Dan Liebling πŸ³οΈβ€πŸŒˆ

Dan Liebling πŸ³οΈβ€πŸŒˆ

@liebling.org

AI for science, HCI+AI research, #nlproc leading w/empathy @ Google Research https://liebling.org Caltech, University of Washington, Microsoft Research alum

210 followers193 following477 posts

Dan Liebling πŸ³οΈβ€πŸŒˆ
@liebling.org

in fact going back as early as 2015, machine translation enabled x-team collaboration despite its frequent mistakes! link.springer.com/article/10.1...

link.springer.comAssessing the impact of real-time machine translation on multilingual meetings in global software projects - Empirical Software EngineeringCommunication in global software development is hindered by language differences in countries with a lack of English speaking professionals. Machine translation is a technology that uses software to t...
hikikomorphism@hikikomorphism.bsky.social

hooooo boy we don't like this no we don't > If you don't speak English, you need to find a different community. We cannot distinguish between AI-written and AI-translated. Since you don't speak English, you can't even verify yourself if it translated correctly. www.reddit.com/r/opensource...

Dan Liebling πŸ³οΈβ€πŸŒˆ
@liebling.org

y/y top public safety issue: enforce traffic laws response: not gonna hire parking officers Seattle even removed parking violations from its issue reporting app

pwalchtaylor.bsky.social@pwalchtaylor.bsky.social

According to the Seattle Times Seattle can’t enforce parking rules because the police want a slush fund/thing to point to when asked how they are saving money? So why did we move parking enforcement from SDOT back to SPD? Maybe we should rethink that. www.seattletimes.com/seattle-news...

seattletimes.comWhy isn’t Seattle hiring parking enforcement officers if they make money?As one of the few revenue-generating city jobs, they should be a budget writer's best friend. Not this year.
Dan Liebling πŸ³οΈβ€πŸŒˆ
@liebling.org

This is exactly what we found in our field work -- many people in the world depend on Google Translate simply to live their daily lives, access healthcare, not get scammed, and so forth.

SE Gyges@segyges.bsky.social

being anti AI is a luxury belief only available to first world people, especially Americans imagine how important machine translation is for people whose first languages aren't that common

Dan Liebling πŸ³οΈβ€πŸŒˆ
@liebling.org

Despite this work being pre-pandemic, and low-resource translation quality improving, barriers to access still exist, particularly for low-literacy populations. arxiv.org/abs/2002.12387

arxiv.orgUnmet Needs and Opportunities for Mobile Translation AITranslation apps and devices are often presented in the context of providing assistance while traveling abroad. However, the spectrum of needs for cross-language communication is much wider. To invest...
Dan Liebling πŸ³οΈβ€πŸŒˆ
@liebling.org

Nice to have concrete names for these phenomena #nlproc #linguistics

Thomas Hikaru Clark@thomashikaru.bsky.social

If you've noticed a distinct dialect emerging in frontier AI models and struggled to understand what they're talking about, you're not crazy! We turned to psycholinguistics to investigate what makes the writing style so weird, and why this might be happening 🚨New preprint! 🧡

Dan Liebling πŸ³οΈβ€πŸŒˆ reposted

Gautam Kamath
@gautamkamath.com

How do you tell if authors fabricated their research results? Sometimes: they just tell you. Read, e.g., this caption carefully. Another banger of a post by my friend Nihar B. Shah on his blog, where he found at least 20+ such #ICLR2027 submissions: researchonresearch.blog/2026/10/20/s...

Dan Liebling πŸ³οΈβ€πŸŒˆ reposted

Ai2
@ai2.bsky.social

Developing a strong research idea means knowing what prior work establishes, where gaps remain, and how your contribution addresses them. Introducing LitPivot, which helps refine early ideas by surfacing relevant literature as they evolve. buff.ly/iYol1M4

Dan Liebling πŸ³οΈβ€πŸŒˆ reposted

Ken McLeod
@kenmcld.bsky.social

Why roads are never officially dangerous "objective language can help agencies communicate clearly while avoiding statements that may unintentionally create or increase liability...subjective terms such as 'dangerous'... can introduce unnecessary risk" www.nationalacademies.org/projects/TRB...

