Minsuk Chang

gemma4 is really great. results comparable to sonnet 4.6 but costs pennies.

my side project budgets are basically $0, so the new gemini flex pricing in AI.dev is actually saving my life 😅 50% discount if you don't need instant responses! perfect for background agents, batch jobs, or evals while you sleep or go watch robot cage fights 💸👇 ai.google.dev/gemini-api/d...

New paper: “𝐀 𝐓𝐡𝐞𝐨𝐫𝐲 𝐨𝐟 𝐀𝐩𝐩𝐫𝐨𝐩𝐫𝐢𝐚𝐭𝐞𝐧𝐞𝐬𝐬 𝐓𝐡𝐚𝐭 𝐀𝐜𝐜𝐨𝐮𝐧𝐭𝐬 𝐟𝐨𝐫 𝐍𝐨𝐫𝐦𝐬 𝐨𝐟 𝐑𝐚𝐭𝐢𝐨𝐧𝐚𝐥𝐢𝐭𝐲” Agent-based models of social order work better when agents act by predictive pattern completion from prefix (culture/context) to suffix (action) than when they act through expected value maximization

To anyone encountering Moltbook this week and wondering about AI personhood, consciousness, sentience, etc---we published a very relevant paper in October: A Pragmatic View of AI Personhood. arxiv.org/abs/2510.26396

🚨 🔬 PhD positions at Google DeepMind in France 🇫🇷 We are advertising Master Level Intern positions at Google DeepMind within our Frontier AI Unit. These could lead to co-advised PhD positions with Google DeepMind and French academic institutions. job-boards.greenhouse.io/deepmind/job...

cool class alert @ UW: Prototyping with AI for Performing Arts docs.google.com/document/d/e...

Heads up! 🚨 Mark your calendar for next week! 🚨 More games to be released on the @kaggle.com Game Arena!!! 🎉🙌 See how foundations models perform at poker ♣️♥️♠️♦️ and werewolf 👫🐺!! And more Chess results too! 🧠♟️ 🤩 🤩 🤩

@waymo.bsky.social Waymo’s biggest missed opportunity is not being the official chauffeur for everyone under 16. Safe, tracked, and never gets tired of "Golden" on repeat. Let’s make it happen. Waymo Junior.

opendoor, opentable, openai, what else are we going to open?

🙌 Join the Gemini 3 global hackathon for early API access to build next-gen apps! We’ve got $100k in prizes + funding opportunities for the best builds. If you've yearned to start a passion project or to kick the tires on an idea for a company, there's never been a better time. gemini3.devpost.com

If you're are an HCI student navigating the industry job market, here’s my unsolicited take on why you should stop defending 'human-centeredness'. I’m rarely right. Read at your own risk: minsukchang.com/blog/2026-01...

People wanted our Open Source Organizations starter pack to include many projects, so we decided to give them their own starter pack. go.bsky.app/HvKFRKa

Recently went to Goodwill for a ceramic planter. Left with Michelle Obama's autobiography for $2 instead (there were three of them???) Still looking for the planter...

All this just to remind the timeline that Eintracht Frankfurt is the greatest football club on earth. I will die on this hill. 🦅✌️ #SGE

In the month of December, Seattle (where I live now) gets 52.9 hrs of sunlight on average, London (where my employer is) gets 52 hrs, Frankfurt (where I grew up) gets 41. Seoul gets 152.6 hrs, more than all three put together.

Please welcome Google's Open Source efforts to Blue Sky at @opensource.google!

Okay people. I've achieved Nobel Trajectory on this Tenure-track game (tenuretrack.vercel.app/game2d.html) Kangwook (kangwooklee.com) made. This is going into my faculty application if I ever apply.

Gentle reminder for the RL folks out there: RLC 2026 → Montreal 🇨🇦 (Aug 16–19) Abstract: Mar 1 | Submission: Mar 5 (AOE) | rl-conference.cc/callforpaper... Looking forward to receiving your submission!

I started a crowdsourced meta-project to list all the other crowdsourced mathematical research projects that are currently active and seeking participants: mathoverflow.net/questions/50...

Trying this place again.

My colleague Sanjay Ghemawat & I have done a fair bit of performance tuning of various pieces of code. We wrote an internal Performance Hints document ~2 years ago as a way of identifying some general principles & we've recently published a version of it externally. Doc: abseil.io/fast/hints.h...

www.cabinetmagazine.org/issues/2/ric... From almost 25 yrs ago!

