OK, I recognize that for some people that question may not make any sense, because after all, ChatGPT is a #generativeai engine and #stackoverflow is a community forum for technology.

Although I consider myself a senior executive first, I also have one foot in technology and #applicationdevelopment. In light of the latter, I spend a decent amount of my time researching and contributing to articles on Stack Overflow which is a community site that helps developers answer questions about technology in general, #programming more specifically, and all sorts of application-oriented questions.

In many cases I have used Stack Overflow to help me resolve syntax errors with particular programming languages such as #JavaScript or #PHP. Answers readily flow from experienced programmers and the amount of contribution these people make to the system has made my coding efforts over the year far more enjoyable.

But I have to be honest. In the last many months, I have been relying more heavily on ChatGPT to solve my programming dilemmas. I simply formulate a question in some cases with mildly sophisticated prompting, and the system will readily come back with code that answers the question.

I have been relying more heavily on ChatGPT to solve my programming dilemmas.

Ah, but is that true?

I recently read an article about how the answers that come back from ChatGPT are statistically more often incorrect and just wrong versus those that come from Stack Overflow. I actually think I believe that, but I also think it sort of misses the point. Creating a prompt in ChatGPT to answer your question is fairly straightforward and easy. Doing the same thing in a search bar on a community forum tends to bring a lot of noise.

But yet I still think that even that misses the point. The point for me is that the proposed code is almost always close enough to what I am looking for that it answers my question. I wouldn't push back too hard at somebody who said it also almost never works because I've also found that to be the case.

The point for me is that the proposed code is almost always close enough

But that, alas, is the point that is missed. The amount of time that it takes to structure a prompt for ChatGPT and get a “close enough” answer that guides you to a solution versus the significant amount of effort it often takes to find the same, mind you correct, answer on Stack Overflow does not compare in my mind.

I have often spoken about a lot of technology #disruption that is going to occur because of generative AI. What I have been surprised about is that even a community forum, and one as valuable as Stack Overflow, could fall into that same category.

Please let me know which source you tend to use more often these days