so far llms have increased my dev productivity by < 1%

id like to increase this but don’t have a good idea of what a realistic goal is

looking for good ideas to up my game, read on for more context on how i’m using them currently

runloop.xyz • memo.ai (acquired by Coinbase), summify.com (acquired by Twitter) • 🇺🇸🇷🇴🇨🇦🇨🇭

Building @warpcast and @farcaster, new dad, like making food

beyond this i’m unsure of what else it can do for me right now

the most time consuming part of the coding work that feels rote is translating from figma files to frontend components

but even this isn’t that time consuming and i have trouble imaging how the llm makes this go faster

would appreciate any suggestions

some more background:
20 years coding
write frontend and backend TS
working on Warpcast and Farcaster

the way i use llms right now is asking questions to claude in a browser (I have @c setup as a  shortcut to search claude in chrome)

its biggest wins for me have been writing short scripts or answering arcane techincal questions about common things like web standards (ex: css stacking contexts or 3rd party cookie behavior under different CSPs)

Software engineer, music maker, creator.

Possibly a dog irl

great advice, big fan of rubber ducking but didn’t think to do with an llm

Use it as a Rubber Ducky that is just smart enough to get in trouble, but also dumb enough to give you realllly bad information. 

Things I use it for:
- naming variables
- boilerplate of docs (verify)
- writing start of tests
- learning some basics for new languages

i have not but this does seem like the biggest potential for unlock

problem is i’d suffer a 5% productivity loss the first few weeks as i’ve invested a decade in my nvim setup

there are plugins though that can achieve something similar with an llm chat aside the editor but requires a bit more setup, though would probably be worth it

mañana unlimited 🧧 icbk.me/zachharris.eth

@deodad you ever try replit or cursor? I’ve heard things about each…