If you could transport yourself to the late 90s, would you try to convince people that the internet's impact on the economy would be no greater than that of the fax machine by 2005.
Overvalued and/or stupid ideas because what doesnt work hadnt been figured out yet. Their failure highlighted what doesnt work and the next round did better. Same deal with ai but probably worse because i see nothing to say open models wont converge to closed models making model companies market power zero eventually. That makes them overpriced gpu reits that wont be competitive at that once the better hardware in the labs right now takea over. The dollar values are also much higher and more painful to write off
If you bought 100 of those "overvalued" companies right before the bubble and 99 of those went to zero and the 100th was Amazon, you would have made bank.
"Being right too early is indistinguishable from being wrong. If I know something’s going to happen in the market, but I’m five years too early, that’s the same thing as being wrong."
Calling AI a "bubble" is not calling it useless or a fad. It's just pointing out that companies in the sector are overvalued. And it's important to recognize and correct bubbles quickly because the bigger the pop, the worse the overall economic damage to everyone.
The instructive part about the comparison is that believing "AI is a bubble" doesn't mean saying that AI is worthless. After all, after the dot-com bubble burst the internet still existed and continued to transform the whole world.
Waymo had a story on this site not long ago. All of the comments were mostly positive - but in real life, no real life human I know wants Waymo. None. And anyone who bikes is vigorously against it.
So- who is the audience for all of this? Reading this AI slop article - I just feel as if a moment of my life was wasted. And not wasted by poor writing, but by meaningless synth.
I hope we have the de-slop button soon. Ad-block -> Slop-block. I would rather read bad writing with meaning, than meaningless organized prose.
Can I ask what makes it read like slop? Because lots of comments here on the style and little on the thinking behind it, which is where I spent my effort.
It seems like a lot of folks arguing about rather or not we need to still read code.
For me, it's mostly in the rhythm and pacing of the sentences -- everything is sooo so punchy. Paragraphs are composed of high-impact, terse sentences. Phrases like "Both perspectives, one person", separated on their own line for dramatic effect. To me it feels like the linguistic equivalent of ultraprocessed food; it's over-engineered for impact and gravitas, to get its point across.
I honestly feel like my actual writing style has drifted towards this, and I'm sure others' have too (although I overused em-dashes before it was uncool!).
But I think that… feeling someone's distinct verbal fingerprint as you read is a signal of authenticity among a sea of lazily generated slop, whereas even a carefully-crafted, thoughtful piece where the final thing is generated by an AI is missing that signal of authenticity.
The fact that it's slop. People know it when they read it.
Put Claude and other LLMs away (for coding, writing, everything - block them on your firewall fully) for a few days and then write a blog post from scratch using your own words. No asking for help, comments, editing, nothing. Then compare against this. It'll jump out at you.
> Put Claude and other LLMs away for a few days and then write a blog post from scratch using your own words. No asking for help, comments, editing, nothing.
Ok, but (and this is the whole point) where _should_ we use LLMs? Because your comment could easily be said to someone writing code.
If slop === LLM involved, that's too reductive.
If LLM does all the work, that's also too reductive.
This blog post is so rambly that I don't think this discussion will go anywhere beyond the usual AI good/bad arguments. This is exactly the sort of thing that frustrates HN.
Since we're not banning these submissions outright, we should maybe move them to their own filter (like "new", but exclusively for "AI" rants)? Somewhere, anywhere, very very far away from the rest of HN.
Is it frustrating that I used Claude to turn detailed outlines, notes, and discussions into a post?
> go anywhere beyond the usual AI good/bad arguments
This is literally the point of the post. AI is not good or bad just like the internet was not all bubble or all value. The goal is to better understand how to use ai well. I'd rather try writing my thoughts than just dump a bunch of notes and diagrams on a blog
It might also be the design of the blog post. It's a wall of text and human made diagrams might have actually helped. I kind of don't like the font used for the body text either. That's more subjective, but sans-serif text tends to tire me out with more length and larger glyph size. You have both going on.
I've seen plenty of LLM-written posts succeed despite the rambling tone (and because they weren't about AI). People don't really care about or notice LLM writing as much as they claim. Like anything else, it's details. I would expect that a human is still needed for details.