nationalacademies.orgTRB Webinar: Words that Work for Liability Neutral Transportation CommunicationsTransportation professionals regularly develop emails, technical guidance, reports, and public-facing materials that communicate an agency’s decisions and actions. Precise, objective language can help agencies communicate clearly while avoiding statements that may unintentionally create or increase liability. TRB will host a webinar on Thursday, October 29, 2026 from 2:00 PM to 3:00 PM Eastern that will examine legal concepts related to liability-neutral language and how they apply to transportation communications. Presenters will share examples of language that can introduce unnecessary risk, including subjective terms such as β€œdangerous,” or β€œsufficient,” and demonstrate how statements can be revised using objective, fact-based language. Attendees will learn practical approaches for critically reviewing their own writing and improving internal and public-facing communications. This webinar is sponsored by TRB’s Standing Technical Committee on Safety, Risk Management and Tort Liability. The registration fee is $60. Purchase a webinar package for discounted pricing. Complimentary registration is available for employees of TRB sponsors, standing technical committee chairs, involved CRP panel members, media, congressional staff, and TRB or National Academies staff. Visit MyTRB.org for details. A recording will be sent to all registrants after the webinar. Webinar agenda and presenters Liability neutral examples and pitfalls – Terri Parker, Missouri Department of Transportation Application of liability neutral principles – Mike Fleming, Washington Department of Transportation Question and answer session moderated by Joshua Root, Minnesota Department of Transportation The first 40 minutes of the webinar will be for presentations and the final 20 minutes will be reserved for audience questions. Learning Objectives At the end of this webinar, participants will be able to: Identify legal considerations associated with word choice in transportation communications and recognize language that may create unintended liability Apply liability-neutral writing principles to develop clear, objective, and fact-based professional communications Professional Development Hours Professional engineers (PEs) who register and attend the entire webinar as an individual will be awarded 1 Professional Development Hour (PDH). TRB’s webinar program awards PDH credits only to the registered individual, and not to other members of a group that view the webinar together. To be eligible for the credits, individuals must remain signed in for the entire webinar, including for any question and answer session following the presentations. This webinar will be delivered as a scheduled – e-learning live event and is categorized within RCEP as Technical, Health and Safety. Please check with your licensing board to ensure that TRB webinar PDHs are approved by your board. TRB will report credits earned to the Registered Continuing Education Program (RCEP). Complaints about registered providers may be sent to RCEP, 1015 15th Street, NW, 8th Floor, Washington, D.C., 20005. Continuing Legal Education Credits TRB is offering 1 CLE credits for this webinar. TRB did not seek approval for this webinar from the state board, so we advise that you contact your state board prior to registering for the webinar to see if credits would be accepted. Registered webinar attendees who sign in and view the entire webinar will receive a link to a certificate that they can fill out and provide to their board. Complete a certificate for each state in which you are required to file. Rules for CLE in some states require the provider to file attendance with the regulator as a service to lawyers. Please confirm jurisdictional reporting requirements with the state regulator. This certificate should be filed with the appropriate MCLE Board(s) or Commission(s) within 30 days of activity. Introductory remarks, keynote addresses, business meetings, breaks, receptions, etc., are not included in the computation of credit. TRB will retain a copy of your attendance; however, TRB is not required to submit attendance to the board. Pre-registration is required. Once the webinar starts, you will be unable to register. Individuals that are registered for a webinar but do not receive their confirmation emails due to their organization’s internal firewalls should contact TRB Helpdesk for assistance in signing in. View our FAQ page for more information. Registration questions? Contact TRB webinar program.

Dan Liebling πŸ³οΈβ€πŸŒˆ reposted

Mark J. Nelson
@mm-jj-nn.bsky.social

MSR internship "focused on creating new interactive experiences that 'blend artificial and human intelligence'". Open to current PhD students in Comp. Sci. or a related field.

apply.careers.microsoft.comResearch Intern - Extended Perception Interaction Cognition (EPIC) | Microsoft CareersGenerative AI: Large Language Models (LLMs), Vision Language Models (VLMs), Stable Diffusion, video/world/action models, real-time models, multi-agent development, and related technologies Human-Compu...

Dan Liebling πŸ³οΈβ€πŸŒˆ reposted

allen institute
@alleninstitute.org

Teach neuroscience to undergraduates? Learn how to incorporate our connectomics data into your courses! 🧠 "Incorporating open connectomics data into teaching neuroscience" πŸ—“οΈ July 23-24 πŸ›« Travel funds available ⏳ Apply by Feb. 2: https://bit.ly/3LAfOdy #neuroskyence

alleninstitute.orgIncorporating open connectomics data into teaching neuroscienceLearn to analyze open neuroscience data and introduce dry lab modules into your existing classes at theΒ Incorporating Open Connectomics data into...

