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

What AI did to stackoverflow in a graph

data.stackexchange.com

Post reply on HN
Print view View original

61–70 of 615 posts

Re: What AI did to stackoverflow in a graph

#61

Post by Ekaros »

hilariously wrote: post #10
mid-kid wrote: post #3

Except for covid, it seems the decline was already there.

Yep, SO was dying before the GPTs - in some ways it was baked in the original SO design - to become the canonical source of information about programming stuff. Many people talk about the negativity, and they are right, but I think the reason more than anything is the waiting. On SO a good question might get answered in minutes (if it was easy and someone was karma farming) but it could be days or weeks for general p…

Somehow I feel that whole idea was 50% broken from start. In field changing as fast as programming you simply cannot have canonical source that is done. There are areas which are constant or slow moving. But too many others are in churn. Trying to apply same rules for both is clearly insanity. And this is really the failure of SO...

Maybe if in some cases stuff actually got deprecated and points did actively decay it might have worked... But can't remove power from those who had it for years...

Re: What AI did to stackoverflow in a graph

#62

Post by adamtaylor_13 »

AI might've delivered the final blow, but Stack Overflow was in decline LONG before LLMs came on the scene.

I read a great article not long ago outlining the full series of events and changes that led to its downfall. I wish I could find that article, but I've forgotten where it was.

Re: What AI did to stackoverflow in a graph

#63

Post by j_maffe »

embedding-shape wrote: post #42
khalic wrote: post #11

Seing a bell curve and singling out a factor that appears only for the 15% of the total time demonstrates some pretty extreme tunnel vision Edit: https://postimg.cc/n9nZGLmb

Yeah, I don't understand the HN title. The "downfall" seems to have began in 2018-2020 sometime, what AI was launched and popularized at that point that would have killed SO? LLMs were basically useless until GPT3 which appeared in middle-2020 sometime, after the downfall seemingly already had begun.

I'd call it significant that the number of questions halved within one year following the release of ChatGPT, the biggest relative or absolute rate of decrease in the timeseries.

Re: What AI did to stackoverflow in a graph

#64

Post by khalic »

j_maffe wrote: post #59
khalic wrote: post #11

Seing a bell curve and singling out a factor that appears only for the 15% of the total time demonstrates some pretty extreme tunnel vision Edit: https://postimg.cc/n9nZGLmb

You don't just fit a Gaussian distribution to a timeseries dataset. That's not what a Gaussian curve is designed for at all. https://www.explainxkcd.com/wiki/index.php/1725:_Linear_Regr...

You are confusing a Probability Density Function (PDF) with a phenomenological curve fit. No one is claiming that time is a random variable drawn from a normal distribution.

Re: What AI did to stackoverflow in a graph

#65

Post by pydry »

embedding-shape wrote: post #42
khalic wrote: post #11

Seing a bell curve and singling out a factor that appears only for the 15% of the total time demonstrates some pretty extreme tunnel vision Edit: https://postimg.cc/n9nZGLmb

Yeah, I don't understand the HN title. The "downfall" seems to have began in 2018-2020 sometime, what AI was launched and popularized at that point that would have killed SO? LLMs were basically useless until GPT3 which appeared in middle-2020 sometime, after the downfall seemingly already had begun.

Add it to the list:

- the downfall of junior devs

- bad hiring market

- layoffs in practically every sector

theres a ton of things where AI took credit for a trend that had already started before it started being even halfway capable.

Re: What AI did to stackoverflow in a graph

#66

Post by adamtaylor_13 »

ahmetson wrote: post #44

When an ecological shoes company pivot to AI, I wonder why StackOverflow executives don't pilot for AI now.

There's still a question of what that would do for them. I haven't used SO in years because I haven't needed it in years.

That's primarily due to AI answering my technical questions.

Re: What AI did to stackoverflow in a graph

#68

Post by j_maffe »

khalic wrote: post #51
airstrike wrote: post #21

Earlier quoted context omitted.

This isn't really a bell curve.

check my other response, it's very much a bell curve, statistically speaking https://postimg.cc/n9nZGLmb

> statistically speaking

That's a very big word you're using there for what is basically making shapes out of clouds. A bell-curve is the amortised function of a random variable with a mean and standar deviation. What does that have to do with a timeseries dataset?

Re: What AI did to stackoverflow in a graph

#69

Post by khalic »

j_maffe wrote: post #63
embedding-shape wrote: post #42

Earlier quoted context omitted.

Yeah, I don't understand the HN title. The "downfall" seems to have began in 2018-2020 sometime, what AI was launched and popularized at that point that would have killed SO? LLMs were basically useless until GPT3 which appeared in middle-2020 sometime, after the downfall seemingly already had begun.

I'd call it significant that the number of questions halved within one year following the release of ChatGPT, the biggest relative or absolute rate of decrease in the timeseries.

"How AI precipitated SO's downfall" would be a correct title then

Re: What AI did to stackoverflow in a graph

#70

Post by j_maffe »

khalic wrote: post #64
j_maffe wrote: post #59

Earlier quoted context omitted.

You don't just fit a Gaussian distribution to a timeseries dataset. That's not what a Gaussian curve is designed for at all. https://www.explainxkcd.com/wiki/index.php/1725:_Linear_Regr...

You are confusing a Probability Density Function (PDF) with a phenomenological curve fit. No one is claiming that time is a random variable drawn from a normal distribution.

> No one is claiming that time is a random variable drawn from a normal distribution.

You are doing that implicitly by fitting a Gaussian curve.

Post reply on HN