Given the soul-crushingly large amounts of money sloshing around in this tiny slice of the industry, the amount of astroturfing, positive-sentiment manufacturing, & etc. is right about where I expect it to be... as is the amount of vehement denial that it exists. ;)
> The other option: people are using AI successfully on the job and are writing based on their own experience.
Those aren't mutually exclusive.
And it's worth noting that people claiming to use "AI successfully on the job" often do not stop at "their own experience." They often make sweeping statements like "LLMs write better code than people," and attack any critiques or skepticism of LLMs.
A careful reader of my post notices that I neither said nor suggested that there were no honest success stories. ;)
cough
It's at least as important as Slashdot was back in the day. I'm not "buying" what you're "selling", dude.
[0] <https://www.quora.com/When-and-how-did-the-meaning-of-the-te...>
Including money from Y Combinator itself (which, may we never forget, was once helmed by Sam Altman).
If you're looking for unbiased tech news and genuine user sentiment, you probably shouldn't be getting it from one of Silicon Valley's favorite VC firms.
Didn't he get fired because enough of the other YC partners couldn't trust him? If so, it's a pity for him that -at that time- he didn't have the CEO of a software megacorp to call on to reverse the decision and shitcan the people who made it.
I don't think suggesting that people are stupid is a very good argument. I think part of the reason people are for it is post like yours which are rude and not in good faith.
I'd imagine why as many people are for it on HN is for many people, it just works. Now imagine this snark about not being as clever as you think you are just because you enjoy using a tool. What's that going to do? Cause you to dig your feet in because you're being told the real reason you like these tools is because you're not very clever or you take things as a personal criticism. Yes, being told you think you are clever but you aren't really is personal.
I don't think you're posting in good faith. I certainly don't think you're going to get the real reasons people support this stuff when you act like they're stupid or suggest a lot of this is being turfed. People are using these tools because they work.
Your post says more about you than it does supporters.
You'll note I didn't say anyone was "stupid," if anything, I was actually saying those people were myopic and arrogant. I recommend you familiarize yourself with the concept of "engineer's disease," once you understand it, you'll see it all the time on HN.
> I don't think you're posting in good faith. I certainly don't think you're going to get the real reasons people support this stuff when you act like they're stupid or suggest a lot of this is being turfed. People are using these tools because they work.
> Your post says more about you than it does supporters.
I don't think suggestion that people are posting in bad faith is a good argument. I also detect a fair bit of binary thinking from you. All I did was post a non-exhaustive list of different reasons for that kind of content to answer the comment I was replying to.
You didn't even try to suggest that a lot of people are using these tools for positive reasons. Who is being myopic and arrogant? Who is engaging in binary thinking?
Consider that some people simplye find that these tools work for them.
I've worked with a lot of people. Many of them are that stupid. Suggesting that these stupid people do not exist, that -if they did exist- none of them are professional programmers, and -even if they were programmers- they do not have the means to voice support for something that they're smitten with is not a good argument.
I'll also vaguely gesture at the various studies that find that the majority of conversation -by post volume- on Internet Forums is done by a small number of people. Never doubt the ability for a few loud folks [0] with lots of time on their hands to dominate a conversation. This goes like 10,000x when there's a huge amount of money sloshing around in said industry and said industry really needs to maintain a positive public reputation.
[0] ...and -sometimes- their sockpuppets...
And I explained why I disagreed! One "subtle" thing to note is that the claim isn't "Everyone who uses LLM-based tools and writes about the success they had is stupid.".
I couldn't disagree more!
If you haven't already, you should watch this topical presentation from Greg K-H that was posted to HN like a week ago. [0] It's about an hour long, but you can probably watch most of it at 120->150% speed.
Notice how his praise for the tools differs from so much of the expressed positive sentiment for said tools here on HN. The "low-horsepower" people I've known tend to praise things in a way that smells very much like astroturfing. The "high-horsepower" people I've known both express very nuanced opinions and tend to know better than to burn a ton of time squabbling with Internet Randos in a place like this... they have much better things to do with their lives.
Me too!
I think that's a good observation, but I wouldn't say those people are actually astroturfing (noting you only said they smell like that).
But that "low-horsepower" vs "high-horsepower" distinction probably has a big bearing on people's attitudes towards LLMs. "Low-horsepower" people are simultaneously much more easily impressed and often lack the skill to properly evaluate the output they're getting. It's often been remarked that LLMs feel awesome for topics you know nothing about. It's kinda like Gell-Mann Amnesia Effect for people who were never knowledgeable about any topic.
[1] "The phenomenon of a person trusting newspapers for topics which that person is not knowledgeable about, despite recognizing the newspaper as being extremely inaccurate on certain topics which that person is knowledgeable about." (https://en.wiktionary.org/wiki/Gell-Mann_amnesia_effect#Engl...)
I agree with everything you've said here [0] as well as this bit I'm quoting specifically. I would strongly disagree with any claim that positive reports from "low-horsepower" people are likely to be astroturf. I'm just noting that a lot of astroturf reads like the way the "low-horsepower" people I've known tend to praise things.
