The acceleration of agentic coding and the plummet in software quality & reliability aren’t two separate stories. They’re the same story. Even if a particular bug/problem can’t be traced back to a slop source, the proliferation of slop is itself a net drag on software teams and leads to overall deskilling and focus diversion.
We have a duty as industry professionals to fight back, say no, and make sure companies are aware this “new normal” is completely and utterly unacceptable.
As I alluded to, when that’s the case it’s still now agentic coding to blame because of the diversion of focus. Companies such as Microsoft, instead of improving their existing human-based craft, introduce literal roadblocks to the improvement of craft both in terms of time/resources and also in terms of corporate culture.
Maybe. But even if everyone banned AI instantly, the culture and (I’d argue) massive mismanagement problems aren’t getting any better.
Hence I think it’s iffy to attribute all these software problems to ‘AI’ so quickly, especially before it’s really had time to affect old systems like Microsoft Windows.
This is now multiple times over multiple post threads where you’ve replied to literally whatever I have said that is critical of LLMs and directly contradicted it. I don’t wish to continue such fruitless conversations.
Sorry. Coming off as a tech bro was not my intent, and let me be clear: these LLMs can, and do, do some shit to code in alien, difficult to trace ways. It’s an existential issue.
…I just don’t want it to mask a problem that’s already there, either. Microsoft’s dysfunction goes way beyond their code completion integrations.
OK that’s fair. And also it’s not my intent to blame AI for all problems of software culture. Just seems in many ways we’re going backwards as an industry.
AI is a perfectly distilled example: pushing an underbaked, sometimes neat tool in completely dysfunctional time-wasting ways, from the top down, because leadership is infatuated with the dream of it, talking to other infuatuated leaders, and feeling FOMO… I guess that’s why I came off so cynical when I read this:
We have a duty as industry professionals to fight back, say no, and make sure companies are aware this “new normal” is completely and utterly unacceptable.
…But this keeps happenening.
AI wouldn’t be the infectious mess it is without underlying structural issues, and pushing back on ‘AI’ directly doesn’t really convey “you shouldn’t have shoved this down our throats in the first place, again!” I dunno how how fix that though.
EDIT: I’m definitely venting at this point, apologies.
The acceleration of agentic coding and the plummet in software quality & reliability aren’t two separate stories. They’re the same story. Even if a particular bug/problem can’t be traced back to a slop source, the proliferation of slop is itself a net drag on software teams and leads to overall deskilling and focus diversion.
We have a duty as industry professionals to fight back, say no, and make sure companies are aware this “new normal” is completely and utterly unacceptable.
But these problems took off way before code assistants were common.
As I alluded to, when that’s the case it’s still now agentic coding to blame because of the diversion of focus. Companies such as Microsoft, instead of improving their existing human-based craft, introduce literal roadblocks to the improvement of craft both in terms of time/resources and also in terms of corporate culture.
Maybe. But even if everyone banned AI instantly, the culture and (I’d argue) massive mismanagement problems aren’t getting any better.
Hence I think it’s iffy to attribute all these software problems to ‘AI’ so quickly, especially before it’s really had time to affect old systems like Microsoft Windows.
This is now multiple times over multiple post threads where you’ve replied to literally whatever I have said that is critical of LLMs and directly contradicted it. I don’t wish to continue such fruitless conversations.
Sorry. Coming off as a tech bro was not my intent, and let me be clear: these LLMs can, and do, do some shit to code in alien, difficult to trace ways. It’s an existential issue.
…I just don’t want it to mask a problem that’s already there, either. Microsoft’s dysfunction goes way beyond their code completion integrations.
OK that’s fair. And also it’s not my intent to blame AI for all problems of software culture. Just seems in many ways we’re going backwards as an industry.
I’d argue its a leadership culture problem.
AI is a perfectly distilled example: pushing an underbaked, sometimes neat tool in completely dysfunctional time-wasting ways, from the top down, because leadership is infatuated with the dream of it, talking to other infuatuated leaders, and feeling FOMO… I guess that’s why I came off so cynical when I read this:
…But this keeps happenening.
AI wouldn’t be the infectious mess it is without underlying structural issues, and pushing back on ‘AI’ directly doesn’t really convey “you shouldn’t have shoved this down our throats in the first place, again!” I dunno how how fix that though.
EDIT: I’m definitely venting at this point, apologies.