Excited to see this finally in print! Scientific American (@sciam.bsky.social) named Lvmin Zhang (@lvminzhang.bsky.social) to it's inaugural list of young scientists recognizing his work on developing AI for digital creation. Congratulations to Lvmin! www.scientificamerican.com/article/lvmi...

Lvmin Zhang: Young American Scientist creating AI tools to help digital artists

Developing artificial-intelligence systems to enhance digital art creation

scientificamerican.com

A big thank you to our amazing organizing team: Mia Tang, Kartik Chandra, Maneesh Agrawala, Judith Fan, and Aaron Hertzmann. And to our sponsors AdobeResearch and Reve for supporting us! We are excited to keep this interdisciplinary conversation going at #SIGGRAPH2026. See you there!

Self-consistency has proven to be an effective method for improving LLM performance on tasks with language outputs (e.g. QA). We sample multiple LLM responses and group them to find the most consistent answer. But how can we apply self-consistency to visual domains? vimeo.com/1178981854?f...

Self-Consistency for LLM-Based Motion Trajectory Generation and Verification

Video figure of the CVPR 2026 paper Self-Consistency for LLM-Based Motion Trajectory Generation and Verification

vimeo.com

New paper in press at Psychology of Aesthetics, Creativity, and the Arts w/ @aaronhertzmann.com interrogating the information bottlenecks that make the drawing of objects & scenes from observation so difficult. This was a really fun paper to write together--we hope it'll be a fun read, too!

Aaron Hertzmann@aaronhertzmann.com · 4mo ago

If asked to draw a person's portrait, many people would say “I cannot draw.” Why is observational drawing so hard? In a new paper, @judithfan.bsky.social and I answer: it's the limitations of human vision. Learning to draw is learning skills to overcome them. aaronhertzmann.com/2026/03/23/d... 1/

Nice spatial study of London that goes a level deeper than what principal component analysis reveals

Lauren Leek@laurenleek.eu · 5mo ago

Sydney named its socio-economic divide the "latte line" and has been arguing about who drew it for 20 years. London has the same divide and calls it "character." I built a machine learning model to do the impolite thing: draw it and blame someone. open.substack.com/pub/laurenle...

Nice Commentary on HCI Education in the Jan issue of Interactions Magazine that mentions our Gulf of Envisoning Work with Maneesh @magrawala.bsky.social Chris Pondoc, Roy Pea, and Colleen Seifert: interactions.acm.org/archive/view...

New demo! youtu.be/rCPLSr79aGg Use Cartokit to map your data using familiar GUI-style interactions—and let Cartokit write the equivalent code for you. Try it out at alpha.cartokit.dev.

Cartokit: A Geo Tool That Writes Code

YouTube video by Sarah E. Chasins and the PLAIT Programming Lab

youtu.be

Didn't expect to run into Chernoff faces in a scifi book about the nature of consciousness and a world where transhumans live alongside vampires. en.wikipedia.org/wiki/Blindsi... en.wikipedia.org/wiki/Chernof...

Blindsight (Watts novel) - Wikipedia

en.wikipedia.org

that is also why i asked about comfyui,regional prompting and controlnet becuse in relation to 1 those and things like img2img do allow for more expression and control then there are project like that by chicano artist coco which are more in depth www.youtube.com/watch?v=envM...

A Love Letter to LA | Behind The Scenes

YouTube video by Cuco

youtube.com

Excited to get back to UC Berkeley this Wed Nov 19. I'll be speaking in the EECS colloquium 4-5pm Soda Hall with a talk entitled: Making is Decision Making Abstract and directions are at this link: events.berkeley.edu/eecs/event/3...

EECS Colloquium | Making is Decision Making

Dedicated to hosting presentations at the intersection of Electrical Engineering and Computer Science, the EECS Colloquium offers a mix of short ta...

events.berkeley.edu

MoVer is a Motion Verification DSL that can check if an LLM-generated motion graphics animation match your text prompts! @jiajuma.bsky.social presented this work at SIGGRAPH 2025 making it easier for designers to specify and verify complex animations. Paper, code and dataset: mover-dsl.github.io

Oof -- still learning to post proper links on Bluesky.

Maneesh Agrawala@magrawala.bsky.social · 11mo ago

Artists often explore designs, undoing and redoing parts of a sketch to try different ideas. Lvmin Zhang's PaintsAlter (presented at SIGGRAPH 2025) is a tool designed to facilitate such exploration Updated link: lllyasviel.github.io/paints_alter... w/ Chuan Yan, Yuwei Guo and Jinbo Xing

Artists often explore designs, undoing and redoing parts of a sketch to try different ideas. Lvmin Zhang's PaintsAlter (presented at SIGGRAPH 2025) is a tool designed to facilitate such exploration Updated link: lllyasviel.github.io/paints_alter... w/ Chuan Yan, Yuwei Guo and Jinbo Xing

Artists often explore possible designs, undoing and redoing parts of a sketch to try different ideas. Lvmin Zhang's PaintsAlter (presented at SIGGRAPH 2025) is a tool designed to facilitate such exploration Paper and code: lllyasviel.github.io/paints_alter... w/ Chuan Yan, Yuwei Geo and Jinbo Xing

If you need to segment sketches into objects, Mia Tang has the tool for you. Enables re-composing the sketch after drawing. She presented her work on InkLayer at SIGGRAPH 2025. Code, data and paper are here: inklayer.github.io w/ @yaelvinker.bsky.social Chuan Yan, Lvmin Zhang

Now that school is starting for lots of folks, it's time for a new release of Speech and Language Processing! Jim and I added all sorts of material for the August 2025 release! With slides to match! Check it out here: web.stanford.edu/~jurafsky/sl...

