How many of these will we have before we do something about it?

The web used to look like this

The robotics revolution is here.

This makes me wondering if num_likes/num_followers could be used as a metric for engagement 🤔

📢🎓 We have open PhD positions in Computer Vision & Machine Learning at @tuda.bsky.social and @hessianai.bsky.social within the Reasonable AI Cluster of Excellence — supervised by @stefanroth.bsky.social, @simoneschaub.bsky.social and many others! www.career.tu-darmstadt.de/tu-darmstadt...

@mmitchell.bsky.social herself provided a plethora of suggestions here: github.com/bluesky-soci...

This is a very good point. It is one of the reasons why I think generic chatbots should probably be outlawed.

Something to consider, along with everything else, when you consider leaving here for there. By posting there, you are contributing to an ecosystem that is designed for the destruction of specific values. You ARE contributing to it. Maybe the tradeoff is worth it to you. For me, it's not.

I agree for most of it, but is it the developer/vendor that needs to list acceptable usecases or a higher regulation authority? 🤔

Just for those not aware, Margaret Mitchell is an AI ethicist who was fired by Google after co-authoring the famous "Stochastic Parrots" paper that criticized the entire project of ever-larger LLMs way before most of the people now blocking her had ever heard of them. www.nytimes.com/2021/02/19/t...

Blocking hatemongers is one thing but this trend of using mass block lists to avoid hearing from people who might or might not hold points of view you expect to disagree with is unhealthy imo and hurts the platform. Bluesky was better off with folks like Margaret Mitchell and Giada Pistilli on it.

What if we could represent events (event cameras) in a way that preserves both asynchrony and spatial sparsity? Exited to share our latest work where we answer this question positively. Spiking Patches: Asynchronous, Sparse, and Efficient Tokens for Event Cameras Paper: arxiv.org/abs/2510.26614

I have a serious idea. Should we start making anti-anti-ai lists? 😈

Something needs to be done ASAP @bsky.app. Great people are leaving the platform due to this nonsense.

Not just an amazing dataset, excellent 3D pose, shape, and eye tracking in the real world! Bravo @trackingskills.bsky.social et al! PS - Exactly this kind of data we predict will be critical for long-term health monitoring @monosovlab.bsky.social et al: www.sciencedirect.com/science/arti...

LLMs can memorize even a phone number seen once in training.🔒 Google’s VaultGemma fixes that, being the first open-weight LLM trained from scratch with differential privacy, so rare secrets leave no trace. ☕ new video explaining Differential Privacy through VaultGemma 👇 🎥 youtu.be/UwX5zzjwb_g

Exactly what I've been saying. @dimadamen.bsky.social if you remember our discussion here.

40M users! 🥳

Turns out NLP is just vision

Important paper on LoRA vs full fine-tuning for continual learning: arxiv.org/abs/2410.21228 Although LoRA restricts weight updates to a low-dimensional subspace, which should reduce forgetting and be better suited for continual learning, this paper shows it may not actually be the case. #MLSky

Launching the Physical AI AV Dataset! 🚀 huggingface.co/datasets/nvi... One of the largest, most diverse & commercially usable open-source datasets for AVs. - 1727 hours of driving data - Camera, LiDAR, & radar - 25 countries, 2500+ cities This is just the beginning, more features to come!

There’s likely to be a wave of X users coming here in the future. These are people who have finally found some decision sufficiently gross to leave. Maybe this time lets try not going with ”get off this site/get back in your locker”

Proud colleague! 🎉

Muon is a (relatively) new optimizer that powered large-scale training of recent foundation models, e.g., Kimi K2 and GLM 4.5. Interested in learning how it works? Check out the video here: youtu.be/bO5nvE289ec

FineVision They unified 200+ data sources into 24M samples, which consists of 17.3M images and 9.5B answer tokens. Project: huggingface.co/spaces/Huggi... Paper: arxiv.org/abs/2510.17269 Dataset: huggingface.co/datasets/Hug...

This marks the end of this year's #ICCV2025! We wish you all the best and hope you had an amazing conference, full of intellectually stimulating conversations and deep insights. We love this community and have enjoyed reporting from the ground (and virtually). Aloha! 🌺🌸

EuroHPC wrote a piece about our love of GPUs 😊 A EuroHPC Success Story | Clear Vision for Self-Driving Cars www.eurohpc-ju.europa.eu/eurohpc-succ...

