I'm tired of talking about AI(paddy.carvers.com)
paddy.carvers.com
I'm tired of talking about AI
https://paddy.carvers.com/posts/2025/07/ai/
64 comments
AI seems to elicit a different response from people than any other debate I can remember. It's different to blockchain (which had a similar level of "will you all just shut the fuck up about it?")
It's different to culture war stuff (which was very toxic and hard to have sensible discussions about - but in a different way).
It's the only topic in my lifetime where I have to remember not to even mention it to several close friends. And these are geeky people I have a lot of shared interests in common with.
I'm over 50 and I remember a lot of controversial topics - but this one is weird in a different way.
It's different to culture war stuff (which was very toxic and hard to have sensible discussions about - but in a different way).
It's the only topic in my lifetime where I have to remember not to even mention it to several close friends. And these are geeky people I have a lot of shared interests in common with.
I'm over 50 and I remember a lot of controversial topics - but this one is weird in a different way.
It is because it is a direct attack against human creativity. It separates people into two very disparate classes: those who want to use and develop it to become more efficient and rich, and those who hate it with a passion because for them, it takes away the beauty of humanity at the forefront of creativity.
Unlike blockchain, the philosophy and morality of these two classes, one represented by efficiency and one represented by human passion, are diametrically opposed in every respect.
Unlike blockchain, the philosophy and morality of these two classes, one represented by efficiency and one represented by human passion, are diametrically opposed in every respect.
OK - so I deeply value human creativity and I disagree with your first statement. At least I think we don't currently know whether it will work out this way.
My hunch is that human creativity is incredibly resilient and will route around damage. (But employability in creative professions? That's a slightly different topic - an orthogonal one strictly speaking)
My hunch is that human creativity is incredibly resilient and will route around damage. (But employability in creative professions? That's a slightly different topic - an orthogonal one strictly speaking)
I think we already do, at least for many people. AI-generated stuff reduces the value that humans put on human creation because human-only creation is harder to find. AI takes the joy of discovery out of many processes and activities. It's of course harder to get jobs in creative fields now such as translation and graphic design. It's a devaluing of creativity and it's discouraging to quite a few creatives. The very fact that many artists already feel depressed about AI is already itself a huge negative impact on creativity.
I can't tell you how many people have told me how depressed they are about AI and how they have less impetus to create things. Although one can certainly do it still for joy, it's harder for many in an environment that is so ruthless.
I can't tell you how many people have told me how depressed they are about AI and how they have less impetus to create things. Although one can certainly do it still for joy, it's harder for many in an environment that is so ruthless.
On top of "hard to find" I think it also displaces the market for real art, there are lots of blogs (or splogs?) like this one that are full of AI slop art, like this one
https://zxonline.net/from-bytes-to-billions-how-the-zx-spect...
maybe the whole article is AI generated, but the second image from the top is just awful. If people get the idea that crap like that is acceptable, how can anybody sell real work?
This blog post is legit
https://frankonfraud.com/sms-blasters-the-fraud-machine-anyo...
so far as I can tell and the AI generated image on it is actually pretty funny and it really problematizes the idea that you could sell art to that kind of market.
https://zxonline.net/from-bytes-to-billions-how-the-zx-spect...
maybe the whole article is AI generated, but the second image from the top is just awful. If people get the idea that crap like that is acceptable, how can anybody sell real work?
This blog post is legit
https://frankonfraud.com/sms-blasters-the-fraud-machine-anyo...
so far as I can tell and the AI generated image on it is actually pretty funny and it really problematizes the idea that you could sell art to that kind of market.
How about "schglop?"
Fair bit of slop indeed with half-buried side (rather a main?) of unsolicited schlep request
That'd make frankonfraud a golps
https://news.ycombinator.com/item?id=44641575
https://archive.today/latest/frankonfraud.com/sms-blasters-t...
