they're Meta. Twitter? Is X. undsoweiter.
The names may have changed, yet the underlying awfulness? Remains the same.
To me? Having considered myself a survivor of the last so-called "A'I' Winter"? I see LLMs and Machine Learning as new names, for the same old bullshit.
Which is to say: "those who have learned from history are doomed to watch others repeat it."
Or maybe why some claim: "history doesn't repeat itself, but it rhymes."
I dunno, maybe in a longer timeline, I will fathom the rhymes? Just seems villainous ad nauseam.