LLMs and performative productivity

Josh Collinsworth asks whether LLMs really are making us more productive (whatever “productive” means).

Think about it: if LLM code was reliably of higher quality than human code, then LLMs would be making software more accessible, more maintainable, more performant, more usable, and more reliable […]

we could expect [all domain experts] to be elated, if LLMs were actually moving the metrics they care about.

But that’s pretty much the opposite of what’s happening, as far as I can tell.

Instead, everywhere I look, specialized craftspeople are overwhelmingly burned out from fighting a losing fight to get people to care

The journey is the work:

A junior who made a mistake is one step closer to being a senior; a junior who let an LLM make a mistake (and had the LLM fix it for them) has probably learned nothing.

Lots in here that resonates with current, day-to-day experience:

If you’re opening PRs faster than anyone can read through them, you’re not increasing productivity; you’re clogging the bottleneck.

I [used the LLM] mainly [to] just check off a bunch of old to-dos, most of which were unfinished because they never mattered that much in the first place.

The more you take a big-picture, holistic view […] the more gains from LLM usage shrink

Code is not the product:

Building things got cheap, but building the right thing didn’t get any easier.

Even perfect code can still make bad products.

Oof.

The less you understand, the more you trust AI. But the more you trust AI, the less you understand.