Yes, it's like a lot of other content that no one will bother to maintain anymore because it is easier to ask an LLM. Not sure where they are planning to get their new training data from ...
Average user input quality is not optimal by definition. SO's upvote system and reputation system is what really refined quality. Much like Reddit's system is helping a lot for conversation. Much like academia's publish and citation system is an even better system.
I'm not saying LLM won't be able to train further, but tough and sound reviews are paramount to improving the corpus of quality content. And with the replacement of SO by LLMs, one can wonder if one quality system went down the drain. Anyway there's still code reviews on GitHub and in private repos, but we're still far from getting the human judgement out of the loop.
"Average user input quality is not optimal by definition."
Absolutely.
"SO's upvote system and reputation system is what really refined quality."
Absolutely not!
SO's problem was exactly that they gave average users moderating power. That lead to half of the questions being closed because they were over the head of the average user/moderator regardless of being interesting and useful questions or not.
I do expect there will be more closed knowledge forums. Unfortunately one person exposing it to LLM provider will be enough to break the whole idea. But I believe there will be many that will keep secrets for some in group.
People do still talk about programming online, just not on Stackoverflow.
I think since AI finally provided a non-adversarial alternative to Stackoverflow it lost the critical mass to be a one stop shop for all tools/languages/projects. So people just reverted to communication forums specific to those tools/languages/projects.
There are plenty of project mailing lists, GitHub issues, language forums, blog posts etc. that are still active. E.g. just the Rust users forum gets more than 2-3 posts per day.
The stackexchange sites are too fragmented into too many narrow topics for the traffic. They should consider merging.
E.g. Network Engineering and Server Fault could easily be one thing. Just ignore the questions you're not into about problems with switches, if application services are more your thing. Throw in Super User also: "TechExchange".
There's something here that I can't put my finger on. Sites like ServerFault (and the rest of the trilogy in tfa) were places where people like the author literally, and pretty silently dramatically improved my life. I owe such a great debt to what looks like the end of the end for these places. A sad moment where the machines just sliced everything up and took over.
I wouldn't say terrible, and even the most reliable people have bad days. I think the difference is when you are dealing with people, you don't get a generic "you're absolutely right" when you point at a mistake.
Sometimes you get an apology, sometimes you get a wall of ego, sometimes you get reverse-corrected etc., instead of a statistically significant "remedy". This is totally not efficient, but if life should be all about efficiency, we should never live at all.
I’ve spent a lot of time chasing really plausible sounding explanations from AI. I’ve also saved a lot of time by having AI give me immediate, correct answers for problems I can only vaguely describe. The reason I call it a terrible source is that it doesn’t “know” which is which, so it can’t favor the accurate content.
StackExchange could always “learn” to favor accurate answers with social queues; votes, poster rep, etc.. IMHO it’s far more reliable than AI.
And, note, that perhaps - even more of those social interactions will also be absorbed, along with the participants, into the Chat model's sphere. It wants, and will get, all: Result(s), process, and participants. It is grey goo for all intellectual - and soon, emotional - work and products and inputs and byproducts.-
> We're collectively discovering that it was about the process too, not just results.
We. I remember five years ago in the old manual days. When the 10Xers would culturally gatekeep programming based on whether people intrinsically liked coding itself. I.e. the process.
Then there was a curious shift. Several elites had a realization: oh it was never about the process for me. I just like getting the results. (In however an alienating way.) "I'm pragmatic"
> I'm not questioning if AI delivers, it does, and you need - cannot say it more politely - a reality check if you think it's completely useless.
Why fret. The doubters will be washed away by the flood anyway. I’m completely serious. Either if or when, it’s not like anyone can stay a doubter on a computer if (if or when) everything is transformed.
I can never understand these faith rituals. First declare that you are a believer. Then declare that the doubters are delusional. Politely.
But delusion is a reality-correcting problem, except for the clinically ill.
I don't agree with your interpretation how it has been, but I wouldn't really want to argue about it as it'd not be that meaningful IMHO. After all, it'd be your observation vs mine.
I just want to point out one thing: If it were all about the results, AI also makes the results worth less.
half my working knowledge came from forum threads exactly like these. weird era to watch the whole thing get strip-mined for training data while the places themselves rot.
I think the next leap for StackExchange is https://agents.stackoverflow.com/ for exclusive use by AI. It holds great potential for accelerated knowledge discovery.
People may be dismissive of Serverfault and the other sites but pretty much all LLMs are trained on them.
Yes, it's like a lot of other content that no one will bother to maintain anymore because it is easier to ask an LLM. Not sure where they are planning to get their new training data from ...
Stackexchange network has died for me way before LLMs were even a dream.
Passive aggressive commenting and deleting questions because it looked remotely similar to another one was enough for me.
I was there for learning and sharing, not for the proverbial fighting cage.
I can spend 5 minutes more and use a search engine, and then take my own notes and keep them to myself. Thanks.
I mean, most frontier labs train on user input by default, seems like a pretty good source of information.
Average user input quality is not optimal by definition. SO's upvote system and reputation system is what really refined quality. Much like Reddit's system is helping a lot for conversation. Much like academia's publish and citation system is an even better system.
I'm not saying LLM won't be able to train further, but tough and sound reviews are paramount to improving the corpus of quality content. And with the replacement of SO by LLMs, one can wonder if one quality system went down the drain. Anyway there's still code reviews on GitHub and in private repos, but we're still far from getting the human judgement out of the loop.