Come be our colleague in the robotics and embodied intelligence center at NYU! 🔷 Professor in Robotics / Embodied AI (Open Rank) apply.interfolio.com/176977 🔷 Faculty Fellow in Robotics / Embodied AI apply.interfolio.com/177077

Physical Intelligence has a recipe for real world RL on top of VLAs and it looks impressive: www.pi.website/blog/pistar06

Hundreds of hours of European driving data from NVIDIA! 1700 hours total

Trying to keep my professional chill but I’m SO excited Carnegie Mellon is launching a cluster hire in computational humanities—MULTIPLE JOBS! 1. Asst Teaching Track Prof in Computational Humanities - apply.interfolio.com/173622 2. Asst Tenure Track Prof in CH - apply.interfolio.com/173626

Our NeurIPS submission arxiv.org/abs/2502.08938 did not get in, but it's one of my favorite papers and I think one of the better papers we've ever put out so I want to highlight it

Do you have a PhD (or equivalent) or will have one in the coming months (i.e. 2-3 months away from graduating)? Do you want to help build open-ended agents that help humans do humans things better, rather than replace them? We're hiring 1-2 Research Scientists! Check the 🧵👇

Archive of archives. worrydream.com/Links2025/

Claude Code, meet your rival: Gemini CLI 🔥🚀 Google just unleashed Gemini 2.5 Pro at the command line—free, open-source (Apache 2.0) , 1-million-token brain baked right into your terminal. Chat, code, search, script—all with zero tab fatigue.

Don't have a robot? Try our newest Gemini Robotics on-Device VLA in simulation! Or become a trusted tester and tune and adapt the model yourself! www.youtube.com/watch?v=nVMY...

I really think ACs can be completely replaced by LLMs. Finding the right reviewers, sending (reminder) emails, nudging them for discussions, summarizing reviews, etc... The only issue is that LLM calls are expensive, whereas human ACs work for free because we all volunteer. Sigh.

Hiring a postdoc to scale up and deploy RL-based planning onto some self-driving cars! We'll be building on arxiv.org/abs/2502.03349 and learn what the limits and challenges of RL planning are. Shoot me a message if interested and help spread the word please! Full posting to come in a bit.

I work in the best university in the world, less than 3 hours from my rural hometown, yet superstar students travel here from all over the world… if we keep treating them badly, will centers of excellence move elsewhere? "If my kids excel, will they move away?" jeffreybigham.com/blog/2025/wh...

👀This is impressive: Using an autonomous agent based on o3-mini and GPT-4.1, a team from Harvard, MIT & other institutions reproduced and updated an entire issue of Cochrane Reviews in two days… saving 12 person-years of work. The AI reviews captured more papers & were more accurate than humans.

Yes!! Originally, generative ML modeled a joint distribution, e.g., P(X,Y), whereas discriminative ML modeled a conditional distribution, e.g., P(Y|X). A famous paper by Ng and Jordan (papers.nips.cc/paper_files/...) compared naive Bayes (generative) to logistic regression (discriminative). 1/

🤖💡 Are LLMs really developing social behaviors—or just reproducing familiar patterns from training? In our new paper, @pettertornberg.com and I take a closer look at recent claims of emergent conventions in AI and offer a different perspective... 🧵

Weave household robot cleaning up. This is teleop, but it's a really cool first step towards a commercial product.

Humans' ability to invent their own games & goals is at the core of open-ended learning. Understanding and modeling computationally how they do it would be enlightening to understand better human cognition and build open-ended AI Great step in this direction in new paper by Guy Davidson et al.

TPUs are amazing.

TIL Canada has its own professional men's basketball league and Vancouver has a team.

Announcing AlphaEvolve, our new LLM coding agent that has - made new scientific discoveries - discovered algorithms that are now deployed at Google (in Gemini, Transformers, TPU hardware design & data centers) Blog: deepmind.google/discover/blo... White paper: storage.googleapis.com/deepmind-med...

The idea of "AI alignment" grew out of a community that thought you could solve morals like it was a CS problem set. Nice to see a more nuanced take.

[1/2] We've released the code for LegoGPT. Our autoregressive model generates physically stable and buildable designs from text prompts by integrating physics laws and assembly constraints into LLM training and inference. Code: github.com/AvaLovelace1... Website: avalovelace1.github.io/LegoGPT/

Coach Pop, please write a book.

:thumbs_up: robhorning.substack.com/p/contentment

How do language models generalize from information they learn in-context vs. via finetuning? In arxiv.org/abs/2505.00661 we show that in-context learning can generalize more flexibly, illustrating key differences in the inductive biases of these modes of learning — and ways to improve finetuning. 1/

This is a fantastic oral history of the last 10 years of NLP and AI. www.quantamagazine.org/when-chatgpt...