let’s see the numbers
-
let’s see the numbers
One thing I do is I feed the various suggestions I see to chatgpt to see if they really break it. Also I sometimes argue with the google search thing when it shows up somehow and pisses me off.
-
let’s see the numbers
@zzt ~four months ago as an experiment i used free claude to make me an emacs package that let me post an org subtree to mataroa. when it worked all the slot machine bells went off in my head at once and i stayed up really late trying to think up other stuff i could make it do. free accounts run out of tokens so fast that i was bouncing these half-functional .el files between like six different accounts made with ancient email addresses. i couldnt get any other packages i "made" to work lol
-
@zzt ~four months ago as an experiment i used free claude to make me an emacs package that let me post an org subtree to mataroa. when it worked all the slot machine bells went off in my head at once and i stayed up really late trying to think up other stuff i could make it do. free accounts run out of tokens so fast that i was bouncing these half-functional .el files between like six different accounts made with ancient email addresses. i couldnt get any other packages i "made" to work lol
@zzt a couple months later i found the mataroa cli tool and jury-rigged a shitty elisp function to do exactly the same thing without a problem lol. and yeah that same weekend i had a minor version of the same thing happen when i asked an llm to troubleshoot something. it worked and then i was just flooded with 'just ask claude it'll be fine' intrusive thoughts for the rest of the week. i was actually legit floored by how bonko it made me immediately
-
@zzt a couple months later i found the mataroa cli tool and jury-rigged a shitty elisp function to do exactly the same thing without a problem lol. and yeah that same weekend i had a minor version of the same thing happen when i asked an llm to troubleshoot something. it worked and then i was just flooded with 'just ask claude it'll be fine' intrusive thoughts for the rest of the week. i was actually legit floored by how bonko it made me immediately
@zzt it seems like there are a lot of people who dont tweak out in that particular way but if even like 1% of users experience that without the critical framework i'd already kind of built up around it...yikes
-
let’s see the numbers
@zzt@mas.to The first one, but I may have to switch to the second in the close future…
-
let’s see the numbers
@zzt Where is the « I don't care LLMs » line ?
-
let’s see the numbers
@zzt When I told my students last semester I try to avoid AI whenever I can they were legitimately shocked. "How can you program or find images or learn anything without AI???" It was a huge culture shock, but I also suspect partially a symptom of me finishing school years before genAI.
This year it seems the new freshmen are much less reliant on it (or at least more of them are also openly anti-AI), so that's reassuring.
-
@zzt When I told my students last semester I try to avoid AI whenever I can they were legitimately shocked. "How can you program or find images or learn anything without AI???" It was a huge culture shock, but I also suspect partially a symptom of me finishing school years before genAI.
This year it seems the new freshmen are much less reliant on it (or at least more of them are also openly anti-AI), so that's reassuring.
@zzt I've considered throwing some of my assignments at them just to see how they respond since it might make catching cheating easier, but so far I haven't really had to. Some of my assignments don't really work with genAI, anyway, since I have students sketch something on paper or explain the logic for how they arrived at their result.
-
let’s see the numbers
@zzt No options apply.
I don't hate the core tech. I think it's interesting. I do hate most of the companies dominating it though, and the way they fuck up so many things in society right now.
My work expects me to use AI, and I do for some tasks, but firing someone for not using it would probably be wildly illegal. We are most definitely being pushed to use it though.
I use AI-tools very sparingly outside of work. Sometimes I run them locally. Most of local use is for video enhancement.
-
let’s see the numbers
@zzt voted 1 but you really need a "hates 99.9% of llms, wont use them, but their one use case of... on keqw 5.2 locally uses less power tham halo" option
to get stats on the probably technically correct, doesnt lie, and refuses to not answer questions even if it makes them look bad crowd -
let’s see the numbers
I hate LLMs so much that I called my old boss before yelling at the VP in front of my people and then I jumped ship.
-
let’s see the numbers
@zzt hate and don't use, but I love 1980s connectionism and I do like playing (emphasis: play not work) with neural nets
-
let’s see the numbers
-
B bogwitch@social.data.coop shared this topic
-
let’s see the numbers
@zzt There is no `I hate LLMs, but I still use them`
-
let’s see the numbers
@zzt I need the option: I have no clue what LLM is but is oddly curious about the voting result anyway
