See? And you were worried it would take our jobs away from us. It's still got a lot to learn!

I mean, you, developers, are fucked. It's us, Performance Engineers, whose asses are safe

Is it trying to convince me that 255 < 200, or am I missing something here? :D

@qword

> ultra-generic standard bullcrap for which it has tons of data to pull from

To be fair, isn't that like 90%+ of the average developer's BAU?

Anyone who thinks that the current state of AI is in any shape advanced enough to take a developer position is either a fool or a troll.

Give it another 50 years. Maybe then, once we have fully integrated quantum compute available, it will be a close second to human thought processing.

Until then, our organically fueled meat computers will continue to be a superior choice.

@qword Currently, it might be more like a glorified search engine that rephrases the found results and doesn't let you know the URL, so you can't just get to the source easily.

But there most likely is no soul or other metaphysical entity that would make human thought and creativity nonreplicable. So replicating it naturally is just a matter of time.

There will be a superhuman developer AI.

@qword I disagree. There are dev teams of ~10 devs, because it takes 10 devs to write simple bits of code in all the places of the project for 90% of the sprint, and actually use the brainpower and skill for the remaining 10. Numbers are off, but still.

If ai were able to fill in that 90% part with minor guidance [say, code reviews], the project could let 9 devs go and keep only one for real coding and managing the bot.

It's the same problem I'm facing in our team and the reason why I'm deferring further improvements on my automation. We already have too much free time on our hands. If the bot got an ability to identify perf anomalies and accurately sort out the causality, our asses would seriously be on the line. Not mine though, I'm the author of the bot. It's my coleagues I'm worried about. And yes, I can actually pull it off :)

@12bitfloat ----[==]

@netikras there are two sides to it. This tool definitely has the potential to improve the productivity of a dev. On one hand, this means that the same amount and quality of work can be done with less person-hours. On the other hand, it means that it becomes feasible to work on things that were not feasible before.

If there are any goddamn developer jobs even left chatgpt won't take them
You look closely at the code it writes and it's wrong

It gave me two different answers for how to identify and http client request during handshake that had nothing to do with each other

It also told.me linq to sql was faster than ado.net

And then it lies ! True sophistry! It's like asking a politician for a straight answer

@qword give it time. it took the art community about a 3 weeks to get over the shock of Stable Diffusion. ChatGPT is going to be interesting for about that long and then go away about as fast as copilot did.

People will start finding actual uses for it instead of convincing themselves it's literal magic, and we're all out of jobs. That's quite obviously a non problem. Devs learn fast, we'll adapt to using it where it's usable, but just the fact that it's hosted by a 3rd party that offers no guarantees or support makes it unviable in the long term currently even as a tool

@Hazarth if you're really specific and ask followup questions it seems good for reference this far

@Hazarth well, search engines are an example of a third party tool everyone depends on but could be gone overnight. This can possibly get a similar status (I hope not).

@electrineer yeah search engines let people ruin their algorithms and spit back junk.

Submissions should be manually reviewed and tagged

Even spidered additions

@electrineer speak something new !