mariaa.bsky.social - Profile | ThreadSky | a Reddit-style client for Bluesky

I interviewed for LLM/ML research scientist/engineering positions last Fall. Over 200 applications, 100 interviews, many rejections & some offers later, I decided to write the process down, along with the resources I used. Links to the process & resources in the following tweets

submitted 6 hours ago • 2 comments

I remember a thread on here abt why citing AI doesn't make sense, b/c citations are about - Authority (why should I believe this) - Reference (where can I go to verify this) - Credit (who deserves credit for this) - ...? Does anyone remember this or have links to similar? #GenAI #AcademicSky 🧪

submitted 5 hours ago • 1 comment

"Search ads accounted for more than half of Alphabet’s revenue and nearly three-quarters of the company’s advertising revenue in 2023, according to company filings." Source: wallethub.com/blog/google-... Incentives matter. This is what makes Kagi different: help.kagi.com/kagi/why-kag...

submitted 5 hours ago • 0 comments

. @mariaa.bsky.social at the 'Impressed by Reading: Measuring the Impact' workshop (Lorentz Centre, Leiden University)

submitted 7 hours ago • 0 comments

Canceling our national surveillance of maternal and infant health - even worse than what they said they’d do in Project 2025.

submitted 1 day ago • 0 comments

I am sad to share that all Retirement & Disability Research centers have been ordered to stop work. My fellowship will no longer exist, and the work of so many talented researchers and dedicated staff will be cut short. www.ssa.gov/news/press/r...

submitted 2 days ago • 17 comments

TIL that if you're using the ACM latex template, you can convert it to a nice preprint format using: \documentclass[acmsmall,screen,nonacm,review=false,timestamp=false]{acmart} Thinking of all the time I've spent carefully modifying the template by hand, oops.

submitted 1 day ago • 0 comments

At least 12 universities now have hiring freezes and nearly 30 have cut PhD admissions this year If you hear of other places you can add them to this google doc:

submitted 1 day ago • 18 comments

This is really good and important, and a model we can point to for what it would mean for the social sciences and humanities to guide AI. A crude version of the argument is: existing approaches to “cultural AI” will model stereotypes (knowing-about-groups) not the sense-making power of culture.

submitted 2 days ago • 0 comments

Time to label some data...

submitted 2 days ago • 1 comment

Postdoc opportunity at the School of Information Sciences, Urbana-Champaign. Teaching one course a semester. Application deadline April 1, but they recommend conversation with potential faculty mentors in advance of that date. Renewable for a second year. Pls RT! ischool.illinois.edu/research/pos...

submitted 2 days ago • 5 comments

👋🏽 We're excited to announce the 💻Call for Workshops🛠️ for #CSCW2025! 📆 Deadline: March 27th, AOE. Workshops will be held on October 18-19th in Bergen, Norway. In-person workshops are strongly encouraged but hybrid is also possible. More details below ⤵️ cscw.acm.org/2025/index.p...

submitted 3 days ago • 0 comments

Pitt pauses grad school admissions while waiting for judge to decide if Trump administration's chaotic attempt to abruptly reduce medical research funding that was already promised is legal www.post-gazette.com/business/hea...

submitted 2 days ago • 1 comment

⚡️US threatens to shut off Starlink if Ukraine won't sign minerals deal, sources tell Reuters. The U.S. first raised the prospect of shutting off Starlink after Zelensky rejected the deal presented on Feb. 12. Starlink terminals are critical in securing Ukraine's battlefield communications.

submitted 2 days ago • 501 comments

Martin Tutek, Fateme Hashemi Chaleshtori, Ana Marasovi\'c, Yonatan Belinkov Measuring Faithfulness of Chains of Thought by Unlearning Reasoning Steps https://arxiv.org/abs/2502.14829

submitted 3 days ago • 0 comments

My @cred.blue score is 453! 🦋 I've been on Bluesky for 666 days, joined during the "invite-only" era, and have a social status of "Moderately Engaging Pathfinder" (I think I am ok with this) Get your score: cred.blue (an interesting new way to check the credibility of an account)