Was recently introduced to Brian Marick's Oddly Influenced podcast. It's an amazing mix of ideas connecting software engineering, cognitive psychology and history of science. Highly recommended! Has significantly extended my to-read list. podcasts.apple.com/us/podcast/o...

Oddly Influenced

Technology Podcast · A podcast about how people have applied ideas from outside software to software.

podcasts.apple.com

Looking forward to this workshop!

@yaelvinker.bsky.social · last yr.

I'm very excited to announce our #SIGGRAPH2025 workshop: Drawing & Sketching: Art, Psychology, and Computer Graphics 🎨🧠🫖 🔗 lines-and-minds.github.io 📅 Sunday, August 10th Join us to explore how people draw, how machines draw, and how the two might draw together! 🤖✍️

This is a good article and I completely agree with its argument for comprehensible research.

Richard McElreath 🐈‍⬛@rmcelreath.bsky.social · last yr.

How can we reform science? I have some ideas. But I am not sure you’ll like them, because they don’t promise much. elevanth.org/blog/2025/07...

For context, AlphaFold is great, but is not a magic wand for making new drugs. (But Hassabis's latest claims go far beyond the hype that accompanied AlphaFold at the time.) www.prospectmagazine.co.uk/ideas/techno...

DeepMind has predicted the shape of every protein known to science. How excited should we be?

The AI-based breakthrough, like the sequencing of the human genome, may not turbocharge the development of new drugs as some hope. But as a resource f...

prospectmagazine.co.uk

The AI for Content Creation workshop is kicking off today at #CVPR2025 - Grand Ballroom A1 - @magrawala.bsky.social Kai Zhang (Adobe), Charles Herrmann (Google), Mark Boss (Stability AI), Yutong Bai (UC Berkeley), Cherry Zhao (Adobe), Ishan Misra (Meta) and @jonbarron.bsky.social ! See you soon!

At CVPR in Nashville for the Workshop on AI for Content Creation. I'll be speaking at 9am tomorrow 6/12. ai4cc.net

AI for Content Creation Workshop

ai4cc.net

There’s still time! If you are in Toronto I’d love to see you

@uoftdsi.bsky.social · last yr.

Get your tickets and join us + @uoftmscac.bsky.social for an upcoming #DataSciencesSpeakerSeries talk: "Unpredictable Black Boxes are Terrible Interfaces" w/ Prof. @magrawala.bsky.social from @stanforduniversity.bsky.social 🗓️April 14 📷Info & reg: datasciences.utoronto.ca/event/prof-m...

Join DSI and @uoftmscac.bsky.social for an upcoming #DataSciencesSpeakerSeries talk: "Unpredictable Black Boxes are Terrible Interfaces" w/ Prof. @magrawala.bsky.social from @stanforduniversity.bsky.social 🗓️Friday, April 14 🔗Info & reg: datasciences.utoronto.ca/event/prof-m...

What will be the linchpin for AI dominance? Read our NSF/OSTP recommendations written with Goodfire's Tom McGrath tommcgrath.github.io, Transluce's Sarah Schwettmann cogconfluence.com, MIT's Dylan Hadfield-Menell @dhadfieldmenell.bsky.social TLDR; Dominance comes from **interpretability** 🧵 ↘️

Looking forward to giving a connected pair of upcoming talks! Apr 14: I'll present Unpredictable Black Boxes are Terrible Interfaces for the Data Science Institute in Toronto datasciences.utoronto.ca/event/prof-m...

Unpredictable Black Boxes are Terrible Interfaces - Prof. Maneesh Agrawala - DSI

Join us for the Inequity Speaker Series with Associate Professor Dustin Duncan, from the Department of Epidemiology and Co-Director of the Social and Spatial Epidemiology Unit, Columbia University.

datasciences.utoronto.ca

Ran into this great article on Navigating Nostalgia from Jonathan Bean. It's about nostalgia for older techniology and interfaces, and it make a lovely reference to LineDrive maps.

Navigating Nostalgia: Technology, Intention, and the Road Ahead | IX Magazine Issue XXXII.2 March - April 2025

I recently took a road trip from Tucson, Arizona, to Portland, Oregon, to spend time with family and to spare our spoiled dog the indignity of flying in the cargo hold. I ended up...

interactions.acm.org

In Barbados for the Bellairs workshop on computer animation!

Hello Bluesky! 🔵 We start our account by having our third guess for Ask Me Anything session #3DV2025AMA! Noah Snavely @snavely.bsky.social from Cornell & Google DeepMind! 🌟 🕒 You have now 24 HOURS to ask him anything — drop your questions in the comments below! Keep it engaging but respectful!

my take on bringing HCI to PL blog.sigplan.org/2024/11/21/e...

Evaluating Human Factors Beyond Lines of Code

Software systems researchers want to make human-centered claims, but don’t have the proper tools to do so. That’s how we ended up with the ubiquitous lines-of-code comparison found in e…

blog.sigplan.org