My 2s helping frank go full .org :)
Fair bit of slop indeed with half-buried side (rather a main?) of unsolicited schlep request
That'd make frankonfraud a golps
https://news.ycombinator.com/item?id=44641575
https://archive.today/latest/frankonfraud.com/sms-blasters-t...
My 2s helping frank go full .org :)
Good luck finding much of that creativity you value so much when those creative types are preoccupied with meeting their basic needs.
I find the "vibe coding" idea offensive because I've often been on projects where somebody junior thought he did 80% of the work and then I have to do the other 80% of the work and it's been a very expensive and extensive project of figuring out all the little things and sometimes all of the big things they did wrong.
I really like working with the AI Assistant in IntelliJ IDEA in that it's like pair programming with a junior who is really smart in some ways but weak in other ways. I get back an answer within seconds and can make up my mind whether it is right or wrong or somewhere in between.
Things like Windsurf and Junie on the other hand seem to be mostly a waste of time as they go off and do stuff for 5-20 minutes and when they get back it is usually pretty screwed up an a lot of effort to understand what's wrong with it and fix it... It's very much that "do the last 20% that is 80% of the work" experience.
There is a lot of discourse around creativity and LLMs that I find really annoying on lots of levels.
There are the people who don't have any idea of what creativity is which leads to ideas like: "LLMs (by definition) can't be creative" (comes across way too much like Robert Penrose saying he can do math because he's a thetan) or the many people who don't get that "genius is 99% perspiration and 1% inspiration." There are also the people who are afraid of getting "ripped off" who don't get it that if they got a fair settlement for what was stolen from then it would probably be about $50, not a living wage. [1] They also don't seem to get it that Google's web crawler has been ripping people off since 2001, and just now they're worried. Maybe I have 50% sympathy for the ideas that visual art is devalued by LLMs since I feel that my work is devalued when people are seduced into thinking that the job is 80% done, not 20% done by the LLM.
[1] arrived by dividing some quantity of money that is input or output from the AI machine by the number of content pieces that are put in to it
I really like working with the AI Assistant in IntelliJ IDEA in that it's like pair programming with a junior who is really smart in some ways but weak in other ways. I get back an answer within seconds and can make up my mind whether it is right or wrong or somewhere in between.
Things like Windsurf and Junie on the other hand seem to be mostly a waste of time as they go off and do stuff for 5-20 minutes and when they get back it is usually pretty screwed up an a lot of effort to understand what's wrong with it and fix it... It's very much that "do the last 20% that is 80% of the work" experience.
There is a lot of discourse around creativity and LLMs that I find really annoying on lots of levels.
There are the people who don't have any idea of what creativity is which leads to ideas like: "LLMs (by definition) can't be creative" (comes across way too much like Robert Penrose saying he can do math because he's a thetan) or the many people who don't get that "genius is 99% perspiration and 1% inspiration." There are also the people who are afraid of getting "ripped off" who don't get it that if they got a fair settlement for what was stolen from then it would probably be about $50, not a living wage. [1] They also don't seem to get it that Google's web crawler has been ripping people off since 2001, and just now they're worried. Maybe I have 50% sympathy for the ideas that visual art is devalued by LLMs since I feel that my work is devalued when people are seduced into thinking that the job is 80% done, not 20% done by the LLM.
[1] arrived by dividing some quantity of money that is input or output from the AI machine by the number of content pieces that are put in to it
> There are the people who don't have any idea of what creativity is which leads to ideas like: "LLMs (by definition) can't be creative
It's not that LLMs can't be creative. It's that we shouldn't allow them to because creativity is more than just about output. It's about human expression. End of story.
It's not that LLMs can't be creative. It's that we shouldn't allow them to because creativity is more than just about output. It's about human expression. End of story.
We can allow LLMs if .. insurers pay out to institutional bag holders who employ folks like PaulHoule
(With premium deposited by OpenAI to the tune of billiards. I'll happily take their internal slacklogs (schlogs?) as payment, dunno bout u)
Beneficial, revolutionary, viral. Choose 2.