submitted 4 days ago • 2 comments

New preprint! Metaphors shape how people understand politics, but measuring them (& their real-world effects) is hard. We develop a new method to measure metaphor & use it to study dehumanizing metaphor in 400K immigration tweets Link: bit.ly/4i3PGm3 #NLP #NLProc #polisky #polcom #compsocialsci 🐦🐦

submitted 4 days ago • 5 comments

UW’s @techpolicylab.bsky.social and I invite applications for a 2-year Postdoctoral Researcher position in "AI Alignment with Ethical Principles" focusing on language technologies, societal impact, and tech policy. Kindly share! apply.interfolio.com/162834 Priority review deadline: 3/28/2025

submitted 5 days ago • 0 comments

Helping you keep up with all the open models and artifacts, one month at a time. * Countless reasoning models + datasets for you to try * Tons of excellent Chinese models * Lots of multimodal models (has been taking off for many months now) Enjoy and let us know what we missed!

submitted 5 days ago • 0 comments

Excited to have two papers at #NAACL2025! The first reveals how human over-reliance can be exacerbated by LLM friendliness. The second presents a novel computational method for concept tracing. Check them out! arxiv.org/pdf/2407.07950 arxiv.org/pdf/2502.05704

submitted 5 days ago • 2 comments

The updated ARR schedule has now been posted! aclrollingreview.org/dates Feb 2025: Februrary 15 May 2025: May 19 July 2025: July 28 Oct 2025: October 6 Jan 2026: January 5

submitted 5 days ago • 1 comment

📡 at very long last, today is the big day! my beloved book, OTHER NETWORKS: A RADICAL TECHNOLOGY SOURCEBOOK, is officially available for pre-order! 1/5 #othernetworks shop.mexicansummer.com/merch/495898...

submitted 6 days ago • 18 comments

This seemed very good! Very actionable! “Our goal is to encourage scientists to return to their hometowns by writing brief opinion pieces in their hometown newspaper, advocating for investment in American science”

submitted 5 days ago • 0 comments

Just got an email saying that my reviewing availability is not registered even though I did register it and it shows up on my console. Dont want to be desk-rejected over this -- who to ask? @aclmeeting.bsky.social

submitted 5 days ago • 0 comments

There's been a lot of work on "culture" in NLP, but not much agreement on what it is. A position paper by me, @dbamman.bsky.social, and @ibleaman.bsky.social on cultural NLP: what we want, what we have, and how sociocultural linguistics can clarify things. Website: naitian.org/culture-not-... 1/n

submitted 6 days ago • 5 comments

🤔 Ever wondered how prevalent some type of web content is during LM pre-training? In our new paper, we propose WebOrganizer which constructs domains based on the topic and format of CommonCrawl web pages 🌐 Key takeaway: domains help us curate better pre-training data! 🧵/N

submitted 6 days ago • 1 comment

wow I do not like those Reddit lines

submitted 6 days ago • 6 comments

Prompting is a weird art that rewards language, logic, and computing experience. Impossible requests get silly answers, but you have to notice that the request is impossible. Somehow I am both: - "you can't just prompt, you need to evaluate" - "oh you think you can't prompt for that? let me try"

submitted 6 days ago • 0 comments

Archive Request xkcd.com/3052

submitted 6 days ago • 124 comments

📢 CSCW 2025 Major Revision Decisions Released! Decisions for the January 2024 cycle (major revisions) submitted in October 2024 have been sent to the contact author. Please check your emails. ✨ The next and only full paper deadline for CSCW 2026 is May 13, 2025, 23:59 AoE.

submitted 6 days ago • 0 comments

Our new piece in Nature Machine Intelligence: LLMs are replacing human participants, but can they simulate diverse respondents? Surveys use representative sampling for a reason, and our work shows how LLM training prevents accurate simulation of different human identities.

submitted 7 days ago • 9 comments

“19. Success requires the study of how others have survived this, but there doesn’t seem to be a way to make this any easier.”

submitted 7 days ago • 1 comment

This looks like a fantastic postdoc opportunity at Cornell, slated to work with a wonderful team. They're looking for someone who does English/cultural studies + DH, with emphasis on computational work.

submitted 7 days ago • 1 comment

A thread on “persuasion” and bias in AI models that’s worth your time

submitted 7 days ago • 0 comments