Hacker News Board Hacker News, browsed like a classic message board
Live data from Hacker News

Total monthly number of StackOverflow questions over time

data.stackexchange.com

Post reply on HN
Print view View original

711–720 of 1001 posts

Re: Total monthly number of StackOverflow questions over time

#711

Post by cyberdick »

Alupis wrote: post #107
palata wrote: post #32

Earlier quoted context omitted.

Agreed. I personally stopped contributing to StackOverflow before LLMs, because of the toxic moderation. Now with LLMs, I can't remember the last time I visited StackOverflow.

People in this thread are missing another key component in the decline of StackOverflow - the more experienced you become, the less useful it is. The harder the problem, the less engagement it gets. People who spend hours working on your issue are rewarded with a single upvote. Meanwhile, "how do I concat a string" gets dozens or hundreds of upvotes. The incentive/reward structure punished experienced folks with chal…

There is also github issues discussions now which also helped in asking these niche questions directly to the team responsible. I dont ask questions about a library on SO I just ask it on the github of the library and I get immediate answers

Re: Total monthly number of StackOverflow questions over time

#712

Post by fireflash38 »

tempest_ wrote: post #375
sky2224 wrote: post #344

Earlier quoted context omitted.

This is a perfect example of an element of Q&A forums that is being lost. Another thing that I don't think we'll see as much of anymore is interaction from developers that have extensive internal knowledge on products. An example I can think of was when Eric Lippert, a developer on the C# compiler at the time, responded to a question about a "gotcha" in the language: https://stackoverflow.com/a/8899347/10470363 Devel…

This type of thing often lives in the issues / discussion tab of a github repo now a days, for better and worse.

A lot of valuable information lived/lives in email threads that might or might not be publicly archived.

Re: Total monthly number of StackOverflow questions over time

#713

Post by wartywhoa23 »

throw-12-16 wrote: post #484
atomicbeanie wrote: post #353

Has AI summarization led to people either getting their answer from a search engine directly, and failing that, just giving up?

RTFM

TFMs are not a thing anymore. Most of them are merely collections of sparse random dots one might join by sheer luck only, granted no other knowledge of the system being attempted to document.

Re: Total monthly number of StackOverflow questions over time

#714

Post by dbobbitt »

NewJazz wrote: post #213
specproc wrote: post #145

Earlier quoted context omitted.

What I always appreciate about SO is the dialogue between commenters. LLMs give one answer, or bullet points around a theme, or just dump a load of code in your IDE. SO gives a debate, in which the finer points of an issue are thrashed out, with the best answers (by and large) floating to the top. SO, at its best, is numerous highly-experienced and intelligent humans trying to demonstrate how clever they are. A bit l…

The fundamental difference between asking on SO and asking an LLM is that SO is a public forum, and an LLM will be communicated with in private. This has a lot of implications, most of which surround the ability for people to review and correct bad information.

The other major benefit of SO being a public forum is that once a question was wrestled with and eventually answered, other engineers could stumble upon and benefit from it. With SO being replaced by LLMs, engineers are asking LLMs the same questions over and over, likely getting a wide range of different answers (some correct and others not) while also being an incredible waste of resources.

Re: Total monthly number of StackOverflow questions over time

#717

Post by chongli »

nerdix wrote: post #495
NitpickLawyer wrote: post #478

Earlier quoted context omitted.

Still gh issues are better than some random discord server. The fact that forums got replaced by discord for "support" is a net loss for humanity, as discord is not searchable (to my knowledge). So instead of a forum where someone asks a question and you get n answers, you have to visit the discord, and talk to the discord people, and join a wave channel first, hope the people are there, hope the person that knows is…

Yeah, I suspect that a lot of the decline represented in the OP's graph (starting around early 2020) is actually discord and that LLMs weren't much of a factor until ChatGPT 3.5 which launched in 2022. LLMs have definitely accelerated Stackoverflow's demise though. No question about that. Also makes me wonder if discord has a licensing deal with any of the large LLM players. If they don't then I can't imagine that wi…

Discord isn’t just used for tech support forums and discussions. There are loads of completely private communities on there. Discord opening up API access for LLM vendors to train on people’s private conversations is a gross violation of privacy. That would not go down well.

Re: Total monthly number of StackOverflow questions over time

#720

Post by noduerme »

bazzargh wrote: post #702
noduerme wrote: post #658

Earlier quoted context omitted.

That's pretty nice ;) I once wrote this humdinger, that's still on my mostly dead personal website from 2010... one of my proudest bits of code besides my poker hand evaluator ;) The question was, how do you generate a unique number for any two positive integers, where x!=y, such that f(x,y) = f(y,x) but the resulting combined id would not be generated by any other pair of integers. What I came up with was a way to g…

It looks like you have typos? (x^2+y^2)+(sin(atan(x/y))*cos(atan(x/y))) reduces to x^2+y^2+( (x/y) / (x^2/y^2 + 1) ) - not the equation given? Tho it's easier to see that this would be symmetrical if you rearrange it to: x^2+y^2+( (xy) / (x^2+y^2) ) Also, if f(x,y) = x^2+y^2+( (x/y) / (x^2+y^2) ) then f(2,1) is 5.2 and f(1,2) is 5.1? - this is how I noticed the mistake. (the other reduction gives the same answer, 5.4…

Hhhhmm. Ok. So I invented this solution in 2009 at what you might call a "peak mental moment", by a pool in Palm Springs, CA, after about 6 hours of writing on napkins. I'm not a mathematician. I don't think I'm even a great programmer, since there are probably much better ways of solving the thing I was trying to solve. And also, I'm not sure how I even came up with the reduction; I probably was wrong or made a typo (missing the +1?), and I'm not even certain how I could come up with it again.

2^x & 2^y ...is the & a bitwise operator...???? That would produce a unique ID? That would be very interesting, is that provable?

Primes take too much time.

The thing I was trying to solve was: I had written a bitcoin poker site from scratch, and I wanted to determine whether any players were colluding with each other. There were too many combinations of players on tables to analyze all their hands versus each other rapidly, so I needed to write a nightly cron job that collated their betting patterns 1 vs 1, 1 vs 2, 1 vs 3... any time 2 or 3 or 4 players were at the same table, I wanted to have a unique signature for that combination of players, regardless of which order they sat in at the table or which order they played their hands in. All the data for each player's action was in a SQL table of hand histories, indexed by playerID and tableID, with all the other playerIDs in the hand in a separate table. At the time, at least, I needed a faster way to query that data so that I could get a unique id from a set of playerIDs that would pull just the data from this massive table where all the same players were in a hand, without having to check the primary playerID column for each one. That was the motivation behind it.

It did work. I'm glad you were curious. I think I kept it as the original algorithm, not the reduced version. But I was much smarter 15 years ago... I haven't had an epiphany like that in awhile (mostly have not needed to, unfortunately).

Post reply on HN