On AI... I get tired of the topic only because there's not much new to say, there is so much being said anyway, all the comments spiral into dumb emotion, and even if I added my own take it will get lost in the noise.
It might get more traction and still get the ideas across if it was only tangential to AI without explicitly saying it. Those discussions are more balanced and everyone tends to be happier. This post could have strictly focused on the investment side of things and how people let that go to their heads.
The irony of using Claude to write this is almost to much to "hold". (Claude loves to tell you to hold things in its writing).
Here's another great Claudism: "The replies had a different read. One of them, in full: {embedded tweet}" lol. Who writes like that? "In full"? You're promising me you didn't fake the embedded tweet or something?
Also this one: "So here’s my first hypothesis, stated so you can hold me to it."
Sentence is utterly redundant, and it reads like Claude first wrote it with a colon and you edited it to be a comma to read less like AI. I think you should've just removed it entirely.
> Here's another great Claudism: "The replies had a different read. One of them, in full: {embedded tweet}" lol. Who writes like that? "In full"? You're promising me you didn't fake the embedded tweet or something?
I originally had the content of the tweet and when posting switched it to have the embedding. So ironically, the error you're pointing out was my manual fix
Because code is instructions to execute a program. It's not meant to be consumed as an end product itself, only its output, which is independent of what it's made from, and if written functionally correct, irrelevant for using the software.
Articles are written to be directly consumed. That's why the writing matters there.
The next question might be, why do people prefer reading human writing to AI writing. There are a few reasons.
One is that because the person writing it is an actual person, they've imagined how the reader will take the content in and have made sure that it reads well. Different people will have different writing styles and you may prefer some over others, but generally they'll make it interesting, engaging, not ramble on, respect the readers attention, get to the point quickly. AI tends to put in a bunch of filler and not really make much of a point, or say something obvious as a conclusion.
A second reason is that people want to feel connected to another person when reading something written by them. They might trust the person and build a mental image of them as they read, and let that further inform whether they agree with their opinions and recognize biases, and refine that image of the person to better understand if they can trust the opinions and perspectives in their writing. You can't do that with the output of an llm.
In summary, LLMs are good for asking questions to when researching facts, producing functional output that is either right or wrong. They are also useful but not completely trustworthy when exploring ideas or asking opinions on things. And terrible when producing text meant for humans to read (including its own output a lot of the time in chats), and any kind of visual design beyond simple templates.
No, AI isn’t like the dot-com bubble and it’s short sighted that people keep trying to compare it.
The dot-com bubble had investment before mass adoption.
AI has mass adoption, there are many industries where usage of AI is now expected.
They have similarities, but claiming they are the same is so fucking stupid.
It doesn't make a difference, the profitability is way overvalued, potential customer companies are firing and closing all over the world, while AI debt is growing in billions (visualize curves going in opposite directions). They'll never be able to repay them on time. Plus, these investments are based on a promise they'll never uphold since the LLMs are a technological dead end for their ambitions.
Yes, there's a bubble and we've known for a while now, thank you.
> It doesn't make a difference, the profitability is way overvalued
That’s your opinion, for many there are clear patterns of loss-leaders and market capture plays.
> potential customer companies are firing and closing all over the world
I’d wager this is due to economics outside of AI.
> Plus, these investments are based on a promise they'll never uphold since the LLMs are a technological dead end for their ambitions.
More speculation and opinion.
> Yes, there's a bubble and we've known for a while now, thank you.
Ok, so analyze it accordingly instead of lazily calling it the same as the dot-com bubble.
There are some things you’ve said that I agree with (to a limited extent) and others that I do not. I have no interest in slop articles calling it a dot-com bubble, put some effort forward and make the share worthwhile.
if you read the post, you'd see I do analyze it and point out one reason calling the internet a bubble was both right and wrong, and that AI is similar. It's not even about the bubble aspect but that you can have wildly divergent perspectives that are somehow both accurate.
>> Plus, these investments are based on a promise they'll never uphold since the LLMs are a technological dead end for their ambitions.
>More speculation and opinion.
You claim I'm just spouting opinions, yet you seem to have miss headlines last year about AI authors/fathers walking out the door claiming just the same... But to you, they're just spouting their own worthless opinion as well.
