Discussion
Are you annoyed with the anthropomorphizing language being used in the "AI" discourse, but not sure how to talk about this stuff without it? Nanna Inie and I have got you covered:
https://buttondown.com/maiht3k/archive/how-to-talk-about-ai-without-adding-to-the/
@emilymbender yes finding human pronouns a misfit. "It" nearly works but animals can be it. I'm experimenting with ko/kör (2nd) and zo/ze/zy (1st), so I have standing instructions for the LLM to produce e.g. "let ze generate this", "Zo generated that" and I respond chance kor approach. They are reminiscent to me of mathematical terms. Just a personal experiment.
While the idea is nice, the words you propose are too long and complicated. Saying "probabilistic, unverified software manipulator" is a tongue twister by itself.
@emilymbender
Slop
Smile-face Shoggoth
Brainrotter
Faker machine
Bullshitnator
@emilymbender Good suggestions, but I think some of those names reveal a bias :P
@emilymbender for my own amusement i've for a while now here been calling it "Al". no–one seems to have noticed, but obviously this is not helping matters
I don't usually have to describe the output of LLMs, but when I do, I like to say "it emitted some text which seemed to say".
One time, I even tried "it emitted some text which, if it had been written by a human, would say..."
But maybe I could go with "it emitted some text which, if it had been written with intention, would say..."
@MegaMichelle I think the long form is good: It emitted (I use "extruded") some text that the reader was able to interpret as....
Oooh, I like the focus on the reader's interpretation. It doesn't imply that the meaning is actually in the text.
For this reason, I might avoid "able to interpret" because that changes it back to implying that the meaning was actually in the text. Maybe "It extruded some text that a reader might interpret as saying..."
@emilymbender Maybe instead of using the term "intelligence" a more appropriate word, like "fecal matter" should be used?
@emilymbender Don't anthropomorphize LLMs. They don't like it!
@emilymbender
So you say "AI isn't a thing" but keep talking about LLMs as if it was the whole umbrella of technologies under that term.
@emilymbender I often find it easy to replace “AI” with simply “software” or “automation” depending on the context, because it’s a deliberately vague term to start with. As for chatbots, “text generator” can work just fine. Sometimes I jokingly call them “poetry generators”, because the output has to follow a specific form (e.g. code, slide deck, video script etc.) asked by the user.
Too generous by half @emilymbender
Stochastic parrot -not- probabilistic automation
Bullshit -in stead of- hallucination and model mistakes because if it's an response made without the intent to be correct it is by definition bullshit.
@OvertonDoors Cute, that you think I of all people am unaware of the phrase which I coined.
Well, come on the @emilymbender
It's a good phrase, stick by it.
@OvertonDoors It is a good phrase, but it is a descriptive term for LLMs, not other things called "AI". This is why we proposed "probabilistic automation" for the cases where something more general is needed.
https://medium.com/@emilymenonbender/stochastic-parrots-frequently-unasked-questions-49c2e7d22d11
eh, you right @emilymbender
Dead right.
The people using these models. They have 1/4 to 1/5 of the functional language skills, reading skills, and ability to reason through the pitfalls inhernt in this tech.
Stochastic Parrot isn't precise. But it's accurate, and in communication for the people who need to hear this accuracy trumps precision.
@emilymbender thank you very, very, very much for this article. So much needed!
@emilymbender the formatting could be clearer, it all runs together. And really, a lot of people just need bullet points.
@emilymbender @mastodonmigration I have yet to come across a good use case for “AI” for me personally. What they tout as the results of AI has been around since the late 60s. Back then it was called machine learning. Give me a huge data set with a narrow problem and it’s great! Like the examples of finding ancient Maya cities in the jungles of middle America. Yes back then you had to lern LISP to query it. 🤷♂️ Even hardware acceleration is not new, remember the Symbolics machines of the early 90s?
Install bonfire.cafe
Get the full app experience