See you at 10:45 today! 🌺

23rd of October. @iccv.bsky.social We present our work on “Large-scale Pretraining for Grounded Video Caption Generation” with Cordelia Schmid and @josef-sivic.bsky.social at Exhibit Hall I #434 on the morning session. We’ll have a search demo for our dataset as well! See you there!! 🚀

Congratulations 👏 to our Principal Investigators Michael @michael-j-black.bsky.social, Gerard @ponsmoll.bsky.social and their SMPL Body Model Team for securing the Everingham Prize at #ICCV2025! @maxplanckcampus.bsky.social

We want to see your #ICCV2025 experience! 📸 Capture your favorite moments—the groundbreaking talks, the innovative posters, the networking, and the fun—and post your pictures! Don't forget to tag us @iccv.bsky.social so we can share the excitement! 🎉

Thank you for organising this @sarameghanbeery.bsky.social ! It was great!! 🤩

Badass birders braving the rain this morning!! We got soaked, saw cool birds, and got a beautiful rainbow as a reward 🌈

ICCV 2025 kicks off tomorrow! We look forward to welcoming everyone to Hawaii 🌺

🚗 Join us for the RealADSim Workshop @ ICCV 2025! 🗓 Oct 19 morning | 📍 Hawaii Conv. Center, 305A Featuring exciting talks and winner presentations from our challenges (70+ regs, ~30 subs). Don’t miss it! 🎉 #ICCV2025

See you in Honolulu y’all! #ICCV2025

Does anyone have a head mount for GoPro with them in Honolulu that could borrow it to me during ICCV? #iccv2025

📣 📣 Postdoc position in #Robotics and #AI at the Department of Engineering, King's College London! www.kcl.ac.uk/jobs/128205-...

I’m definitely going! #iccv2025

Are you coming to @iccv.bsky.social?? Do you like BIRDS?!?! The #CVBirdWalk is happening!!! Oct. 21, 7am, Ala Moana Park. Kick off your conference right!! #ICCV2025 #AIforConservation

Our recent research will be presented at @iccv.bsky.social! #ICCV2025 We’ll present 5 papers about: 💡 self-supervised & representation learning 🌍 3D occupancy & multi-sensor perception 🧩 open-vocabulary segmentation 🧠 multimodal LLMs & explainability valeoai.github.io/posts/iccv-2...

I’m in paradise

Thank you all who shared your open-sourced #ICCV2025 works! I wrote and published the article at Medium medium.com/@yoshitomo-m... HF🤗 huggingface.co/blog/yoshito... Did you miss my previous post? No worries, I will keep the Google Form open until next week and try to update the article

& “Online Reasoning Video Segmentation with Just-in-Time Digital Twins” by @mathias-unberath.bsky.social & more introduces a just-in-time digital twin where an LLM plans a low-level scene representation from high-level video using specialist vision models: arxiv.org/abs/2503.21056

Wufei Ma, Guofeng Zhang, Yu-Cheng Chou, Jieneng Chen, Alan Yuille, @celso-de-melo.bsky.social, and more present the first comprehensive 3D spatial reasoning benchmark in “3DSRBench: A Comprehensive 3D Spatial Reasoning Benchmark”: 3dsrbench.github.io (6/10)

In “VideoAuteur: Towards Long Narrative Video Generation—A Case Study in How-to-Cook Videos,” Junfei Xiao, Alan Yuille, & researchers from @bytedance.bsky.social present a large-scale cooking 🥘 video dataset designed to advance long-form narrative generation: arxiv.org/abs/2501.06173 (5/10)

Interested in doing a PhD in machine learning at the University of Edinburgh starting Sept 2026? My group works on topics in vision, machine learning, and AI for climate. For more information and details on how to get in touch, please check out my website: homepages.inf.ed.ac.uk/omacaod

💡 The new European HPC Application Support Portal is live! Developed under the @epicurehpc.bsky.social, it offers one gateway to support across EuroHPC systems, including LUMI 💪 Simpler access, guidance, and training — all in one place! 🔗https://lumi-supercomputer.eu/epicure-portal-launched/