Please spend a bit of time researching the subject before posting anything here.
> You claim I'm just spouting opinions, yet you seem to have miss headlines last year about AI authors/fathers walking out the door claiming just the same... But to you, they're just spouting their own worthless opinion as well.
It is speculation and theory. Many of said same authors have been surprised at how far transformers have taken the general technology. Yes, transformers have inherent limitations and not all LLMs are powered by transformers.
Besides that, many experts don’t want their entire career defined by strictly transformers based LLMs, it’s very reasonable they have “walked out”.
> Please spend a bit of time researching the subject before posting anything here.
> Hold that exchange still for a second. Nobody in it is stupid. That’s what makes it worth understanding.
The room falls silent. Except there is no room. There is no scene to dramatize. So I have to instruct you.
Two things. Both of them true. Yet I’ve decided that your imbecile mind can’t imagine that. How can they both be true? One word. Nuance.
I would know. On one hand I haven’t written a line of code in a year. But that’s table stakes. Everyone does not write code any more. I also haven’t written even a four-minute blog post in a year. (Arthiritis? AI? Both. It’s both.) On the other hand I am critical of AI. I think fully autonomous coding and writing is not here. Yet. Nuance.
My pieces used to be about writing. Now? They are about understanding.
My pieces used to be about individual inquiry. Now? I string you along for a series of posts. That’s collective inquiry.
By the end you won’t need me to spell out a conclusion. Bubble. Revolution. Dark. Light. Contrasts. It’s not discrete. It’s analog.
Was dot com a bubble?
If you could transport yourself to the late 90s, would you try to convince people that the internet's impact on the economy would be no greater than that of the fax machine by 2005.
Or was it just some companies were overvalued?
Overvalued and/or stupid ideas because what doesnt work hadnt been figured out yet. Their failure highlighted what doesnt work and the next round did better. Same deal with ai but probably worse because i see nothing to say open models wont converge to closed models making model companies market power zero eventually. That makes them overpriced gpu reits that wont be competitive at that once the better hardware in the labs right now takea over. The dollar values are also much higher and more painful to write off
If you bought 100 of those "overvalued" companies right before the bubble and 99 of those went to zero and the 100th was Amazon, you would have made bank.
"Being right too early is indistinguishable from being wrong. If I know something’s going to happen in the market, but I’m five years too early, that’s the same thing as being wrong."
https://www.mckinsey.com/featured-insights/quote-of-the-day/...
> Or was it just some companies were overvalued?
That's what a bubble is.
Calling AI a "bubble" is not calling it useless or a fad. It's just pointing out that companies in the sector are overvalued. And it's important to recognize and correct bubbles quickly because the bigger the pop, the worse the overall economic damage to everyone.
The instructive part about the comparison is that believing "AI is a bubble" doesn't mean saying that AI is worthless. After all, after the dot-com bubble burst the internet still existed and continued to transform the whole world.
Same argument was used for crypto 5 years ago, people have just moved onto to using it for AI now.
Funnily enough a lot of the same players are involved, they’ve just managed to fool a bigger group of people this time.
At least the thought got through to one person
AI on AI
Bot farms, AI stories, AI comments on AI.
Waymo had a story on this site not long ago. All of the comments were mostly positive - but in real life, no real life human I know wants Waymo. None. And anyone who bikes is vigorously against it.
So- who is the audience for all of this? Reading this AI slop article - I just feel as if a moment of my life was wasted. And not wasted by poor writing, but by meaningless synth.
I hope we have the de-slop button soon. Ad-block -> Slop-block. I would rather read bad writing with meaning, than meaningless organized prose.
We need a report AI slop button like LinkedIn
This might be interesting, but it reads so strongly like AI slop I had to stop.
Can I ask what makes it read like slop? Because lots of comments here on the style and little on the thinking behind it, which is where I spent my effort.
It seems like a lot of folks arguing about rather or not we need to still read code.
For me, it's mostly in the rhythm and pacing of the sentences -- everything is sooo so punchy. Paragraphs are composed of high-impact, terse sentences. Phrases like "Both perspectives, one person", separated on their own line for dramatic effect. To me it feels like the linguistic equivalent of ultraprocessed food; it's over-engineered for impact and gravitas, to get its point across.
