Yes. I agree those are all terrible. And I agree with your assessment of almost all of them. But I’m not naive enough to think that we can put the evils back in and switch back to a non-ai world. We live, we grow, we adapt. This is the future, no matter how bleak it is.
I didn’t say it improves quality of life. I said it can lower barriers of entry to some things, and it can improve a workers productivity when you compare two workers of equal skill.
improve a workers productivity when you compare two workers of equal skill.
It makes you seem more productive in most metrics that corporate america seems to measure, but that doesn’t really mean anything. Even if most of what the current AIs produce would be good quality (and it’s not, its mediocre at best), one hallucination costs you more time than everything it saved you.
Hard disagree. Random utility scripts, most, terraform, and ansible, brainstorming architecture decisions, throwing an error log at it and spitting back the plausible solution. All faster. But the engineer needs to be capable of knowing when it’s plausible or not.
I don’t think replacing sr engs with interns+claude is the solution. But many sr+principal engineers can be accelerated on a majority of tasks. That’s why was careful to say equally skilled engineer with/without.
Yes. I agree those are all terrible. And I agree with your assessment of almost all of them. But I’m not naive enough to think that we can put the evils back in and switch back to a non-ai world. We live, we grow, we adapt. This is the future, no matter how bleak it is.
I didn’t say it improves quality of life. I said it can lower barriers of entry to some things, and it can improve a workers productivity when you compare two workers of equal skill.
It makes you seem more productive in most metrics that corporate america seems to measure, but that doesn’t really mean anything. Even if most of what the current AIs produce would be good quality (and it’s not, its mediocre at best), one hallucination costs you more time than everything it saved you.
Hard disagree. Random utility scripts, most, terraform, and ansible, brainstorming architecture decisions, throwing an error log at it and spitting back the plausible solution. All faster. But the engineer needs to be capable of knowing when it’s plausible or not.
I don’t think replacing sr engs with interns+claude is the solution. But many sr+principal engineers can be accelerated on a majority of tasks. That’s why was careful to say equally skilled engineer with/without.