Altman’s remarks in his tweet drew an overwhelmingly negative reaction.
“You’re welcome,” one user responded. “Nice to know that our reward is our jobs being taken away.”
Others called him a “f***ing psychopath” and “scum.”
“Nothing says ‘you’re being replaced’ quite like a heartfelt thank you from the guy doing the replacing,” one user wrote.
As a developer on a team of 20ish that use AI heavily, daily, I’d argue there are proper ways of using it as a tool that substantially increases productivity and quality while also reducing bugs and improving maintainability.
But I would also say that 75%+ of people I encounter aren’t leveraging it properly or even close to properly. Also too many people use ChatGPT and Copilot which are just bottom of the barrel garbage and then wonder why the output is also garbage.
But I agree that too many people and companies lean into AI too heavily and incorrectly and there will be a reckoning. I’m all for it.
everyone using ai says some variation of this
acting like they do proper code review on this ai gen code but we’ve had so many examples of how people well before ai carried massive tech debt and cut corners so I have zero faith
much less the numerous examples of slop laden bugs making it to live with major companies in the us
ai just amplifies and obfuscates the problem
You do you I guess. AI slop can suck a dick and we’ve fired people for less.
AI causes problems, AI can amplify problems, but AI can legitimately assist, too.
imagine instead of ai it was your compiler
we all know they can have weird quirks but imagine the fail rates were like we are experiencing with current ai
would people keep these tools in their work flow?
if big business wasn’t so hard on to cut labor the answer would be a hard no
Its because your perspectives clash. They are saying the problems you have are caused by people behaving poorly, and you are saying its because the AI is inherently flawed.
They do have a point, if I were to pick up a hammer and throw it through a window, I can’t really blame the hammer for the broken window.
I do think we have to hold each other to some higher moral standard where we put people above profit, which would reduce a lot of abuse. Its astonishing that so many programmers would justify AI use knowing it may be causing harm to many others in their field.
except using your metaphor, everyone using the hammer has the same problems, and it is using said hammer as advertised, regardless of what we all know should be best practices
there have been multiple big fuck ups now and i think the reason why matters less than the fact the cause will not improve any time soon
I do try to give it a chance and use it every once and a while (most recently Claude Code; last year, Cursor), and it has been my experience that it personally decreases my productivity and quality. I found that even CoPilot’s autocomplete would introduce bugs if I decided to “trust” it and try to work too fast without meticulously reviewing every token generated. I have seen people I work with use AI to quickly create decent looking prototypes (i.e. bog standard, boring design), but I think this is still detrimental because they lose the full benefit of exploratory programming (and of course, the prototypes just have all kinds of faked data and functionality, glaring security problems, bad architecture). I’ve also experienced people submitting nonsense vibe-coded pull requests that would break tons of things they shouldn’t have even touched for the issue. I could see a less interested or overworked reviewer letting stuff like that through, which is why I think we’re seeing all these failures and bugs at these big tech companies. So for me, at least, I haven’t seen the benefit. Using CoPilot in VSCode actually caused me to go back to using nvim and lsp plugins :)
Agreed, Copilot auto complete is awful. I wouldn’t wish that on my worst enemy.
I don’t know what these people are doing who seem to claim it’s wonderful. Every single time I have tried to use one, it’s been completely clueless about the problem and wastes my time producing slop. I almost want to keep my code closed source because of how awful it is at generating anything. Maybe they’re just doing very simple web design or something, I don’t know
I even have started to hate Google and felt like their search engine is becoming very bad. Yandex has been returning more results and Google feels censored and replaced by generative AI answers.
Copilot isn’t bad, but generally I agree.
It’s a tool that can be helpful, or you can just create problems for yourself down the road.
It’s a lot like building a house. After all the drywall is up, it’s hard to tell if the studs are 18 inches apart or five feet apart, but you’re gonna find out eventually.
Would love to know how you finagle Copilot to be useful, the code it generates and the things it suggests make me throw my hands up like 9/10.
Once in a blue moon I try to use it again, and it’s laughable. It’s like I told someone non-technical to do something super technical and… it’s just 100% discard or 50% rewrite. I guess it just feels like it isn’t saving me any time so…why?
For context I’ve only used it in VS Code or Visual Studio. If there is some other avenue or process that’s better, let me know.
I’m usually just very targeted in what I ask it to do. I keep it to things I know will be in the basic reference books or on stack overflow. It basically just saves me from having to look up and apply existing examples to my code.
It makes for a pretty good ORM.
Are you talking about copilot, that bar that is on the side of Microsoft edge or something else? Maybe GitHub Copilot?
Because that copilot bar on Microsoft edge is literally a pointless waste of electricity
Oh, I was talking about GitHub copilot that’s in visual studio.