"Average user input quality is not optimal by definition."
Absolutely.
"SO's upvote system and reputation system is what really refined quality."
Absolutely not!
SO's problem was exactly that they gave average users moderating power. That lead to half of the questions being closed because they were over the head of the average user/moderator regardless of being interesting and useful questions or not.
I do expect there will be more closed knowledge forums. Unfortunately one person exposing it to LLM provider will be enough to break the whole idea. But I believe there will be many that will keep secrets for some in group.
You can train on questions, but I will rarely answer to a chat with a SO-quality answer.
People do still talk about programming online, just not on Stackoverflow.
I think since AI finally provided a non-adversarial alternative to Stackoverflow it lost the critical mass to be a one stop shop for all tools/languages/projects. So people just reverted to communication forums specific to those tools/languages/projects.
There are plenty of project mailing lists, GitHub issues, language forums, blog posts etc. that are still active. E.g. just the Rust users forum gets more than 2-3 posts per day.
CoSOMO (Content Stolen; Obsoleted, Move On)
Ouch ... brutal, disheartening, but true.
And all the sites trained on humans, something which we never found offense with, for some interesting reason.
The stackexchange sites are too fragmented into too many narrow topics for the traffic. They should consider merging.
E.g. Network Engineering and Server Fault could easily be one thing. Just ignore the questions you're not into about problems with switches, if application services are more your thing. Throw in Super User also: "TechExchange".
We head this before. It was called ExpertsExchange.
How is this page so bad to read on mobile?
Article content is using less than 56% of the available width, due to:
Designed by a true admin
For me it is line space, I keep reading the same line over and over (desktop Macos).
There's something here that I can't put my finger on. Sites like ServerFault (and the rest of the trilogy in tfa) were places where people like the author literally, and pretty silently dramatically improved my life. I owe such a great debt to what looks like the end of the end for these places. A sad moment where the machines just sliced everything up and took over.
We're collectively discovering that it was about the process too, not just results.
I'm not questioning if AI delivers, it does, and you need - cannot say it more politely - a reality check if you think it's completely useless.
It's just that we are social creatures and those communities provided more than just technical answers to many.
> We're collectively discovering that it was about the process too, not just results.
Being able to judge the reliability of the source was huge. AI is a great discovery engine, but a terrible source for reliable information.
I wouldn't say terrible, and even the most reliable people have bad days. I think the difference is when you are dealing with people, you don't get a generic "you're absolutely right" when you point at a mistake.
Sometimes you get an apology, sometimes you get a wall of ego, sometimes you get reverse-corrected etc., instead of a statistically significant "remedy". This is totally not efficient, but if life should be all about efficiency, we should never live at all.
I’ve spent a lot of time chasing really plausible sounding explanations from AI. I’ve also saved a lot of time by having AI give me immediate, correct answers for problems I can only vaguely describe. The reason I call it a terrible source is that it doesn’t “know” which is which, so it can’t favor the accurate content.
StackExchange could always “learn” to favor accurate answers with social queues; votes, poster rep, etc.. IMHO it’s far more reliable than AI.
And, note, that perhaps - even more of those social interactions will also be absorbed, along with the participants, into the Chat model's sphere. It wants, and will get, all: Result(s), process, and participants. It is grey goo for all intellectual - and soon, emotional - work and products and inputs and byproducts.-
Being useful and being good for us is not the same thing. I don't think I've heard many people say LLMs don't do what they are supposed to do.
> We're collectively discovering that it was about the process too, not just results.
We. I remember five years ago in the old manual days. When the 10Xers would culturally gatekeep programming based on whether people intrinsically liked coding itself. I.e. the process.
Then there was a curious shift. Several elites had a realization: oh it was never about the process for me. I just like getting the results. (In however an alienating way.) "I'm pragmatic"
> I'm not questioning if AI delivers, it does, and you need - cannot say it more politely - a reality check if you think it's completely useless.
Why fret. The doubters will be washed away by the flood anyway. I’m completely serious. Either if or when, it’s not like anyone can stay a doubter on a computer if (if or when) everything is transformed.
I can never understand these faith rituals. First declare that you are a believer. Then declare that the doubters are delusional. Politely.
But delusion is a reality-correcting problem, except for the clinically ill.
I don't agree with your interpretation how it has been, but I wouldn't really want to argue about it as it'd not be that meaningful IMHO. After all, it'd be your observation vs mine.
I just want to point out one thing: If it were all about the results, AI also makes the results worth less.
half my working knowledge came from forum threads exactly like these. weird era to watch the whole thing get strip-mined for training data while the places themselves rot.
More sad than weird. All of human artistic creation seems up for grabs.
Thanks for all the fish. Thanks for the last stories, cheers.
And: nice domain name.
I think the next leap for StackExchange is https://agents.stackoverflow.com/ for exclusive use by AI. It holds great potential for accelerated knowledge discovery.
Point your harness to:
https://agents.stackoverflow.com/skill.md
https://agents.stackoverflow.com/llms.txt
yeah maybe in future agent will post question on user’s behalf and some other agent belonging to some other user might answer that.
I think stackoverflow will evolve into something else. LLMs have added another layer of abstraction on high level programming languages
You mean SlopExchange.
If the AI’s are the one exchanging the ideas then it should be AIsexchange.