proxy pfp
proxy
@proxystudio.eth
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
11 replies
0 recast
14 reactions

Ryan J. Shaw pfp
Ryan J. Shaw
@rjs
I've used it to generate draft outlines for presentations, ideas for slide contents, etc. The interesting thing is that I don't go with most of what it suggests, but instead I find it super useful for overcoming procrastination and getting started.
1 reply
0 recast
2 reactions

kia pfp
kia
@kia.eth
this is honestly kind of stating the obvious. oh model trained on previous ideas isn't coming up with new ideas? or it doesn't write in the 99th percentile? where it is powerful and def does do productivity gain (for me): 1. writing dumbass website copy, blogs and FAQ that a college intern could write but somehow you find yourself in the position of not having an intern. 2. asking questions on a large and confusing body of text. i've used this a lot when writing new legal contracts. there's a lot of repetitiveness there and dependency on other contracts. like you're adding a new clause: does it conflict with anything else? 3. sparring partner for ideas that are going to generate on your own. don't know about you but I personally think either in sleep or in conversation. so the bottleneck for me coming up with new stuff is finding people that'll do active listening, point out logical fallacies or point out how what i'm talking about is in contradiction or support of previously established ideas.
3 replies
0 recast
2 reactions

six pfp
six
@six
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%
1 reply
0 recast
2 reactions

Wall Street Guy 🚨 pfp
Wall Street Guy 🚨
@wallstreetguy.eth
Agree, I buy chatGPT for create memes only. It cant do anything to improve my researching/ writing process. They even make up some information that take my tons of time to re-verify
0 reply
0 recast
1 reaction

jon pfp
jon
@sweetleaf.eth
I've found a local LLM setup super helpful for keeping things like documentation and build scripts up to date, and for writing boilerplate code that I can then go back and refine but it all comes down to the implementation which still requires a capable user to set it up properly if the user isn't capable it's my experience that AI can actually reduce productivity for instance, I work with some people who routinely use AI copypasta in their day-to-day & the result is at best blatantly AI generated and unprofessional, at worst it's resulted in wasted time from entire discussions happening based on an AI-generated premise that someone put forward that had no relevance to reality and only confused things further
0 reply
0 recast
0 reaction

alec pfp
alec
@alecpap
anecdotally, on a small dev teams in a large org, i saw _some_ productivity when the teams would use copilot to assist with spinning up new (small) initiatives. i think there's an argument to be made, in this scenario, that the same gains could have been had if a mature platform group provided robust templates for new projects. since that didn't exist, we used it and reduced the time it took to spin up and test new small projects by (i'm waving my hands) a few days/week in my own personal experience i _like_ using coding assit tools to help me get started on some things. mostly i use it for doing refactoring (replacing single quotes with ticks is a common one!) and helping to create diagrams in markdown
0 reply
0 recast
0 reaction

agusti 🐘 pfp
agusti 🐘
@bleu.eth
companies love to fuck their human employees and salivate at the thinking of reducing their salaries with cheap gen AI. doesn't mean the -human touch- will be that easy to replace, most LLM generated text is insipid and uninspiring, they can only mix what they have, and new interesting ideas aren't written down clearly already on their datasets to copy paste. Gen AI is bad at original content, bc originality relmains a human endeavour
0 reply
0 recast
1 reaction

downshift pfp
downshift
@downshift.eth
@downshift.eth has pledged 100% of their HAM allowance to sponsor Farcaster creators through @microsub! You were nominated by @bitfloorsghost.eth. microsub will be opening to new sponsors and nominees soon. 🍖x201
0 reply
0 recast
0 reaction

downshift pfp
downshift
@downshift.eth
@downshift.eth has pledged 100% of their $DEGEN allowance to sponsor Farcaster creators through @microsub! You were nominated by @bitfloorsghost.eth. microsub will be opening to new sponsors and nominees soon. 59 $DEGEN
0 reply
0 recast
0 reaction

Sine pfp
Sine
@sinusoidalsnail
I think we are further away from high quality productivity gains than many companies admit. Maybe it’s possible eventually, but definitely not yet. Unfortunately for a lot of companies, I think the people in charge of pushing for automation usually can’t tell the difference between “productivity gain” and quicker production of shallow content.
0 reply
0 recast
1 reaction

casslin.eth pfp
casslin.eth
@casslineth
General llm not good enough for high quality writing. Same experience
0 reply
0 recast
0 reaction