most of my hatred of ai comes from like, things that are all somewhat unrelated to ai
and probably a few others that i can’t remember
which is most of what i see of AI so im kinda. biased against it
Meow nya purr
Why did may make me remmeber I used to use this phrase now I’ll do it constantly
he threw for 6 touchdowns if you count the two that were for the other team
Fair enough. I just like spend a lot of time talking to people who use chat gpt in places they shouldn’t and presenting them with bad arguments just makes them dig in their heels harder so I tend to get frustrated by it
very hard to get over this bias because its very strong especially when you factor in the “this can probably escalate to the point where it just permanently ruins a lot of my creative interests” (which might not even be true but it’s definitely a perception i have)
How is this my fault I just asked what I should say
just went on the tumblr popular tab
this is a loaded enough statement to make its own post
are there uses? sure. are they basically anything that they’re teaching? haha, nope! i think the only ai thing I was like “oh i actually see the use” was the google ai thing where u can copy & paste ur notes and it’ll synthesize, make flashcards, and make a “podcast” for you to listen to of 2 ai voices basically just talking about ur voice but in a like informal manner. and like yes there is “use” in other things but these are all things that are just Done Better By Humans
I don’t recommend it
hm
it like shows a little peak of the first post and i saw something that looked really odd so i scrolled down to see what it was
it was the thing i thought it was and im sad
I don’t find the machines useful in any learning contexts. I find the normal existing schooling process perfectly adequate, I just do the homework I’m required to do and sometimes glance at a back test and I’m always fine
I opened Tumblr trending to see what scared Atlas but I saw nothing scary. Those posts all seem lame to me. Good way to make me fall asleep from boredom
its different on each refresh and for each person, i think?
yeah that’s fair. i have spent frankly an inordinate amount of time reading about it, reading scholarly papers about it, i took a class on machine learning and neural networks. the algorithms modern LLMs are using were made in like the 70s and virtually haven’t changed. i have a pretty good idea of how chatgpt works excluding the part of it related to conversational context. however, i still think that summation is still an area ripe for hallucination given its output generation is stochastic; it picks the next word from a pool of likely next words. so it irks me to really say that it’s doing any kind of synthesis at all; it’s not. it is a seeming of synthesis, a semblance of coherence. it fundamentally does not do what humans do when we synthesize information. i also think that divorcing someone even further from actual sources is harmful. i think its authoritative tone and reinforcement of biases is harmful. and if a lot of academic literature was wrong about something for years and years and years, and we only recently corrected it, there could be a huge corpus of just wrong text that it has a probability to pull from. it could easily be more likely to give the wrong answer in that case. etc. etc. etc.
I found a reference to one of things I was thinking about. Thank fuck
In a series of blind assessments, the generative AI summaries of real government documents scored a dire 47 percent on aggregate based on the trial’s rubric, and were decisively outdone by the human-made summaries, which scored 81 percent.
