After experience on both ends of the spectrum, I arrived at the position that what we basically need is to be a "Responsible Human in the Loop (RHITL)" [0]

> You may not write the code by hand but you understand it enough to investigate and fix it when it fails. It is how I think we should leverage AI instead of becoming a meat proxy.

[0]: https://raahelbaig.com/entry/responsible-human-in-the-loop/

People keep saying this, but how realistic is that in reality? Even say a person is doing their best to do that, we all know that the alarm that doesn't trip 99% of the time isn't going to be closely monitored. Just look at what happens when people get used to "hands free" cruise control.

Not only that, but even if we assume best effort on the engineer, the business pressures don't often allow that. I'm under constant pressure to deliver more, faster with less people. The performance eval ladder at my company was just reworked to double the amount of deliverable features expected per job level per year. Our CEO told us we should be able to deliver what took us the past decade to deliver in a quarter, every quarter going forward.

How could you possibly have a human anywhere near that loop with those demands?

Don't touch the code for 10 years and even the brightest will experience serious dip in skill. Most probably much earlier, our minds are lazy by default since creative thinking is energetically draining. Then try to debug some harder problem llm created in its own code. Look at what its doing to students.

Its like saying you can keep fluent level of French as a foreign language to you by just reading it. Nope you won't - do it for 2 years solidly and you will have problems forming basic sentences correctly if you don't use language actively, forming foreign language constructs in your mind.

Those few of us who are largely apart from current llm craze (due to my mega corporation and its obscure rules and procedures) are not missing much I see. I still happily code changes by hand and at most consult basic available llms via prompt, its beautiful process to create fix to a problem which seems cryptic at the beginning, and knowing it all came just from my mind. By far the best part of my corporate job, will keep hanging to it for as long as possible.