Lately, an odd habit seems to have caught on among my peers when they’ve done something with an LLM.

When asked “how many support requests have we received about the new Feature X?”, they answer “according to [LLM tool], we’ve received 17 requests since it was released.” When posting a pitch document or a pull request, they volunteer “I asked [LLM tool] to write this for me” or “[LLM tool] wrote the unit tests”.

Odd. Like a writer disclosing the spellchecker they used alongside every piece, or a pilot announcing their version of avionics software over the PA while taxiing to the runway.

Why did I need to know that?

At first, I put it down to a bit of novelty and excitement. But as this wore on, I started to wonder why someone might feel the need to declare their use of an LLM.

Perhaps:

1. You believe in ascribing credit where credit is due. Using an LLM feels like cheating so disclosing it makes you feel less like an imposter for having produced something amazing.
2. The LLM’s performance exceeded your wildest expectations, ergo the work is impressive by association. Never mind that it is only 80 percent complete and no one knows which 20 percent is missing.
3. You outsourced the work to an LLM and didn’t bother to review it before hitting send. There may be mistakes but spotting them is left as an exercise for your recipients.
4. LLM use has been mandated by Up Above. Claiming you’ve used one (even if you haven’t) buys you a higher tolerance for sloppiness since you are presumably facilitating your own eventual redundancy.

For 2 to 4, all I can say is I’m sorry and I hope things work out for you.

But for the first reason, I’d challenge the notion of sharing credit with a tool, let alone feeling like a cheat for using one. Credit and accountability are two sides of the same coin. A tool can only take as much credit as it can be held accountable for, so crediting it inadvertently dilutes your own accountability for the work.

If you’ve created something profound or remarkable, take complete credit for it. If it misses the mark or turns out to be utter slop, also take full responsibility and improve from there. This is how every meaningful or valuable thing was ever created long before the advent of “AI can make mistakes” and will remain so long after, with or without LLMs.

HN discussion