[arrowsmith]

> Were there similar gnashing of the teeth and wails of despair when compilers were first introduced?

Yes, at least according to ChatGPT:

"Compilers didn’t arrive to universal applause; they arrived into a world where a chunk of programmers absolutely believed the machine could not be trusted to write “real” code—until the productivity wins (and eventually the performance) became undeniable."

Damn that sounds familiar.

[falloutx]

compiler is deterministic, coding models are not. compilers have been unit tested and will generate same output for a given input. They are not the same things.

[arrowsmith]

So?

[Anamon]

So, there's a big qualitative difference in whether you can trust the output.

You can either "just believe", and prepare for inevitable, nasty surprises down the line. Or you can verify in ways you don't have to with stable compilers, eating up most of, if not more than all the efficiency gains you felt you had by using the LLM.

The two aren't comparable even remotely. One is a tool, the other is a slot machine. One allows for a new layer of abstraction, the other allows for a new layer of imprecision and hoping for the best.