I mean, for the bugs in the screenshot, it is more than plenty, if even just 1% of bad code slips through.
And AI-generated code is extremely time-consuming and tricky to review, because you can’t assume there to be rhyme and reason to the changes, so I would be surprised, if they actually put in all the effort to properly review.
I suspect it’s largely more the result of failing internal organization. Like, a detached from reality and ideologically motivated faction with in corporate leadership has seized control of the company and fired anyone who told them they were being idiots or opposed their initiatives. People are probably getting promoted or hired to management positions based on their ability to tell leadership what they want to hear rather than their ability to actually run things. Everyone lower down has internalized that telling the higher ups what’s going on will get them fired and only is telling them what they want to hear. Resources and people got diverted away from projects that the management doesn’t care about (have no potential to drive growth), and they’re just assuming that the “increase in productivity of AI” will make up the difference. Now everything is melting down and their core product is losing market share while the new products intended to drive growth are failing to see meaningful adoption. Heads will probably role, but it’s unlikely it will be the people who are causing the problem.
30% is last year news. Now windows 11 in entirely developed by ai. How it works:
1: Ai hallucinates a new feature.
2: Ai agents generate enough code to meet investors demands.
3: Ai Qa agents are being gaslighted into verifying code.
4: The result is a feature which does not even work, while also ten breaking changes in core system features, which worked completely fine, but ai rewrote them anyway.
You forgot part 5 : AI scrapes MSlop forums for reported issues with the new update, then posts someone’s workaround as the official fix while the bots in the back hallucinate another fix for next week
Its a mix. Some devs are definitely lazy but from what I have heard there is also a big push for devs to deploy faster and to actively use AI or be punished. So there is incentive to just get code out to meet deadlines/expectations and move on to the next task. The amount of work being put on an individual dev is rising with these accelerated expectations in mind, and getting another dev to review your code takes time from them and their own stack of tasks so code review quality has fallen greatly. Not to mention the high likelihood that AI is also doing code reviews to make up for this and we can only guess how many reviews get approved saying “you’re absolutely right!”
just 30%?..
that should cause much less harm.
are the devs there that lazy? do they just not review the code?
I thought that especially these companies do code review and such…
I mean, for the bugs in the screenshot, it is more than plenty, if even just 1% of bad code slips through.
And AI-generated code is extremely time-consuming and tricky to review, because you can’t assume there to be rhyme and reason to the changes, so I would be surprised, if they actually put in all the effort to properly review.
I suspect it’s largely more the result of failing internal organization. Like, a detached from reality and ideologically motivated faction with in corporate leadership has seized control of the company and fired anyone who told them they were being idiots or opposed their initiatives. People are probably getting promoted or hired to management positions based on their ability to tell leadership what they want to hear rather than their ability to actually run things. Everyone lower down has internalized that telling the higher ups what’s going on will get them fired and only is telling them what they want to hear. Resources and people got diverted away from projects that the management doesn’t care about (have no potential to drive growth), and they’re just assuming that the “increase in productivity of AI” will make up the difference. Now everything is melting down and their core product is losing market share while the new products intended to drive growth are failing to see meaningful adoption. Heads will probably role, but it’s unlikely it will be the people who are causing the problem.
That’s what it looks like to me from the outside.
They got rid of their entire QA staff team. So the cancer spreads much more unchecked than usual.
30% is last year news. Now windows 11 in entirely developed by ai. How it works:
1: Ai hallucinates a new feature.
2: Ai agents generate enough code to meet investors demands.
3: Ai Qa agents are being gaslighted into verifying code.
4: The result is a feature which does not even work, while also ten breaking changes in core system features, which worked completely fine, but ai rewrote them anyway.
You forgot part 5 : AI scrapes MSlop forums for reported issues with the new update, then posts someone’s workaround as the official fix while the bots in the back hallucinate another fix for next week
… i assume this is you hallucinating a pipeline?
EDIT: what i meant was: i assume this is your guess?
Yes, this is my guess, but it is indeed accurate.
OP was making a joke. It was sarcasm
:o
oki <3
Its a mix. Some devs are definitely lazy but from what I have heard there is also a big push for devs to deploy faster and to actively use AI or be punished. So there is incentive to just get code out to meet deadlines/expectations and move on to the next task. The amount of work being put on an individual dev is rising with these accelerated expectations in mind, and getting another dev to review your code takes time from them and their own stack of tasks so code review quality has fallen greatly. Not to mention the high likelihood that AI is also doing code reviews to make up for this and we can only guess how many reviews get approved saying “you’re absolutely right!”