if quantitative social science is automatable, it's because many of the fields in question have already reorganized their research in such a way to be indistinguishable from machine output, and already optimized for machine measurement.
This is a great thread, and makes me think about how the affordances of LLMs (largely text-only chat) preclude the kind of epistemic vigilance we were able to develop around tech like search. I think we can build LLM-based tech that's good here, but chat ain't it
Rename LASSO and Ridge to kiki and bouba
Let's not forget about apparently Italy, Greece, and *checks notes* Afghanistan having one school together?
Congratulations!! 🥳🥳🎉
*a situation
lol
This does feel like situation where open models are held back by the state of benchmarks. For an open model, there isn't as much of a theory of use outside of benchmarks. For closed models, the companies putting them out can build for business cases and their associated internal benchmarks
Imo, this is also a major problem at ACL venues, and I think it also creates a formulaic discussion of ethics that is sometimes only tangentially related to the paper :/ I'd love the removal of sections like this (and checklists!), but to emphasize the discussion of ethics as a review priority.
Most tech workers, even at Palantir (probably), are pro-social compared to the tech owner and VC classes. They need unions so their voices matter for company governance.
www.wired.com/story/palant...
Vivid mental image of cellino and barnes being like penn and teller, where one is always completely silent and if you advance a surprising legal theory your consultation is free
Firewatch is one of the most lovely games I've played. I've taken so much joy from playing it with multiple friends and seeing how different a game it is for them. I didn't know games could tell stories like this before it
(It also played a small part in my fiancée and me getting together!)
I also had exactly this experience looking for a paper I could swear I saw about the difference between AI usage and topline productivity metrics, it seems to have just vanished
I remember this, I think? I also cannot find it, though
But I think it *would* be helpful overall to have more people willing to consider what society the day job they leave behind to rock climb is building
Right now, it only takes a few karps to steer the direction of the whole industry, because a lot of people have not been socialized to be critical
As someone that *did* liberalize after studying some humanities, I've done a full 360 on this. I went from saying it, to agreeing with the Alex Karp example, to kinda being back?
Like, I don't think studying the humanities would push people to the left or fix all of tech's issues
Yeah, I think what I'm getting at though is that there is a *lot* of stereotype sharing between diasporic authors and non- or maybe questionably diasporic authors like Anita Desai, and I'm wondering if the audience is sometimes a more important factor than the author
This is one of the things that I found arresting about the Immortal King Rao (which is diasporic, tbf) — there's like, a formal consistency with prestige Indian lit in the cultural things presented as authentic. Everyone's breath smells sour, the way the village/coconut farm is presented, etc.
They do, but I think this is complex than emic/etic representation!! Like, even among non-diaspora Indian authors in English, things like the scent of jasmine are _really_ common, and I think it's because of the writing for an outside audience
This is a real banger of a paper. The example of a model being weirdly focused on jasmine (lol) makes me increasingly think that single-point-of-access models don't really consider who their audience is. Jasmine is a super legible cultural marker for people outside, but is so, _so_ generic.
Also, for other beautiful and stressful, I've heard really great things about both The Last of Us games (too stressful for me, though). For beautiful and (maybe?) less stressful, I can't recommend Chants of Sennaar and Return of the Obra Dinn enough - just lovely, lovely puzzle games
This is definitely one to avoid spoilers on!! The narrative is really beautiful, and really well done in game. I also definitely second the recommendation for the sequel, both so fun and narratively great
“Position paper” is just a label to make some kinds of interdisciplinary theoretical work fit into the CS publishing schema.
ty!! This looks so good, and I'm excited to do some of this reading!
extremely on brand, love this town
That'd be great, I'd love to take a look!
I would take this class in a second if I could 👀
Here if you'd like a pair of eyes!
Fwiw, I really enjoy your point of view on here, so would love to hear about what you see as worth doing.
If you're willing, I'd love to hear more about what the thing you'd like to be doing is, as someone who's in a similar position. I think I know what I'd like to be doing, but I have no idea whether it'd be economically viable for me.
Best use of this meme i’ve ever seen
