I'm something of an LLM researcher myself, but there's just too many LLM-related posts on HN imo. Almost anything with the word "LLM" in the title finds its way to the front page, while other actually programming/hacker-related posts get ignored.
Yes and no. A lot of them are low value, but there are some that have been hugely useful to me, either examples of how to do things, or tools that I now use regularly, or posts that have made me view problems I’m trying to solve through another lens.
It’s a hot topic and so of course there are going to be a lot of posts. A few years back there were lots of blockchain posts.
I think anyone could pick a topic to which they feel a certain level of antipathy and build a case that there are too many posts of that type.
Au contraire, it's an incredible efficient way to gain high level insights into a new field that I just don't have time to dive into otherwise. And HN spans all the layers, from low level instruction set developments to the commercials practicalities of hosting your own AI. There is no comparable source of distilled wisdom that I know of.
Of course if you just hate AI instinctively like my wife then yeah, I'm sure there's way too much AI content.
It’s more that that they aren’t innovating or doing anything novel. Eg. “Tensorcracked is hiring to build document processing with LLMs” are fairly irritating
I am indeed tired of them, but as retired engineer I have the luxury of living (in my imagination) in the pre-AI world. Being a programmer today is like being someone in the early 20th century who enjoyed the craft and mastery of being a cold type typesetter. You could continue to do letterpress printing even as wave after wave of new technologies came along, but something which was necessary and state of the art was transformed into either a boutique small business or a quaint hobby, like calligraphy.
I understand what you’re saying but while artisan printing and calligraphy are visible articulations of the craft, skill, care and attention that go into their making, do you think the same can be said for “artisan, human-crafted code”?
It’s a genuine question, and whether your answer is yes or no I’d be interested to hear your opinion on why.
My opinion is it’s unlikely that human-crafted code will have the same inherent value that human crafted typography does, for all sorts of reasons to do with the way we can consume physical artefacts of language vs the way we consume code.
But I’ll also admit that my thinking in this - as regards code - is not particularly nuanced because I commission code rather than write it, so I’m interested in “does it do the job effectively” vs “is it inherently beautiful”.
When I was in the workplace, I enjoyed the actual practice of analyzing problems and writing code to address those problems. There was a certain satisfaction in knowing that I wasn't engaged in some behavior I enjoyed that was less than efficient for my employer -- there didn't exist an AI that I could unleash on the problem or that I could consult to help me write the code. Now (I imagine) a programmer who enjoys the act of programming cannot in good conscience write code "by hand" -- he is wasting time getting the product out unless he utilizes the AI. But from subjective standpoint of the programmer, using an AI is like using an AI to do the New York Times crossword for you -- where's the fun in that?
That’s an interesting point that I’d not considered. You’re essentially saying “code for code’s sake”, or “writing code in order to achieve a personal rather than industrial objective”, right?
By which I guess we are talking “personal code projects that don’t necessarily serve a commercial end”?
Someone could code for the same reasons a person might make art.
Professional artists don’t always ignore the commercial aspect of their work. But often those projects allow making work that is not driven by commercial concern.
That doesn’t make an artist’s paying work any less art when an artist considers making art the important outcome rather than the made artifacts.
Not only here but now people as close as my family have heard of Deepseek through other channels and are telling me I need to get on top of the trends now in order to stay relevant in my industry, or I'll be left behind.
The combination has sapped away any interest I might have had in pivoting to an AI career.
I systematically hide them since I'm not interested in them and they just pollute the main page, and you're right that they've become a bit excessive. Looking at my list of hidden posts, there are 26 AI-related items from the last 2 days alone.
Edit: 10 minutes after this comment I checked the front page and had to hide another half-dozen posts.
Especially comments on productivity boosts and outright lies when someone “builds a few side projects in the last month with AI”. It feels either like a sponsored roleplay or a competence level where building a jsx form is a side project.
Yes, to an extent; but I've seen enough hype cycles come and go not to be too worried about it. A few years ago I was tired of all the blockchain posts, and a few years before that I was tired of everything being about the cloud; in a few more years to come I am sure I will be tired of seemingly everything on HN being about whatever the next obsession is.
Yes, there is little to no major breakthroughs why keep posting them? Like wow this AI can run 5% better than the last one, jump from 68/100 score to 71/100 score because it uses 100% more GPU power, like, ok?
I just skip most of them but if you wanted to filter them out uBlock Origin can do it with a couple rules. [1] Adjust as desired, e.g. LLM|GPT|openai|deepseek and so on...
No. I find them - and the broader societal changes around their emergence - the most fascinating thing happening in tech at the moment.
I'm old enough to have watched the internet become taken from granted despite how fundamentally it changed our lives. Now, I'll watch humanity create intelligence on demand, transform our lives again, and take that for granted, too.
This isn't bad, either. It feels like true progress for people to become bored with things that would have astonished us only a few years ago.
Get ready for the rest of 2025, as this is the least we'll be talking about LLM's (or genAI for that matter) for the rest of the year and likely this tech epoch
Yes. There was a time Literally the first 2 page are mostly AI / LLM. I really wish all the post about FOSDEM would get instant front page listing.
And then politics, there is a guidelines about no or little politics submissions on HN. There used to be a grey area where it is politics Relating to tech gets a pass. Somewhere along the line this was gone.
I think opportunities to engage with the unfamiliar is one reason for HN’s stickiness among HNers.
If LLM’s are familiar at a professional level, then internet status points aside, engaging with LLM’s at a “lay level” is probably not why HN sticks for you.
Even if many other people are excited by them…particularly since endless-September is a feature not a bug on HN.
Submitting content that interests you is the most reliable way to get content that interests you on HN. Good luck.
Yes, I am -- but, admittedly, not only on HN. There's far too much of that stuff everywhere on the Internet nowadays. So, OK, to the extent that HN is supposed to be a reflection of the larger 'Net... But is it, really? I'd say not: It's supposed to separate the wheat from the chaff. And there's too much "AI" / LLM stuff, both elsewhere and here, for it all to be wheat.
(Side note on usage: Dunno whether OP is a native English speaker or not, but I think I would have put that as "Are you tired of all the LLM posts on HN?" Yes, it's only a tiny definite article -- but idiomatically, it changes the meaning from an absolute "Are you tired of each and every LLM post on HN?" to a more general-collective "Are you tired of the huge amount of LLM posts on HN?". And isn't that really what we're tired of? There might still be some few of them we could tolerate.)
It’s a hot topic and so of course there are going to be a lot of posts. A few years back there were lots of blockchain posts.
I think anyone could pick a topic to which they feel a certain level of antipathy and build a case that there are too many posts of that type.
reply