It feels like we’ve forgotten all the lessons from management metrics over the last 100 years. I’m beginning to think it’s not just the hype, it’s the dopamine hit we get when we send a prompt and get a load of impressive blurb or code back. We’ve developed a reflex around it that clouds judgement. I do think the DORA type metrics are still working well, they measure overall outcomes not reductionist indicators. Seem to remember they’ve added a new rework rate to deal with the slop.
> I’m beginning to think it’s not just the hype, it’s the dopamine hit we get when we send a prompt and get a load of impressive blurb or code back.
For me at least it hits different back in the day there was a certain dopamine loop around figuring out how to make thing work, now is a "cheap" hit; it doesn't register the same.
> I do think the DORA type metrics are still working well, they measure overall outcomes not reductionist indicators.
I tend to agree here, although maybe the scale is different.
> Seem to remember they’ve added a new rework rate to deal with the slop.
The hard thing in my mind to quantify right now is to underlaying security issues that might explode at any moment.
It feels like we’ve forgotten all the lessons from management metrics over the last 100 years. I’m beginning to think it’s not just the hype, it’s the dopamine hit we get when we send a prompt and get a load of impressive blurb or code back. We’ve developed a reflex around it that clouds judgement. I do think the DORA type metrics are still working well, they measure overall outcomes not reductionist indicators. Seem to remember they’ve added a new rework rate to deal with the slop.
Thanks for the comment!
> I’m beginning to think it’s not just the hype, it’s the dopamine hit we get when we send a prompt and get a load of impressive blurb or code back.
For me at least it hits different back in the day there was a certain dopamine loop around figuring out how to make thing work, now is a "cheap" hit; it doesn't register the same.
> I do think the DORA type metrics are still working well, they measure overall outcomes not reductionist indicators.
I tend to agree here, although maybe the scale is different.
> Seem to remember they’ve added a new rework rate to deal with the slop.
The hard thing in my mind to quantify right now is to underlaying security issues that might explode at any moment.