[bhouston]

I disagree. StackOverflow grew to be many times the size of Fog Creek software. It was highly successful for a time.

You can not anticipate back then that AI would train on its Q&A and then provide the same service but fully integrated into IDEs. StackOverflow is being destroyed by Copilot and ChatGPT.

[bachmeier]

> You can not anticipate back then that AI would train on its Q&A and then provide the same service but fully integrated into IDEs.

What an interesting interpretation of events. For years (almost since the beginning) users were complaining about the experience. The SO leadership took pride in ignoring those complaints. Remember Jeff Atwood repeatedly boasting, without a shred of evidence, that the lack of discussion was what made SO popular? They dumped resources into automating the closing of questions in the coldest possible manner. They had out of control mods that everyone hated, but the answer was always that those very mods were the only ones that mattered for the site. They knew about the ridiculous dup trolling, and they did nothing to prevent it. The list goes on...

Now that they're in trouble, the explanation is that a new technology came along that nobody could have anticipated. No. They lost their base of new users long ago to Reddit and Discord. They had enough existing users that put up with their crap because SO was a standard part of the toolbox, and it's where Google took you. They did everything to make the site awful and now they're attributing their problems to new technology.

[busterarm]

The way we measure success is different. The history of the web is littered with companies that had a big exit and died/became-irrelevant.

SO destroyed itself several times over (first by its own moderators and then by the company itself). ChatGPT just nailed the lid shut.

Success to me looks like a lot of real users paying a lot of real money. SO never really had that. It tried to be a VC unicorn and was acquired by Prosus. Most of Prosus' other investments also look really questionable if we're headed for a global recession like it looks like we are...

[bhouston]

> The way we measure success is different. The history of the web is littered with companies that had a big exit and died/became-irrelevant.

Given a long enough timeline every companies dies or they pivot so extreme that they resemble nothing that they were to begin with.

[xmprt]

That might be true but it's a very nihilistic way to look at things. I believe SO had many more successful years ahead of it but decisions taken by the company and moderators sped that process up.

[mywittyname]

I don't think it nihilistic. The world changes; it changes rapidly; it changes dramatically. Successful businesses pivot, and as a result, often don't resemble the company that they used to be.

The end game of successful businesses is to become a conglomerate of sorts. Once a company owns enough brands/products across enough markets, they can shed the under-performers and expand into new segments and industries.

[xmprt]

I think we're in agreement. But the original commenter was stating that all companies die or change given a long enough timeline so criticizing a company for dying/changing is pointless since that happens to everyone.

[wizzwizz4]

Criticism of the company has been done to death, but I'd be interested to hear criticism of the moderator actions. (I'm not even sure which actions you and the grandparent are referring to.)

[busterarm]

The moderators seemed to collectively go insane with maintaining the purity of their individual fiefdoms.

Questions and answers both got moderated out of existence with high frequency with it being clear that the moderator didn't understand either. I remember this behavior becoming rampant throughout the community around 2012 and the moderators being incredibly toxic in Meta SO. I removed my profile and never looked back then.

[mschuster91]

> Success to me looks like a lot of real users paying a lot of real money. SO never really had that.

Or you follow the lean engineering model, like Whatsapp did. Millions of users paying a $ each are still millions of $, and Whatsapp prior to FB acquiring them had very very low operating costs - similar to Stackoverflow by the way. Both engineered their systems to run on incredibly low resources.

[eastbound]

> AI would train

No, Google made StackOverflow live and StackOverflow die by stopping its referencing.

SO should really pivot to the search business. They have everything: The culture of how to build a website that answers users’ questions, the culture to index pages, the culture to get an invested community. The only faulty mechanism here is Google, and SO would be a much better steward of how people want webpages to look like, than Google.

[berkes]

Somehow they never managed to get search right.

It's decent, but most often DuckDuckGo gives more relevant stack overflow answers than stack overflow itself does for the same query. Even though SO has far more information to tune the results on than DDG.

Another "proof" of how bad their search is, can be witnessed in the millions of duplicates. SO basically got an army of volunteers who know how to operate this search within their niche, to close duplicates, something that a decent search result would solve for a big part.

[bhouston]

> No, Google made StackOverflow live and StackOverflow die by stopping its referencing.

Maybe that as well. But I have definitely reduced my Google searches for technical answers and focus them on ChatGPT. So to me it doesn't matter if Google stopped referencing them or not. ChatGPT is phenomenal at answering technical questions with context specific answers and code snippets.