HN Explorer

Ask HN: Has HN become too skewed towards AI/LLM?

9 points by khaledh ↗ HN
This is more of a rant rather than a question. I find HN has become overwhelmed with AI and LLM content. For someone who has little interest in this area (I gravitate more towards programming/systems topics), I find myself jaded due to the large number of submissions in this single area. Don't get me wrong, I do like what LLMs can do and find their potential to be on par with inventing the computer itself, but I feel HN has lost the balance of topics it once had.

19 comments

[ 2.9 ms ] story [ 68.0 ms ] thread
mvkel
HN is tech-centric, and AI is dominating tech.

What would you rather see instead?

GiorgioG
AI is not dominating tech, hype about AI is dominating tech. I'm still knee-deep in the same shit that I've been doing for 20 odd years and AI hasn't changed a thing about my job. Occasionally ChatGPT helps me do something in 2 minutes that might have taken me 30 minutes to figure out, but it is not an earth-shattering productivity multiplier that the industry is hyping it to be.
mvkel
> AI hasn't changed a thing about my job

> ChatGPT helps me do something in 2 minutes that might have taken me 30 minutes to figure out

Increasing productivity by 1500% sounds pretty earth-shattering to me.

What would you rather see instead?

huesatbri
The average productivity increase is not 1500% though.
anonzzzies
Copilot saves my team enough time that we could split it up in 3 teams instead of hire 2 more teams of our team size to do that work. And we also didn’t hire new team members and are not planning to as we actually are running out of tasks. If this stuff gets any smarter, the firing will happen. Instead I am thinking of a new product instead to stop that. Pretty insane imho.
mvkel
Ok.

A "mere" 20% would be a four day work week with zero productivity loss.

MakeThemMoney
A four day work week requires a 25% productivity increase to result in zero productivity loss.

100 * 1,20 * 0,8 = 96

100 * 1,25 * 0,8 = 100

bartonfink
If AI increased your productivity 1500% I shudder to imagine how little you get done without it.
altdataseller
So what else would you like to see on the front page? Java, boring Postgres and HTML all are useful but there aren’t exactly any new innovative things that are happening everyday that warrant posting on the front page
codingdave
Fair, but HN is also about interesting discussion, and the same old day-to-day doesn't make for interesting discussion.

Even if AI is hype, it is more interesting than other hype because it is not just empty. There is nuance to this hype. AI is real, and can do real things. People are hyping it beyond what is real, but finding where that line falls is exactly why it is interesting.

matt_s
The amount of content about JS frameworks is certainly down. The hype cycle trend is very predictable I would hardly call it “dominating” in usage, certainly hype. Are we at the point yet where a knitting business is using AI?
dredmorbius
Discussion of AI is increasingly prevalent in many areas. I'm not convinced that the pitches for what AI can or might do will pan out, but I've no doubt at all that there will be a huge scramble in any number of areas (technology, investment, startups, media, governance, ...) to adopt, promote, explore, regulate, and abuse the tech.

Having been around and in the tech world at the time, the general feeling is reminiscent of the buzz, hype, and cons around the World Wide Web in the mid-to-late 1990s, as people were starting to hear about it, not understanding much of what it might do, and making all kinds of bold predictions of the utopia or dystopia which might result. The reality ... matched little of any of that discussion.[1]

The Web largely addressed distribution, taking the place of both the printing press and various channels (physical and broadcast) of media distribution. Generative AI is more about how content itself is generated and evaluated, and has the added twist that it is very much a black box, including to those who create the models and systems, which will undoubtedly change what its own ultimate effects and impacts are, though again there's a likely wide range of positive and negative factors.

I follow numerous media and podcasts, on a wide range of topics and languages, and AI has been popping up with extreme frequency in many of them.

Whether it's good or bad, this is undeniably big.

________________________________

Notes:

1. An exception to the poor predictions is Andrew L. Shapiro's The Control Revolution (1999), which I'd read for the first time a year or two ago. It seems to me to have done a quite good job of exploring both the upsides and downsides of the Web, and remains worth reading, though as a historical document now.

atleastoptimal
Sorry but this is the least people will be talking about AI for the rest of human history
moomoo11
lol. You missed the crypto nonsense, endless js shitworks I mean frameworks, fictional I mean functional programming, and so on.
smcleod
I'm glad there's a lot of AI/LLM talk - it's rapidly changing our industry and lives like we haven't seen in decades, I want to stay ahead or at least on top of the wave. This is quite different from the empty hype of crypto etc...

However, I would appreciate a new topic tag/stream in the top menu bar of HN a bit like 'Show', 'Ask' etc... specifically for AI/LLM content while keeping the main page more general. I think this would benefit everyone.