@lauraknelson
Associate Professor @ UBC computational sociology machine learning is feminist You only have to look at the Medusa straight on to see her. And she’s not deadly. She’s beautiful and she’s laughing. www.lauraknelson.com
I don't see this current stage as being all that different from previous stages of the "science of culture". The mechanics of using generative models to research culture are largely the same as what folks have been doing for a decade. We just have more options now.
The furthest back I could trace this to led me to this passage 👇 where they claim that things like feeling irritated counts as a decision. So yeah, not just should I look up first, but the decision to recognize feeling hungry at all, and then do something about it, and then
bsky.app/profile/laur...
As we delve more deeply into the topic of decision making, we must keep in mind that nearly everything we do and many things we feel are the result of decisions—even though we often do not think of them as such. For instance, consider this example. You are watching a football match at a pub, and a drunk patron is staggering around and being obnoxious. What do you do? Do you avert your eyes? Do you ask the bartender to eject the patron? Do you shout, ‘Sit down’? These actions—or inactions—are manifestations of decisions. More subtly still, what you feel may well be the result of a decision. Do you feel irritated? Do you feel threatened? Do you start to feel irritated and then recall your own boorish behaviour during university and decide you have no right to judge? It is impossible to catalogue all of the possible decisions in even the simplest situation, and we do not intend to. We merely want to emphasize how ingrained the decision-making process is to everyday experience, and how the cognition of decisions can encompass changes that are not perceptible to anyone else. When we speak of decisions, it is important to remember how much of our experience and identity is connected to these seemingly simple events. Sahakian, Barbara, and Jamie Nicole LaBuzetta. Bad Moves : How Decision Making Goes Wrong, and the Ethics of Smart Drugs, Oxford University Press, Incorporated, 2013. Page 14.
The best I can find in that book is this passage. Clearly articulating that we indeed likely make thousands of decisions every day, but I can't find anywhere where they do the actual calculation. The mystery continues.
Ok, this is helpful!
Yeah me neither. I now just say there's no agreement, and cite this paper: journals.sagepub.com/doi/10.1177/...
That happens to me every time I write a paper. Please tell me you have citation for it!
Ok I think the statistic might be reported in this book. I'm now trying to get access to that book to figure out of it's actually reported in it, and what they cite to back it up.
www.amazon.ca/Bad-Moves-de...
If you factor in sleeping, that about one decision every two seconds.
Ok I'm in a rabbit hole. If you search "how many decisions do we make in a day" the reported number is almost always 35,000, often reported that this is according to "multiple sources". Yet I can't actually find a single source that backs up that number. Anyone know where this number comes from?
I worked at a centre and the director, like you, was inundated with emails. Her assistant Amy had access to her email. Amy sorted her email. Some she needed to answer personally. Some, Amy drafted a response that she just needed to approve. The reciever read it because it conveyed information.
I'm curious what jobs, or forms of writing, you're thinking of here?
Ah! It's nothing to do with prompting. ML, the method, the math, is feminist because it allows for high dimensionality in the way it represents and processes data - people, language, images. Rather than reduce people/text/etc to a category (e.g. surveys), ML uses ✨vectors✨ with all their dimensions
Umm, there were never prizes for washing dishes tho, even when done by hand? A better analogy is photography. Most pictures are artistically bad but serve a purpose. Everyone can take a (bad) picture, no skill at all involved. Just click. There is still a demand, and prizes, for skilled photography.
Absolutely. Writing can be beautiful! It can give you chills, take your breath away. There will always be a place for that.
Oh yeah for sure. But even those who most love the art of language ought to be able to accept that writing means many different things to people, and serves many functions
This is the needed corrective to the new round of "when humans write it's like music, every word carefully chosen and precisely placed" which is simply not the reality for most people and situations. Much of writing is a tedious task that we have to slog through to simply keep our job/day/etc going.
Mimeograph, Linux, Napster
Higher ed programs should teach more about history/philosophy/sociology of science
The history of AI in China vs the US is so interesting. Or even East Asia vs North America. And it goes way back! I figured someone would have published on this by now.
Is there a good piece to assign on this? Preferably with some historical context? I start my class with the Xiaoice/Tey example and would love to follow it up with something contemporary.
Our department is hiring an Assistant Teaching Professor!! This is a joint-appointed position with Computational Social Sciences (css.ucsd.edu). It's 75+ degrees F and sunny today, just thought I'd mention apol-recruit.ucsd.edu/JPF04461
Now out in the American Sociological Review
We present the first large-scale assessment of the structure and evolution of temporalities expressed in U.S. climate change news coverage (2000 to 2021). For this, we analyzed more than 23,000 statements about climate change effects and actions. 🧵 1/
Oh this is a fantastic paper.
I wonder if we can leverage the overly simplified way LLMs simulate opinions to do controlled experiments of some sort. Can we turn this into a positive, counterfactual-type feature of LLM simulations?
Incredible. Fantastic research design.
🧵on my new paper "Synthetic personas distort the structure of human belief systems" w Roberto Cerina I'm v excited about...
🚨 Do synthetic samples look like human samples?
We compare 28 LLMs to the 2024 General Social Survey (GSS) to find out + develop host of diagnostics...
The IC2S2 deadline is right around the corner!
We really need more reviewers to make this conference work. I know (I KNOW) you all get a lot of requests but please consider signing up, especially if you submit. It’s just a few abstracts, we’ll keep the review load light. Promise!
So awesome to see this international bestseller data *out in the world*! That's the goal!
Thanks to @jamesfolta.com, @literaryhub.bsky.social, and F. Poretti for these great pieces.
From time to time I mutter about a secret project that involves benchmarks and historical language models. Here's a formal announcement of the Schmidt Sciences grant. Other PIs include @dmimno.bsky.social , @lauraknelson.bsky.social, @andrewpiper.bsky.social, and @mattwilkens.bsky.social. And +
For a long time education has been turning into a "fill in the blank" activity. Ace the test, write the paper according to my prof's rubric, impress the right people, and you'll get the grade. But of course we know that's not learning. AI is a very stark illustration of why that's not learning.
Oh no (but I get it). I'm optimistic because students are actively and deeply grappling with what it means to learn now. Like ok, I'm here in uni to get credentialed, but I also value myself as a person. It's golden opportunity for us to engage students about the process of learning itself.
