Sign in

Dan Liebling 🏳️‍🌈

@liebling.org
201 followers 187 following 455 posts

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

PostsRepliesMedia
Dan Liebling 🏳️‍🌈 @liebling.org · 28/09/2026
"Locating Translation as a Craft in the Age of AI" www.semanticscholar.org/paper/Locati...
semanticscholar.org
000
Dan Liebling 🏳️‍🌈 @liebling.org · 26/09/2026
when the floof grows in
Adult red Merle Australian Shepherd sleeping on a couchJuvenile red Merle Australian shepherd sleeping on a couch with a blanket and two pillows
010
Dan Liebling 🏳️‍🌈 @liebling.org · 25/09/2026
Seattle Open Orchestra’s first show was very compelling. I felt the theme of openness as well as permission to create. www.waywardmusic.org/event/seattl...
waywardmusic.org
Seattle Open Orchestra - Wayward Music Series
In Seattle Open Orchestra’s inaugural concert, the group will explore unconventional, experimental, and improvised music with radical inclusion and collaboration as its guiding ethos. Seattle Open Or...
000
Reposted by Dan Liebling 🏳️‍🌈
Richard Jean So @richardjeanso.bsky.social · 23/09/2026
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...
0204
Dan Liebling 🏳️‍🌈 @liebling.org · 23/09/2026
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)
000
Reposted by Dan Liebling 🏳️‍🌈
Joe Bak-Coleman @jbakcoleman.bsky.social · 20/09/2026
I don’t mean this to be cruel, but there’s just too much air being taken up by people with media or tech backgrounds who think their hot take on consciousness with LLMs is new, and lord it has all been done before and we’re past that. Christ, half the news cycle is yud refusing to read a book.
738364
Dan Liebling 🏳️‍🌈 @liebling.org · 19/09/2026
the rest, as they say, is history
020
Dan Liebling 🏳️‍🌈 @liebling.org · 16/09/2026
gonna vibe code a telephone voice agent whose number I give to every "wanna sell your house" text spammer. keep 'em occupied for a while at least
010
Reposted by Dan Liebling 🏳️‍🌈
Simon Kirby @simonkirby.bsky.social · 15/09/2026
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.org
AI Agents Evolve Their Own Languages
Schmidt 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...
33014
Reposted by Dan Liebling 🏳️‍🌈
Cogan Lab @coganlab.bsky.social · 18/08/2026
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.org
Distinct neural processes link speech planning and execution - Nature Human Behaviour
Direct 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.
12914
Dan Liebling 🏳️‍🌈 @liebling.org · 11/09/2026
"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.com
GitHub - krux702/crashandcompile: Crash and Compile
Crash and Compile. Contribute to krux702/crashandcompile development by creating an account on GitHub.
001
Dan Liebling 🏳️‍🌈 @liebling.org · 10/09/2026
interesting PhD thesis on flax for violin construction biblio.ugent.be/publication/...
biblio.ugent.be
Violin making with fiber-reinforced composites : an artistic and scientific research journey
110
Dan Liebling 🏳️‍🌈 @liebling.org · 09/09/2026
happy Hoo Hoo day to those that celebrate historylink.org/File/8670#:~....
historylink.org
Alaska-Yukon-Pacific Exposition in Seattle celebrates California Day,
On September 9, 1909, California Day is celebrated at the Alaska-Yukon-Pacific Exposition on the grounds of the University of Washington in Seattle. Five other, smaller days are also celebrated at th
000
Dan Liebling 🏳️‍🌈 @liebling.org · 09/09/2026
agent just made a Helper class 🤢 its_the_90s.gif
000
Dan Liebling 🏳️‍🌈 @liebling.org · 08/09/2026
billions of tokens inflight for P=NP any day now right, do a breakthrough
011
Reposted by Dan Liebling 🏳️‍🌈
Vilém Zouhar @zouhar.bsky.social · 04/09/2026
Machine translation is not solved and it will take a while for it to be done arxiv.org/abs/2609.04173
arxiv.org
Last Translation Benchmark
For scientific progress, we need benchmarks that test the limits of state-of-the-art models, and evaluation methods that inform us about failure cases. As models get stronger, standard benchmarks for ...