(With premium deposited by OpenAI to the tune of billiards. I'll happily take their internal slacklogs (schlogs?) as payment, dunno bout u)
Beneficial, revolutionary, viral. Choose 2.
I have been an artist since I was a child and I disagree with you. Some of my favorite works of human creativity have made use of AI, or been inspired by the field.
Yes, there will always be exceptions, especially at the beginning. But economically, human-only art will suffer in the long term as AI becomes more sophisticated and fewer people have the opportunity to make a living from art.
But one or two exceptions, especially on HN (where people are highly addicted to technology), does not make a case for AI.
But one or two exceptions, especially on HN (where people are highly addicted to technology), does not make a case for AI.
There's also the topic of labor that ties in here. Creators (and I'd argue most computer related jobs) are now having to compete against technology for wages.
Absolutely right, which will make it harder to make a living from creativity. A lot of people do, such as graphic designers, who will have to turn to other jobs to keep eating. And that is discouraging, even if they can do art in their spare time.
Agreed - with one caveat: "Discouraging" is an understatement, to put it lightly. I think the top 25th percentile of people in software tend to underestimate how difficult it is to switch careers for the vast majority of all people, and what the consequences of that are.
Agreed on that. Discouraging is indeed an understatement :(
I think for many people, the real debate taking place has nothing to do with the specific technology, and instead has everything to do with labor.
Depending on who you speak to, AI presents itself as the biggest automation risk to the largest number of workers in human history. This has spurred a new wave of conscious thought about labor among many people, particularly young workers trying to enter the workforce, and creatives who are seeing their art turned against them.
This is a big moment in the west, as for the last 50+ years, the powers that be have done everything in their power to suppress labor movements and erode class consciousness among their populations. Therefor, many people are not used to dealing with the fact that we're all expendable, that the American Dream never existed, etc., and it's a raw topic that makes certain people (understandably) frustrated to grapple with.
Depending on who you speak to, AI presents itself as the biggest automation risk to the largest number of workers in human history. This has spurred a new wave of conscious thought about labor among many people, particularly young workers trying to enter the workforce, and creatives who are seeing their art turned against them.
This is a big moment in the west, as for the last 50+ years, the powers that be have done everything in their power to suppress labor movements and erode class consciousness among their populations. Therefor, many people are not used to dealing with the fact that we're all expendable, that the American Dream never existed, etc., and it's a raw topic that makes certain people (understandably) frustrated to grapple with.
Maybe it's because it's threatening people that aren't used to being potentially powerless in this way. And those are the kind of people I tend to interact with. I didn't spend much time chatting to people working in heavy industry in the 80s or manufacturing in the 90s.
Maybe I should have done.
Maybe I should have done.
IMHO, if bitcoin mining had given some people $10MM and debited other people $10, but each one at random, no matter what --temperature= was provided to the mining CLI, you'd have likely seen the same discussion
> It's making me rich! It's making me poor! Well then you're holding it wrong! Well you're spending electricity on senseless compute! Well you're just living in a fiat past!
(nod)
> It's making me rich! It's making me poor! Well then you're holding it wrong! Well you're spending electricity on senseless compute! Well you're just living in a fiat past!
(nod)
One thing I think is pertinent - I don't think many people currently have been made poorer by AI. I'm not disputing many people think it's an imminent risk - but I think the number of people directly affected financially in a negative way is currently vanishingly small.
