do we have companies attributing worker productivity gains to gen AI? 

my anecdotal experience of trying to use LLMs for writing is... zero productivity gains, no experience of writing (or thinking) better, faster. can generate mediocre writing & conventional ideas endlessly, but turning that mid text into anything I'd want to publish feels harder than just being thinking & writing about something myself from the jump

I had a similar experience when I had some ghostwriting clients. Tried writing with AI and I was just like fuck this itll be faster and better if i just do it myself

For coding there are productivity gains though, 100%

where do you think we're seeing an impact of coding productivity gains from gen AI?

exqctly this. for simple, repetitive stufd, its a huge value add. think tasks like, update all the log statements in Class A, Class B, and Class F, at info level to include the request_id property

for nonstandard things it hallucinates very frequently, and oftentimes (in my experience) adds time rather than reduces time. it will make up apis, try and use libraries that dont exist, do implementations incorrectly in subtle ways, etc.

yeah I definitely hear lots of devs talking about using AI copilots

feels like its probably useful for simple stuff?

Just speaking anecdotally and from conversations with peers, idk if software companies themselves are more productive or efficient/how to measure that

It dramatically speeds up boilerplate code, doc strings, simple unittests, and a lot of the easy stuff. You can save your coworkers time by asking AI pain in the butt questions.

But it will slow you down if you don't know how to identify its bad suggestions and you roll with them.