Dan Liebling πŸ³οΈβ€πŸŒˆ reposted

Alfie Kohn
@alfiekohn.bsky.social

Fascinating account showing how a false claim can become widely accepted (and propagated by multiple independent sources) as a result of a common academic shortcut: citing a source you haven't actually read after seeing it cited by another source: journals.sagepub.com/doi/full/10....

is.gdAcademic urban legends - Ole BjΓΈrn Rekdal, 2014Many of the messages presented in respectable scientific publications are, in fact, based on various forms of rumors. Some of these rumors appear so frequently,...

Dan Liebling πŸ³οΈβ€πŸŒˆ reposted

Seattle Streets Alliance
@streetsalliance.bsky.social

The idea of a car-free Ave has been around since 1955. This Oct, we get to see it in action! Join The Ave for All for 2 Saturdays of music, games, outdoor seating & space to enjoy the U District without car traffic. πŸ“… Oct. 10 & 17 πŸ•š 11 AM–8 PM πŸ“ University Way NE ℹ️ theaveforall.udistrict.org

Dan Liebling πŸ³οΈβ€πŸŒˆ reposted

Marzena Karpinska
@markar.bsky.social

Can you tell whether the #ideas in a piece of writing came from a human or an AI? As AI writing becomes common, policies on AI use focus on who came up with the ideas in the text rather than who wrote it. AI detectors answer the latter, not the former. We release #IdeaLens, a detector of AI ideas.

Dan Liebling πŸ³οΈβ€πŸŒˆ reposted

Richard Jean So
@richardjeanso.bsky.social

This Fri + Sat, Humanistic AI conference at Duke! Great line-up of speakers from lit studies, history, philosophy & computer science, working out what gen AI means for humanistic inquiry & what humanistic knowledge means for gen AI development. Schedule in link: sites.duke.edu/humanisticai...

Dan Liebling πŸ³οΈβ€πŸŒˆ
@liebling.org

shower thought today was that a whole generation of students have never trained a multiclass classifier, which generated some of the excitement (I know Jev is more complicated obvs)

David Mimno@dmimno.bsky.social

It's possible for Jev/Laya/Decision Models to be not that big a deal as tech and massive as a new paradigm. Here's why I'm really excited from an NLP history perspective (thread)

Dan Liebling πŸ³οΈβ€πŸŒˆ reposted

Simon Kirby
@simonkirby.bsky.social

We uncovered spontaneous evolution of new languages in populations of AI agents. This creates extraordinary scientific opportunities but also safety risks. New blog post with about how we created a platform for studying this safely. www.schmidtsciences.org/glossogen/

schmidtsciences.orgAI Agents Evolve Their Own LanguagesSchmidt Sciences’ new AI Agents Evolving Communication and Coordination pilot program works toward advancing foundational research on multi-agent communication and coordination, and building an open-s...

Dan Liebling πŸ³οΈβ€πŸŒˆ reposted

Cogan Lab
@coganlab.bsky.social

How does a speech plan become fluent movement? In Nature Human Behaviour, led by Kumar Duraivel: Intracranial recordings reveal how hierarchical speech plans become continuous motor sequences across planning, articulation, and monitoring networks. doi.org/10.1038/s415... 1/5

doi.orgDistinct neural processes link speech planning and execution - Nature Human BehaviourDirect recordings from the human brain reveal a hierarchy in planning to speak, with neural activity organizing whole syllables before the individual sounds that compose them are sequenced.
Dan Liebling πŸ³οΈβ€πŸŒˆ
@liebling.org

"Crash And Compile is a ACM-style programming contest crossed with a drinking game, where teams of two people try to solve as many programming problems as they can. ... each time their code fails to compile, produces the incorrect output or segfaults, the team must drink." github.com/krux702/cras...

github.comGitHub - krux702/crashandcompile: Crash and CompileCrash and Compile. Contribute to krux702/crashandcompile development by creating an account on GitHub.

Dan Liebling πŸ³οΈβ€πŸŒˆ reposted

tor
@torwager.bsky.social

Launching scienceofplacebo.org: a living bibliography of placebo and nocebo research. New papers screened & tagged daily by design, population, outcome and findings, with a searchable database (currently about 5,000 papers), an interactive researcher map, events, news and discussion.

scienceofplacebo.orgScience of PlaceboA research hub for the science of placebo and nocebo effects: a daily-updated feed of new papers, a searchable tagged database, bibliometrics, events, resources and discussion.
Dan Liebling πŸ³οΈβ€πŸŒˆ
@liebling.org

shower thought: AI building software systems falls apart because we rarely had examples of large systems being built from zero anyway. Systems nearly always evolve from simpler systems. ask anyone who has tried to will a new large system into existence