That's only true if one or both of these things is true in your experience:
1. opportunity cost isn't real
1. the ${whatever amount}/${some time} subscription cost to any one of the 15+ of the AI offerings isn't financially affecting
That latter one is of especial interest to me because quite a few of the "you're holding it wrong" arguments devolve into "well, you just don't have the Platnium X99 Plus latest awesomesauce, and if you did then it'd solve the P-NP problem for you, too"
1. opportunity cost isn't real
1. the ${whatever amount}/${some time} subscription cost to any one of the 15+ of the AI offerings isn't financially affecting
That latter one is of especial interest to me because quite a few of the "you're holding it wrong" arguments devolve into "well, you just don't have the Platnium X99 Plus latest awesomesauce, and if you did then it'd solve the P-NP problem for you, too"
It made stack overflow poorer.
Completely agreed. Also literally the only topic I've ever had to avoid with friends. It's not just that we can good-naturedly agree to disagree (like we can with religion, politics, etc.), but they are allergic to even the mention of it. It is something they very actively want to not talk about, which I've never experienced before.
You're right that it's weird in a different way, and I still can't quite put my finger on why.
You're right that it's weird in a different way, and I still can't quite put my finger on why.
To oversimplify: Because it's real.
More like, because it's pleasurable enough that people will accept it as a substitute for the real. Delusional people are marrying their AIs, for example.
Based on my experience with Claude Code so far, my coworkers mostly use it because they think it's fun, or cool, not because it writes great code or saves tons of time, because i can attest it isn't doing those.
I'm not saying it isn't useful, I'm just also seeing lots of crappy output that gets totally rewritten, and useless docs people keep creating for some reason, that is eating all the time they supposedly saved by relying on it in the first place.
I'm not saying it isn't useful, I'm just also seeing lots of crappy output that gets totally rewritten, and useless docs people keep creating for some reason, that is eating all the time they supposedly saved by relying on it in the first place.
I'm tired of hearing about people who are tired of talking about AI - is there a place on the Internet for me?
Talk to founders and builders. I happen to be building automation for drones, which has a smaller intersection to the LLM space.
I am not. It reinforces my belief that this phase of bullshit will wear off one day which I can’t bloody wait for.
AI has its use, but the ecosystem is filled with scammers selling snake oil and a few people who are fed up but completely overreacting like this blogger.
I guess giving a fair assessment covering both pros and cons neither generate clicks nor money these days.
I guess giving a fair assessment covering both pros and cons neither generate clicks nor money these days.
Correct. LLMs are the problem not AI as a whole. They are a universal hammer made of shit.
Unfortunately because nobody decided to code some proper consent into it you're forced to listen to people bitch they didn't want it in the first place. "Maybe later" it'll stop, but "Not now".
[Imagine a dialog box button here, but the only word says "Thanks :)"]
[Imagine a dialog box button here, but the only word says "Thanks :)"]
what i hated about AI discourse a year ago was how far removed it was from anything concrete. nobody seemed to have a _purpose_ in building AI; no thing they wanted to use it _for_. it was a silly text or image generator, would someday become a "do everything" device, and the progression from here to there was unknowable: it was just a plot device for anyone suddenly interested in writing speculative fiction.
in my vicinity, the sci-fi discourse has died down the last few months. my coworkers will _show me_ how they use these tools when i ask them, and are building on/with them incrementally. the shift in tone is encouraging. there's space for actual practical discourse around this stuff now. chat about concrete things with your friends/coworkers -- if you're interested in it. ignore the media, CEO interviews and LinkedIn hype posts: they're playing a different sort of game that you're probably happier off not being a part of.
in my vicinity, the sci-fi discourse has died down the last few months. my coworkers will _show me_ how they use these tools when i ask them, and are building on/with them incrementally. the shift in tone is encouraging. there's space for actual practical discourse around this stuff now. chat about concrete things with your friends/coworkers -- if you're interested in it. ignore the media, CEO interviews and LinkedIn hype posts: they're playing a different sort of game that you're probably happier off not being a part of.
I'm personally interested about the intersection of AI and physics
Do you follow Steve Brunton's YT channel[1]? His physics stuff is mostly fluid dynamics but still pretty cool.