I honestly feel like my actual writing style has drifted towards this, and I'm sure others' have too (although I overused em-dashes before it was uncool!).
But I think that… feeling someone's distinct verbal fingerprint as you read is a signal of authenticity among a sea of lazily generated slop, whereas even a carefully-crafted, thoughtful piece where the final thing is generated by an AI is missing that signal of authenticity.
> Can I ask what makes it read like slop?
The fact that it's slop. People know it when they read it.
Put Claude and other LLMs away (for coding, writing, everything - block them on your firewall fully) for a few days and then write a blog post from scratch using your own words. No asking for help, comments, editing, nothing. Then compare against this. It'll jump out at you.
> Put Claude and other LLMs away for a few days and then write a blog post from scratch using your own words. No asking for help, comments, editing, nothing.
Ok, but (and this is the whole point) where _should_ we use LLMs? Because your comment could easily be said to someone writing code.
If slop === LLM involved, that's too reductive. If LLM does all the work, that's also too reductive.
I'm just talking about a few days, so that you see the contrast. You can do that, right?
Giving out homework is ridiculous. Just say you are a slop oracle and be done with it.
This blog post is so rambly that I don't think this discussion will go anywhere beyond the usual AI good/bad arguments. This is exactly the sort of thing that frustrates HN.
Since we're not banning these submissions outright, we should maybe move them to their own filter (like "new", but exclusively for "AI" rants)? Somewhere, anywhere, very very far away from the rest of HN.
Is it frustrating that I used Claude to turn detailed outlines, notes, and discussions into a post?
> go anywhere beyond the usual AI good/bad arguments This is literally the point of the post. AI is not good or bad just like the internet was not all bubble or all value. The goal is to better understand how to use ai well. I'd rather try writing my thoughts than just dump a bunch of notes and diagrams on a blog
It might also be the design of the blog post. It's a wall of text and human made diagrams might have actually helped. I kind of don't like the font used for the body text either. That's more subjective, but sans-serif text tends to tire me out with more length and larger glyph size. You have both going on.
I've seen plenty of LLM-written posts succeed despite the rambling tone (and because they weren't about AI). People don't really care about or notice LLM writing as much as they claim. Like anything else, it's details. I would expect that a human is still needed for details.
On AI... I get tired of the topic only because there's not much new to say, there is so much being said anyway, all the comments spiral into dumb emotion, and even if I added my own take it will get lost in the noise.
It might get more traction and still get the ideas across if it was only tangential to AI without explicitly saying it. Those discussions are more balanced and everyone tends to be happier. This post could have strictly focused on the investment side of things and how people let that go to their heads.
The irony of using Claude to write this is almost to much to "hold". (Claude loves to tell you to hold things in its writing).
Here's another great Claudism: "The replies had a different read. One of them, in full: {embedded tweet}" lol. Who writes like that? "In full"? You're promising me you didn't fake the embedded tweet or something?
Also this one: "So here’s my first hypothesis, stated so you can hold me to it."
Sentence is utterly redundant, and it reads like Claude first wrote it with a colon and you edited it to be a comma to read less like AI. I think you should've just removed it entirely.
> Here's another great Claudism: "The replies had a different read. One of them, in full: {embedded tweet}" lol. Who writes like that? "In full"? You're promising me you didn't fake the embedded tweet or something?
I originally had the content of the tweet and when posting switched it to have the embedding. So ironically, the error you're pointing out was my manual fix
Everyone got lazy in a big hurry.
Everybody was already lazy. Now they have a means of demonstrating it.
> My argument, Claude’s sentences. It won some of the arguments along the way. The mistakes are still mine.
From the footer.
lol. lmao, even.
how is that any different than claude generating code? honest question
Because code is instructions to execute a program. It's not meant to be consumed as an end product itself, only its output, which is independent of what it's made from, and if written functionally correct, irrelevant for using the software.
Articles are written to be directly consumed. That's why the writing matters there.
The next question might be, why do people prefer reading human writing to AI writing. There are a few reasons.
