New bill will make it a crime to download DeepSeek in the U.S., punishable with up to 20 years in prison.

As if techbros haven’t damaged society enough…

There should be much more work on policy + technical mechanisms for avoiding extreme concentration of power among a few decision-makers at AI companies (or in governments, with control over AI companies). Open source may help but is far from a complete solution to the problem.

multi-generational Black new orleans historian here for the comments.. 1, stop telling us how to self-identify. we whatever we said we are. 2, yes this is a Black american city. but context is that new orleans had a fully established [Black] creole culture predating "america."⬇️

frontier ai today

Excited to have our AI research published in nature today. Proud of the Profluent team and the extensive final version available under open-access. OpenCRISPR is a milestone. It's the first successful demonstration of editing the human genome with a molecule fully designed

THREAD. A recent poll shows that people in the United States suffer from mass delusion about crime. The results are alarming for Democrats. It should be a massive scandal for mainstream news, and it's a pillar of the authoritarian zeitgeist.

THREAD. A subtle but pervasive propaganda technique is when the news adopts as the *actual* motive the *stated* motive of people in power.

I’m at #COLM2025 🇨🇦presenting “Hidden in Plain Sight: VLMs overlook their vision representations” as a Poster and Oral! Also honored to win Outstanding Paper here and Best Paper @ CVPR EVAL-FoMo 2! Come chat at poster 12 (Wed AM) about building perceptual representations! (1/3)

Main idea: VLMs underperform their visual encoders on vision tasks. We diagnose this failure + present evaluations for future investigations into visual understanding in VLMs. Paper📄: arxiv.org/abs/2506.08008 Blog🌐: hidden-plain-sight.github.io Code💻: github.com/stephanie-fu/h… (2/3)

Congratulations to BAIR Researchers from trevordarrell lab whose paper "Hidden in plain sight: VLMs overlook their visual representations," was awarded an Outstanding Paper Award at #COLM2025 this week in Montreal, Canada. Stephanie Fu tyler bonnen Devin Guillory

Did anyone else know that AI could surveil students and send automatic messages to police with no warning or protections. This child almost lost his life

Bill Madden Terrible things are happening outside.

I am recruiting Ph.D. students at UMD Department of Computer Science starting Fall 2026! I am looking for students in three broad areas: (1) Physics-integrated computer vision (2) VLMs with constraints (2) Dual-use AI policy We're ranked #3 in AI on CSrankings! Specific details in 🧵

Not an opportunity to miss!!

LMAOOOO this dataset has been the gold standard for a decade

starting fall 2026 i'll be an assistant professor at Penn 🥳 my lab will develop scalable models/theories of human behavior, focused on memory and perception currently recruiting PhD students in psychology, neuroscience, & computer science! reach out if you're interested😊

Neurips fomo is too strong!

Researchers consider themselves very successful if they win one test-of-time award (and one is more than enough). Ross Ross Girshick has been winning them nonstop over the past year: CVPR 2024, ICCV 2025, and now NeurIPS 2025, because winning just one was too easy for him!