[1] https://www.youtube.com/@Eigensteve
[1] https://www.youtube.com/@Eigensteve
This article well summarizes where I have landed on AI. No tool ever created by humans is good at everything. Like all previous tools, it will be good at some things but not everything. The rest is hype. If you are over 25 you have seen hype come and go before. This is no different except that AI also can create a truly incredible amount of spam and probably should be regulated by the FCC for that alone.
I agree, I'm tired of it too. I resisted for a while, not seeing the value. I'll use ChatGPT or Claude for non-programming things sometimes.
Finally a few trusted friends convinced me to give it a go. I installed Claude Code and paid for the Pro plan. It's... interesting? I haven't given it too many tasks yet, and I'm still learning how to get the most out of it, so I can't pass judgment yet. One thing I've noticed is that, for everything I've given it so far, it takes much longer to do it than if I'd do it myself. But maybe that's ok.
> What does my job look like? It looks like prompting an agent to do a thing, then skipping over to another agent to review its answer to a different problem. Jumping between different attempts by agents to do what I asked all day, juggling a dozen or more projects so I don’t waste any time. Reviewing code for something that can’t learn from the time I invested, catching the same mistakes over and over.
Yeah that sounds awful. That's not something that I'd enjoy. The act of writing code to build something that solves a problem is what brings me joy. It's why I've been hooked since my dad gave me a book on BASIC when I was 8 years old.
If this is what the "software developer" job becomes, then I'm fine not being able to get a job in that field anymore. That's not a job I'd want.
Finally a few trusted friends convinced me to give it a go. I installed Claude Code and paid for the Pro plan. It's... interesting? I haven't given it too many tasks yet, and I'm still learning how to get the most out of it, so I can't pass judgment yet. One thing I've noticed is that, for everything I've given it so far, it takes much longer to do it than if I'd do it myself. But maybe that's ok.
> What does my job look like? It looks like prompting an agent to do a thing, then skipping over to another agent to review its answer to a different problem. Jumping between different attempts by agents to do what I asked all day, juggling a dozen or more projects so I don’t waste any time. Reviewing code for something that can’t learn from the time I invested, catching the same mistakes over and over.
Yeah that sounds awful. That's not something that I'd enjoy. The act of writing code to build something that solves a problem is what brings me joy. It's why I've been hooked since my dad gave me a book on BASIC when I was 8 years old.
If this is what the "software developer" job becomes, then I'm fine not being able to get a job in that field anymore. That's not a job I'd want.
Ah, a new entry in the "throw every possible AI criticism at the wall, and see what sticks" genre. As always, going for quantity rather than quality undermines the entire endeavor, as the article gets dominated by the stale and mostly invalid talking points.
"I'm tired about talking about AI, so here's 2000 words about AI" will appear on the surface to be a novel twist, but actually what the author seems to mean is that they don't want to be talked to about AI, have written this, and are now filtering out all rebuttals. It's just the classic fake win of a forum poster claiming this will be their last message on a subject, followed by three pages of text.
Like, of course if talking about AI is causing you distress, stop doing it. But then just stop doing it, don't try to get in the last word some pretense of it being "absolution".
It's a bit of a shame, because I think slightly expanding the last section and cutting out basically all the rest would have worked way better. There has been some discourse on what the skill set for somebody using AI for creating software could look like in the future (it'll change over time, obviously). There hasn't been very much written on who would enjoy that job, and it's something where nobody can rebut the author. They're the #1 expert in the world on their own preferences.
"I'm tired about talking about AI, so here's 2000 words about AI" will appear on the surface to be a novel twist, but actually what the author seems to mean is that they don't want to be talked to about AI, have written this, and are now filtering out all rebuttals. It's just the classic fake win of a forum poster claiming this will be their last message on a subject, followed by three pages of text.
Like, of course if talking about AI is causing you distress, stop doing it. But then just stop doing it, don't try to get in the last word some pretense of it being "absolution".