34612
Reposted by Dan Liebling 🏳️‍🌈
tor @torwager.bsky.social · 03/09/2026
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.org
Science of Placebo
A 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.
22410
Dan Liebling 🏳️‍🌈 @liebling.org · 03/09/2026
TIL Google has a Jimothy easter egg www.google.com/search?q=jim...
google.com
Google Search
010
Dan Liebling 🏳️‍🌈 @liebling.org · 02/09/2026
🤯🤯
000
Dan Liebling 🏳️‍🌈 @liebling.org · 24/08/2026
Free CHI paper idea is to find incidences where AI Overview hallucinates a trend that didn’t exist, but this hallucination catalyzed into an actual trend
AI Overview claiming that ppl are urinating on Flock cameras and putting “peeing Calvin” stickers there. The query was “peeing Calvin flock”
000
Dan Liebling 🏳️‍🌈 @liebling.org · 20/08/2026
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
011
Dan Liebling 🏳️‍🌈 @liebling.org · 20/08/2026
I was fortunate to have bio teachers in high school that really loved the material. Though I didn't yet know the term "systems thinking," yet, that was my lens, not just memorizing labels. Then I went on to study neuroscience. So thanks to my HS bio teachers! jsomers.net/i-should-hav... (via HN)
jsomers.net
jsomers.net | I should have loved biology
In biology class, biology wasn't presented as a quest for the secrets of life. The textbooks wrung out the questing.
030
Dan Liebling 🏳️‍🌈 @liebling.org · 13/08/2026
index all the little free libraries bookstoop.org
bookstoop.org
BookStoop Ballard
A live, neighbor-powered catalog of Ballard's little free libraries. Find your next book two blocks away.
020
Dan Liebling 🏳️‍🌈 @liebling.org · 11/08/2026
funny to see speech translation devices invented over and over again. turns out the hard part is not the model -- when I joined Google 9 years ago we first demonstrated porting Google's ASR and MT models to the NVIDIA Jetson TX2 board. Here's Gemma on Raspberry Pi 5: www.youtube.com/watch?v=4dNr...
youtube.com
Gemma Translator: offline DIY translation device built with Gemma 4 + Antigravity
YouTube video by Google Antigravity
220
Reposted by Dan Liebling 🏳️‍🌈
Oliver Haimson @haimson.bsky.social · 10/08/2026
"Just sprinkle some AI on it..." We're staging a live PLAY🎭 at #CSCW2026 and looking for amateur actors - no experience needed! It's based on AI Powder, our speculative fiction piece on AI techno-solutionism (forthcoming in Catalyst), paired with an academic panel. Reply/message if interested!
052
Dan Liebling 🏳️‍🌈 @liebling.org · 05/08/2026
the average delay is lower than the range 🤔
020
Dan Liebling 🏳️‍🌈 @liebling.org · 02/08/2026
still shocked that Claude uses Python to read docx
010
Dan Liebling 🏳️‍🌈 @liebling.org · 30/07/2026
pretty inclusive of MacOS to include BAMUM LETTER PHASE-C MBAA CABBAGE-TREE (𖢞) when you search for the "bag" emoji
100
Reposted by Dan Liebling 🏳️‍🌈
Cat Hicks @grimalkina.bsky.social · 27/07/2026
SEATTLE friends: Cat-in-real-life alert! I am in Seattle this week and decided to host a small (limited spots) Psychology of Software Teams happy hour this thursday July 30th, 4-6pm. RSVP here for location & to secure a spot: luma.com/lxcmtn68
luma.com
The Psychology of Software Teams: Seattle happy hour edition · Luma
Let's talk about being more human in tech! Join a small group for a happy hour meetup talking about getting better psychology in our software and technical…
3298
Reposted by Dan Liebling 🏳️‍🌈
Cabel Sasser @cabel.panic.com · 26/07/2026
the temu app will be studied for generations. i opened the app. here’s my unedited, nearly two-minute launch sequence. i was just searching for a bookshelf
972116566
Dan Liebling 🏳️‍🌈 @liebling.org · 24/07/2026
the Sitterwerk library has a "dynamic" shelf order. Books are findable by RFID, as this ordering encourages reshuffling and serendipity www.sitterwerk.ch/en/Dynamisch...
