siree.sh
PhD student @ltiatcmu.bsky.social. Working on NLP that centers worker agency. Otherwise: coffee, fly fishing, and keeping peach pits around, for...some reason
https://siree.sh
120 posts
2,372 followers
2,745 following
Regular Contributor
Active Commenter
comment in response to
post
All of these sound great! I would also maybe suggest rather than (or maybe in addition to) a starter pack, a list of attendees - that way, the list can back a custom feed, capturing things that aren't explicitly tagged for the conference, and the feed also becomes interest-based after the conf
comment in response to
post
This is some excellent viz 🤩 also love that there's a Sam Learner on the team that built it!
comment in response to
post
OpenAI has effectively conned people into thinking that Chatbots & AI "Assistants" are The FEWTCHA of AI. Friends, they are most likely *not.* Neither are the big cloud-based Generative AI services.
Small, purpose-fit, on-device models that make your existing activities easier/better? There you go.
comment in response to
post
The MA lottery is the follow-up season - it's the same folks and the same feed!
comment in response to
post
I absolutely love the work coming out of WGBH, about the big dig and the MA lottery!! www.wgbh.org/podcasts/scr...
Both 10ish parters, and lovely to listen to in series.
comment in response to
post
I'm excited to read (and download/cite, ofc) this when it's out!!
comment in response to
post
👀
comment in response to
post
🤩🤩
comment in response to
post
I don't know if conditions being bad in PA is a weather thing, but if you haven't, consider cherry springs state park! www.pa.gov/agencies/dcn...
comment in response to
post
I also emailed PCs, I guess we'll see 🤷♂️
comment in response to
post
Yeah, I think I read this (and got burned) the same way :/
D&B also had additional spots that implied our original reading, like the dates page:
comment in response to
post
Did you figure this out I also couldn't add authors that didn't have an OR profile, and now adding authors has been disabled on the portal.
comment in response to
post
I love this frame!! Excited to read this paper post NeurIPS deadline!
comment in response to
post
These look great, thank you!!
comment in response to
post
I actually just stumbled on this (and Briet) independently, in a field where data (but only quantitative!) is often unquestioned in exactly the way you describe.
Do y'all have any recommendations for more stuff in this vein?
comment in response to
post
I like "bullshit" as a descriptor, and another one that is maybe easier to use (and is also more correct) is "confabulation"
comment in response to
post
Idk, man, I like the g at the end.
Then again, maybe this would get North Versailles to reconsider how they pronounce their name 🤔
comment in response to
post
Hey, if it's good enough for the guy that founded the town...
comment in response to
post
I don't doubt that there will (eventually?) be some applications of AI that present a more compelling vision, one that doesn't involve e.g. crazy environmental costs, in the same way that short-form video is now a thing — just not in the way Facebook pitched it then.
comment in response to
post
This is great! And as someone at NAACL now, there's still a few days left to take the advice!
comment in response to
post
We also had a lovely discussion afterwards, that got at a number of issues that @davidthewid.bsky.social and I (along with @abertsch.bsky.social, @clarana.bsky.social and @strubell.bsky.social) addressed in our previous work: aclanthology.org/2023.emnlp-m...
comment in response to
post
Buddies! bsky.app/profile/did:...
comment in response to
post
I am also at NAACL this week, presenting a materials science-based multimodal IE benchmark we're developing at the AISD workshop. Hit me up if you'd like to talk AI for science (or scientists), or user-centric, local NLP.
comment in response to
post
Can't wait for you to read my paper "Scaling laws for memey paper titles: mad libs is all you need" 🚀
comment in response to
post
@mariaa.bsky.social would probably have some great pointers!
comment in response to
post
Fwiw, this seems true of AI-adjacent fields in general. When we interviewed people about NLP, one of our participants literally said they found it difficult to publish papers with ideas in them in *CL conferences anymore
comment in response to
post
Well, I definitely *start* with a conference in mind 🥲
comment in response to
post
Also, apologies if you've seen this already, but I love the kind of work in e.g. arxiv.org/abs/2309.15084 — I think there's real value in making this explicit/readable/citeable!
comment in response to
post
Especially bc so many tech companies are pivoting to jingoistic "national security"/"defense" applications that a) is a real historical return to form and b) a cynical pivot towards the only relatively easy source of hyper scale money left
comment in response to
post
Fwiw, this is not actually my primary area of research! Corporate involvement was def the elephant in the room when we were looking at the grands dataset, and I think it would be *super* useful!
comment in response to
post
If you haven't seen it, the Mike Schur series Man on the Inside is really lovely and affirming of being human
comment in response to
post
Heavy same!!
Even as someone for whom writing and thinking don't feel as closely aligned, it still feels like there needs to be more work on the role of friction, whether writing or otherwise, in thought
comment in response to
post
This is so cool! I love libraries as the bridge to evaluation and training for really specific things ❤️
comment in response to
post
Oops, @xblock.aendra.dev
comment in response to
post
If you don't use it already, highly recommend xblock.aendra.dev! You can hide post screenshots from other social media services
comment in response to
post
I'd really love to see these slides, if you're willing to share!
Your point about water usage, also, 💯
comment in response to
post
If we did just refer to them as (zero-shot) classifiers, it'd be a lot more obvious that you *do* need validation, and papers like this wouldn't need to exist: arxiv.org/abs/2406.184...
I think you're right that the terminology just makes it sound as though it's a new, totally legit thing.
comment in response to
post
So much AI *deserves* to be sneered at!! I wish reviewers would understand that criticism is still engagement, and is better for everyone than hypebeasting
comment in response to
post
Yeah, genuinely awful, and probably a dealbreaker in the age of the app store? But as you said, everyone shipping their own model would also be bad