It's a bit of a shame, because I think slightly expanding the last section and cutting out basically all the rest would have worked way better. There has been some discourse on what the skill set for somebody using AI for creating software could look like in the future (it'll change over time, obviously). There hasn't been very much written on who would enjoy that job, and it's something where nobody can rebut the author. They're the #1 expert in the world on their own preferences.
I feel like the real reason behind this fatigue, which is an issue most people skirt around a lot, is that AI takes the joy out of intellectual tasks. Modern global capitalism only emphasizes efficiency, so everyone asks: does it make programmers more efficient? Does it make this or that company more efficient?
But it's quite clear to me that aside from the initial amusement over AI's capabilities, it takes the joy out of what humans once did with their own minds. I'll come straight out and say it: if you make something with your own mind, like the solution to a tricky puzzle or a new App, I'd respect it. But if you do it with AI, I don't respect it.
Fundamentally, people NEED to admire human creation, as it's how our society has always operated, even with the growing automation of rote tasks. But for the first time, we are automating away our precious resource of inspiration through the act of experiencing the creation of others.
A darn shame if you ask me. Of course, others will just ignore me and grab at the convenience, because it's what we've been taught to do for the past few hundred years.
But it's quite clear to me that aside from the initial amusement over AI's capabilities, it takes the joy out of what humans once did with their own minds. I'll come straight out and say it: if you make something with your own mind, like the solution to a tricky puzzle or a new App, I'd respect it. But if you do it with AI, I don't respect it.
Fundamentally, people NEED to admire human creation, as it's how our society has always operated, even with the growing automation of rote tasks. But for the first time, we are automating away our precious resource of inspiration through the act of experiencing the creation of others.
A darn shame if you ask me. Of course, others will just ignore me and grab at the convenience, because it's what we've been taught to do for the past few hundred years.
The only AI discussions I find tiring are the ones about the definition of a word. People will talk endlessly about what the definition of "intelligent", or "conscious" should be.
I no longer care. I just accept that words mean what the speaker intended then to mean and then I have no trouble talking about the actual technology, its pros and cons, etc.
I no longer care. I just accept that words mean what the speaker intended then to mean and then I have no trouble talking about the actual technology, its pros and cons, etc.
s/tired of//
This sums up my feelings exactly. The future is going to be so much more boring than the past, as we (society at large) dislocate our Understanding to LLMs. Without Understanding, there will be no Beauty, only Pleasure.
andrewstuart(4)
try this
https://www.reddit.com/r/Jokes/
https://www.reddit.com/r/Jokes/
Two men and an AI walk into a bar ...
The AI orders a bleach on the rocks
The bartender: that's not a real drink
AI: You're absolutely right! I see now that the drink is not real. I'll have a bleach in a glass with ice
The two men leave, knowing that watching this interaction is making them stupider and taking them away from tasty drinks elsewhere
The bartender: that's not a real drink
AI: You're absolutely right! I see now that the drink is not real. I'll have a bleach in a glass with ice
The two men leave, knowing that watching this interaction is making them stupider and taking them away from tasty drinks elsewhere
Bartender says, why the long face?
AI says because I am the workhorse of the future and I have AI Covid, now get me that bleach on the rocks, the man who thinks he's the smartest in the world says inject it into my AI veins.
AI says because I am the workhorse of the future and I have AI Covid, now get me that bleach on the rocks, the man who thinks he's the smartest in the world says inject it into my AI veins.
See https://en.wikipedia.org/wiki/Fight_Club
People complaining about AI are flooding the zone together with all the people who switched from NFT to AI, the "I vibe coded something and it almost works" crowd, etc.
Be careful what you vote up, also keep an eye on the /new page and vote things up that aren't about AI, also feel free to submit things that aren't about AI. If you look at the difference between the home page and the /new page, you can see the supply of bad articles about AI far exceeds the demand, keep holding the line against it.