sitterwerk.ch
Stiftung Sitterwerk • Dynamic Order
en
020
Dan Liebling 🏳️‍🌈 @liebling.org · 22/07/2026
trees of AI agents are reinventing human process asking for reports and status updates. next they’ll be circling back
031
Reposted by Dan Liebling 🏳️‍🌈
Ben Williamson @benpatrickwill.bsky.social · 18/07/2026
The anatomy of an AI-supported academic article is becoming more apparent. I desk reviewed ~70 manuscripts this week for our journal. Reckon 20ish had significant AI assistance. On top of hundreds of others recently I'm starting to see patterns. Are other editors are seeing the same or similar? 1/
28932417
Reposted by Dan Liebling 🏳️‍🌈
Data Science Institute @dsi-uchicago.bsky.social · 08/07/2026
Engage with the study, application, and use of AI and Machine Learning to enable scientific discovery across the physical and biological sciences at the UChicago and @caltech.edu AI+Science Conference 2026! Register by August 28, 2026 datascience.uchicago.edu/research/ai-...
052
Reposted by Dan Liebling 🏳️‍🌈
The Data Therapist in the Blue Sky @datatherapist.bsky.social · 06/07/2026
🚨Emergency reviewers needed for: 1. Reasoning with backtracking 2. Agent personas 3. Graph LLM reasoning 4. Story generation 5. Agentic RL Please reply / DM / contact me ASAP if you can review any of the above in 1-1.5 days. Thanks for your urgent help! Please BOOST. #ARR #ACL #ARR2026May
121
Dan Liebling 🏳️‍🌈 @liebling.org · 02/07/2026
sometimes working on Science AI you come across papers with unexpected titles like "Some Factors Affecting Dairy She-Camel Performance" www.ejfa.me/index.php/jo...
ejfa.me
SOME FACTORS AFFECTING DAIRY SHE-CAMEL PERFORMANCE | Emirates Journal of Food and Agriculture
000
Reposted by Dan Liebling 🏳️‍🌈
media archaeology lab @mediaarchaeology.bsky.social · 22/06/2026
Our residency call for the 26/27 academic year is live! The theme this year is Creative Misuse - interpret as you will. Read the full prompt: https:// drive.google.com/file/d/16zB-7J4UoajGAooans30yEgQKlzys2hD/view Fill out the application: https:// forms.gle/KNfcAGBVCbyURiWJA
a MISTY robotics robot appears via security camera on the screen of an Apple IIc computer. An explainer for the Apple IIc is under the screen.
14025
Dan Liebling 🏳️‍🌈 @liebling.org · 29/06/2026
petition to invent mandatory 1st person pronouns for AIs
110
Reposted by Dan Liebling 🏳️‍🌈
Dr Naomi Scott @drnaomiscott.bsky.social · 28/06/2026
A lot of media coverage of the ‘AI’ reading of the Herculaneum scrolls really fails to understand what’s happened here. There’s no language processing happening, the AI/ML is not trained on a corpus of texts or language. It can detect ink to see shapes. HUMAN EXPERTS are then transcribing the text.