[0] ...at least with whatever the comment said when I wrote this one ;)...
I think we're in agreement.
> I'm just noting that a lot of astroturf reads like the way the "low-horsepower" people I've known tend to praise things.
The "why" behind that would be interesting to know. It could just be that we assume astroturf would be low-effort, and that praise is low-effort (even if that's the most effort the "low-horsepower" person can muster), so they read the same.
Ditto!
> The "why" behind that would be interesting to know.
I'm going to briefly noodle on about why I think astroturfing generally Sounds Like That, and then attempt to answer your implied question.
So my hunch about the tone of astroturfing is that it's _kind of_ like 419 scams. [0] Perhaps none of this is new to you, but I write it down anyway:
1) Creating detailed praise that withstands scrutiny is a lot of effort, whereas expressing positive sentiment in very broad strokes is extremely easy. If your strokes are broad enough, you can probably often reuse the praise, saving even more time and effort. Similarly, defending criticism of detailed praise is much harder than defending broad-strokes positive sentiment.
2) If you want to hire live people to astroturf on message boards, people able to effectively defend detailed praise are surely going to be much more expensive than those who can only handle the vague broad-strokes stuff.
3) The target of this astroturfing deliberately isn't someone who can scrutinize detailed praise... in part because of how much work it is to author that praise, and in part because the people with power and/or purchasing authority are the ones you want to target... and it's often but not always the case that those folks lack one or more of the time, background information, or horsepower required to perform that detailed scrutiny in the first place. So, why waste the time, effort, and money?
Now please excuse the dogshit writing style of the following attempt to address your question, I'm definitely having a "low-horsepower" day today...
But why do "low-horsepower" people tend to praise in that way? speaking as someone who definitely has "low-horsepower" days, and who has occasionally been an absolutely godawful teacher, it might have something to do with the whole "You can't effectively teach someone something if you don't have a solid understanding of it yourself." thing? Both detailed praise and good instruction require a detailed understanding of the thing in question.
So part of the reason might be that "low-horsepower" folks tend to have just a surface understanding of the thing they're praising, so they cannot really create detailed praise. [1]
Both praise and instruction invite questions from the folks who are around to hear it. Folks who are capable of feeling shame often feel shame when when they present themselves as someone who is far more knowledgable about something than they actually are and questioning reveals the deception.
So another part of the reason might be to avoid in-depth questioning -both to determine the justifications for their praise, as well as to gain an in-depth understanding of parts of the thing being praised that the praiser is entirely unequipped to provide-. If one gets a vague "IDK, it's just good, right? It makes my work so much faster! I just like it!" response to your request for praise then that's a signal that you're not likely to get much more information if you ask probing questions. _Maybe_. This paragraph I'm super not confident in.
I'm happy to chat more about this, and will spend some time over the next day or so trying to phrase the gunk that's in my head better... but I can't make any promises that I will succeed at that.
[0] ...AKA "Nigerian Prince" scams...
[1] To downvoters: There's absolutely nothing wrong with this! Knowing just enough about your work and how to do it to satisfy your boss and keep your paychecks flowing is a solid way to live one's life. Precious few people do work that they like, let alone work that they're interested in... and it's my bone-deep belief that the best future for all humanity is one in which no one has to work unless they either want to or if having a job of some kind truly makes their life far better than not having one. Work sucks for most folks and putting in just enough effort to get the work done to the standard required is a fine way to engage with it.
People are writing entire apps in days. You're just jaded if that isn't impressive to you.
You:
You're again reverting to the only reason people like these is because they're stupid.
If you haven't already, you should watch this topical presentation from Greg K-H that was posted to HN like a week ago. ... Notice how his praise for the tools differs from so much of the expressed positive sentiment for said tools here on HN. ... The "high-horsepower" people I've known both express very nuanced opinions ...> people are simultaneously much more easily impressed and often lack the skill to properly evaluate the output they're getting
Yes, this is the exact sentiment expressed. That the praise you see is because of these people with low-horsepower who aren't capable of evaluating their own experiences.
You really should! It's positive sentiment about LLM-based tools that comes from a "high-horsepower" individual. Given your commentary here, it really seems like you've not encountered much of that sort of thing, so it'd really do you some good to load it into to your databanks for use in future discussions.
Do yo believe there are positive reasons people are adopting these models? Because there is nothing in your responses that suggest you do. I, as somebody who rarely uses these models but understands why people might, don't see any attempt at being unbiased.
Can you please review https://news.ycombinator.com/newsguidelines.html and stick to them in the future? We'd appreciate it.
Frankly I don't think you've understood what I was saying but it doesn't matter. You keep saying I'm misunderstanding you but your bias is very obvious and I don't think you've answered why people are praising these models.
Can you please review https://news.ycombinator.com/newsguidelines.html and stick to them in the future? We'd appreciate it.
With that many elements you'd be better off storing the elements as a tuple or a struct, but people often reach for a generic array container regardless.