One is that because the person writing it is an actual person, they've imagined how the reader will take the content in and have made sure that it reads well. Different people will have different writing styles and you may prefer some over others, but generally they'll make it interesting, engaging, not ramble on, respect the readers attention, get to the point quickly. AI tends to put in a bunch of filler and not really make much of a point, or say something obvious as a conclusion.
A second reason is that people want to feel connected to another person when reading something written by them. They might trust the person and build a mental image of them as they read, and let that further inform whether they agree with their opinions and recognize biases, and refine that image of the person to better understand if they can trust the opinions and perspectives in their writing. You can't do that with the output of an llm.
In summary, LLMs are good for asking questions to when researching facts, producing functional output that is either right or wrong. They are also useful but not completely trustworthy when exploring ideas or asking opinions on things. And terrible when producing text meant for humans to read (including its own output a lot of the time in chats), and any kind of visual design beyond simple templates.
No, AI isn’t like the dot-com bubble and it’s short sighted that people keep trying to compare it.
The dot-com bubble had investment before mass adoption. AI has mass adoption, there are many industries where usage of AI is now expected. They have similarities, but claiming they are the same is so fucking stupid.
It doesn't make a difference, the profitability is way overvalued, potential customer companies are firing and closing all over the world, while AI debt is growing in billions (visualize curves going in opposite directions). They'll never be able to repay them on time. Plus, these investments are based on a promise they'll never uphold since the LLMs are a technological dead end for their ambitions.
Yes, there's a bubble and we've known for a while now, thank you.
> It doesn't make a difference, the profitability is way overvalued
That’s your opinion, for many there are clear patterns of loss-leaders and market capture plays.
> potential customer companies are firing and closing all over the world
I’d wager this is due to economics outside of AI.
> Plus, these investments are based on a promise they'll never uphold since the LLMs are a technological dead end for their ambitions.
More speculation and opinion.
> Yes, there's a bubble and we've known for a while now, thank you.
Ok, so analyze it accordingly instead of lazily calling it the same as the dot-com bubble.
There are some things you’ve said that I agree with (to a limited extent) and others that I do not. I have no interest in slop articles calling it a dot-com bubble, put some effort forward and make the share worthwhile.
if you read the post, you'd see I do analyze it and point out one reason calling the internet a bubble was both right and wrong, and that AI is similar. It's not even about the bubble aspect but that you can have wildly divergent perspectives that are somehow both accurate.
>> Plus, these investments are based on a promise they'll never uphold since the LLMs are a technological dead end for their ambitions. >More speculation and opinion.
You claim I'm just spouting opinions, yet you seem to have miss headlines last year about AI authors/fathers walking out the door claiming just the same... But to you, they're just spouting their own worthless opinion as well.
Please spend a bit of time researching the subject before posting anything here.
> You claim I'm just spouting opinions, yet you seem to have miss headlines last year about AI authors/fathers walking out the door claiming just the same... But to you, they're just spouting their own worthless opinion as well.
It is speculation and theory. Many of said same authors have been surprised at how far transformers have taken the general technology. Yes, transformers have inherent limitations and not all LLMs are powered by transformers.
Besides that, many experts don’t want their entire career defined by strictly transformers based LLMs, it’s very reasonable they have “walked out”.
> Please spend a bit of time researching the subject before posting anything here.
Uncalled for.
> Hold that exchange still for a second. Nobody in it is stupid. That’s what makes it worth understanding.
The room falls silent. Except there is no room. There is no scene to dramatize. So I have to instruct you.
Two things. Both of them true. Yet I’ve decided that your imbecile mind can’t imagine that. How can they both be true? One word. Nuance.
I would know. On one hand I haven’t written a line of code in a year. But that’s table stakes. Everyone does not write code any more. I also haven’t written even a four-minute blog post in a year. (Arthiritis? AI? Both. It’s both.) On the other hand I am critical of AI. I think fully autonomous coding and writing is not here. Yet. Nuance.
My pieces used to be about writing. Now? They are about understanding.
My pieces used to be about individual inquiry. Now? I string you along for a series of posts. That’s collective inquiry.
By the end you won’t need me to spell out a conclusion. Bubble. Revolution. Dark. Light. Contrasts. It’s not discrete. It’s analog.