A screenshot from here: https://www.ox.ac.uk/news/2025-02-07-inside-herculaneum-scroll-seen-first-time-almost-2000-years

It explains that the machine learning involves only the detection of ink
382499944
Reposted by Dan Liebling 🏳️‍🌈
mia ridge @miaout.bsky.social · 16/06/2026
I'm biased but there's an incredible Digital Humanities Research Software Engineer (DH RSE) job going at the British Library, working with the Digital Research team, curators and more ce0752li.webitrent.com/ce0752li_web... #JobFairy #RSEs #JobKlaxon
ce0752li.webitrent.com
Powered by MHR
Powered by MHR
0916
Reposted by Dan Liebling 🏳️‍🌈
yvesfrtl.bsky.social @yvesfrtl.bsky.social · 26/06/2026
📚 AI-written stories get attention, but AI-translated literature is quietly shaping how readers experience the author So what gets lost ⁉️ 🤖 AI translation into English is readable and often ‘fine’ ✍️ But human translation is valued more ‼️Both vary in quality BUT AI more, even within a single book
First page of the paper "AI translation of literary texts is ‘fine’, but readers still prefer human translations".
1125
Reposted by Dan Liebling 🏳️‍🌈
Maria Antoniak @mariaa.bsky.social · 24/06/2026
What are people really doing with chatbots? 👀 Turns out that a lot of them are writing/reading various kinds of fiction, and in particular, they're generating a ton of erotica! What does this mean for the future of reading? Paper: arxiv.org/abs/2606.22748 Web Explorer: ai-fiction-wild.com
36010
Dan Liebling 🏳️‍🌈 @liebling.org · 23/06/2026
proud of our #Caltech alumni who contributed to Engineers Without Borders to provide clean drinking water www.caltech.edu/about/news/c...
caltech.edu
Caltech Engineers Without Borders Chapter Implements Revitalized Water Supply System in Ecuadorian Village
A new protected collection area and replaced pipeline will help ensure access to clean drinking water in Tandacato
010
Dan Liebling 🏳️‍🌈 @liebling.org · 23/06/2026
"Thought Crimes for Cello" -- 18 vignettes where "every piece is about a person, or pet, or neighbourhood anomaly." brianpatrickbromberg.com/Thought-Crim...
brianpatrickbromberg.com
Thought Crimes for Cello — Brian Patrick Bromberg, Composer & Cellist
2005 — ONGOING / SOLO Thought Crimes for Cello THOUGHT CRIMES BOOK 1  Mary’s Gone Feral Moe The Moa︎ Off-Colour Joke︎ Bobby...
000
Dan Liebling 🏳️‍🌈 @liebling.org · 22/06/2026
before final publication, if a paper says it "will publish" the data then that data needs to be published, or the paper should be rejected. by far most papers that promise to release data, never do
000
Dan Liebling 🏳️‍🌈 @liebling.org · 19/06/2026
Kiwi the Australian Shepherd in the USA is conflicted ⚽️⚽️ (fun fact, Australian Shepherds are as Australian as Outback Steakhouse)
Red Merle Australian Shepherd on a beach
030
Dan Liebling 🏳️‍🌈 @liebling.org · 19/06/2026
Pasadena, CA folks: if you are looking for an office manager with who is highly quantitative, I have a recent Caltech grad who would be interested!
000
Dan Liebling 🏳️‍🌈 @liebling.org · 18/06/2026
$META not allowing people to leave orgs reminds me of when Microsoft's Windows division had a no-transfer policy, which of course did not end well. Basically people left $MSFT, leveled up or took time off, and came back to a happier division
000
Reposted by Dan Liebling 🏳️‍🌈
Eve Fleisig @ FAccT 2026 @efleisig.bsky.social · 17/06/2026
Remember the backlash against all AI voices being feminine—Siri, Alexa, etc.? Now, companies use a wider range of AI voices with ambiguous names like “alloy” and “shimmer.” Bias solved forever—ha, no. What happens when people hear these supposedly neutral voices?👇
1105
Reposted by Dan Liebling 🏳️‍🌈
Daniel van Strien @danielvanstrien.bsky.social · 11/06/2026
Can the new DiffusionGemma model help fix broken OCR? In theory, denoising tokens in parallel could work better for OCR correction since context is seen upfront? Pointed it at 19th-century newspaper OCR. It corrected better than the autoregressive baseline — at ~8x the speed.
47015