Related: the Rationalists display this very well (though it's certainly not unique to them). The jump from "I think my beliefs through carefully, therefore I am reasonably rational" (likely already false, or at best true in a limited domain) to "I am rational, therefore my beliefs are well-thought through" is actually very small in the human mind, despite the latter being the opposite of rationality.
Also, the existence of Gell-Mann amnesia should be taught to everyone!
> 3. I think there's a certain amount of defensiveness from AI enthusiasts.
I've encountered this as well. I think this is also a huge part of why you see so many people loudly exclaiming "every programmer uses LLMs now" despite that not being even close to true. If you criticize LLMs, it (socially, to them) counts as criticism of their users as well. As well, bosses are probably the most excited about LLMs because they hope to fire us all soon, so it can be pointlessly self-destructive to speak negatively about LLMs in a professional context right now.
That, plus AI boosters tend to be incredibly annoying about it, bringing it up at every possible opportunity, changing every technical discussion to be about them, suggesting them for the stupidest use-cases, loudly proclaiming nonsensical productivity boosts (perhaps you feel 5x as productive because you're forcing other people to spend their time fixing your shitty slop code?) etc. So as someone who does not use LLMs I avoid talking to them.
I'm not a booster, I barely use these tools but what I notice is the two apparent "sides" that people have shifted into and really I don't think that' necessary because these are just tools that have adnvantages and disadvantages.
I will cede this: it indeed seems like you can throw together a shitty half-working prototype very quickly using LLMs. There's been a deluge of half-working prototypes, but years in and where are all these projects now? No study has concluded anything near a 5x improvement.
Besides, my actual opinion is that even if these actually gave a large productivity improvement, it's wouldn't be worth it. The negatives are simply too large:
- the concentration of power they are pushing for;
- the environmental effects of training;
- the disinformation they spread at scale (which in practice is their primary use-case);
- the adverse effect on a shared reality from the fact that they cannot reliably answer questions, yet are presented as doing so;
- the adverse effect on a shared reality from their sycophancy (do the rich surrounded by yes-men seem in touch with reality?);
- their anti-social nature, as they discourage engaging with other people in favour of solipsism masked by a poor simulacrum of personhood;
- their weakening of labour power because of the managerial class' belief that they will soon replace workers (whether or not they actually can is not particularly relevant);
- the large-scale plagiarism they enable;
- lastly, their weakening of already incredibly lax standards of engineering from our field.
The promise of more software does not even come close to outweighing all of that.
I can't speak to you other issues but these are here to stay and I think the only realistic way forward is to learn how to use them.
I guess let's just ignore the evidence, then.
> I can't speak to you other issues but these are here to stay and I think the only realistic way forward is to learn how to use them.
People only say this about things that they want. Ignoring the consequences of one's actions is not a neutral position. Societies are perfectly capable of banning things (yeah, you can probably sneak a local model through a hypothetical law, but it would end the industrial scale problems).
There isn't any evidence. It's pretty much all anecdotal. I can only speak from my experience.
> People only say this about things that they want. Ignoring the consequences of one's actions is not a neutral position. Societies are perfectly capable of banning things (yeah, you can probably sneak a local model through a hypothetical law, but it would end the industrial scale problems).
We're not going to ban these models, though. Which means we have to figure out how best to live with them.
> A recent study of actual coding practices across hundreds of firms finds that human code review forms a significant “bottleneck” for the overall efficiency of AI coding tools, resulting in “little evidence that firms increase software output or reduce employment” by using them.
https://metr.org/blog/2025-07-10-early-2025-ai-experienced-o...
> When developers are allowed to use AI tools, they take 19% longer to complete issues—a significant slowdown that goes against developer beliefs and expert forecasts. This gap between perception and reality is striking: developers expected AI to speed them up by 24%, and even after experiencing the slowdown, they still believed AI had sped them up by 20%.
LLM boosters always cry foul with this one because newer models exist (see next item), but the fact is this is evidence that programmers are remarkably bad at estimating their own productivity.
https://metr.org/blog/2026-02-24-uplift-update/
> For the subset of the original developers who participated in the later study, we now estimate a speedup of -18% with a confidence interval between -38% and +9%. Among newly-recruited developers the estimated speedup is -4%, with a confidence interval between -15% and +9%.
With the rather large caveat:
> Unfortunately, given participant feedback and surveys, we believe that the data from our new experiment gives us an unreliable signal of the current productivity effect of AI tools. The primary reason is that we have observed a significant increase in developers choosing not to participate in the study because they do not wish to work without AI, which likely biases downwards our estimate of AI-assisted speedup. We additionally believe there have been selection effects due to a lower pay rate (we reduced the pay from $150/hr to $50/hr), and that our measurements of time-spent on each task are unreliable for the fraction of developers who use multiple AI agents concurrently.
> We're not going to ban these models, though.
How could you possibly be confident of that?
> How could you possibly be confident of that?
I just don't see any indication that we're anywhere close to that. It would take a rather big shift in governance.
I see a lot of people going even further and claim that "it's impossible not to use LLMs now."
Not just among programmers talking about code either, you see it with non-tech people defending their overly long emoji-ladden AIspeak text posts.