Eugene Vinitsky 🍒
@eugenevinitsky.bsky.social
Anti-cynic. Towards a weirder future. Reinforcement Learning, Autonomous Vehicles, transportation systems, the works. Asst. Prof at NYU
https://emerge-lab.github.io
https://www.admonymous.co/eugenevinitsky
https://emerge-lab.github.io
https://www.admonymous.co/eugenevinitsky
For those at NeurIPS, Percepta, NVIDIA and GC are hosting an event.
Join the wait list:
t.co/arIJkVvs7h
Orrrrr solve some puzzles for a chance to skip ahead:
t.co/L4tjbr1qNi
Join the wait list:
t.co/arIJkVvs7h
Orrrrr solve some puzzles for a chance to skip ahead:
t.co/L4tjbr1qNi
https://percepta-neurips.web.app/
t.co
November 28, 2025 at 8:26 PM
For those at NeurIPS, Percepta, NVIDIA and GC are hosting an event.
Join the wait list:
t.co/arIJkVvs7h
Orrrrr solve some puzzles for a chance to skip ahead:
t.co/L4tjbr1qNi
Join the wait list:
t.co/arIJkVvs7h
Orrrrr solve some puzzles for a chance to skip ahead:
t.co/L4tjbr1qNi
Would people be cool with an LLM trained entirely on synthetic data and data sourced by consent?
November 27, 2025 at 8:34 PM
Would people be cool with an LLM trained entirely on synthetic data and data sourced by consent?
Reposted by Eugene Vinitsky 🍒
At the risk of starting the flame war to end all flame wars...
Modern LLMs (GPT-5.1, Claude 4.5, Gemini 3) produce excellent code and can be a significant productivity boost to software engineers who take the time to learn how to effectively apply them - especially if used with coding agent tools
Modern LLMs (GPT-5.1, Claude 4.5, Gemini 3) produce excellent code and can be a significant productivity boost to software engineers who take the time to learn how to effectively apply them - especially if used with coding agent tools
November 27, 2025 at 7:55 PM
At the risk of starting the flame war to end all flame wars...
Modern LLMs (GPT-5.1, Claude 4.5, Gemini 3) produce excellent code and can be a significant productivity boost to software engineers who take the time to learn how to effectively apply them - especially if used with coding agent tools
Modern LLMs (GPT-5.1, Claude 4.5, Gemini 3) produce excellent code and can be a significant productivity boost to software engineers who take the time to learn how to effectively apply them - especially if used with coding agent tools
Reposted by Eugene Vinitsky 🍒
*surveys discourse and taps sign*
and you can block! just block them! it’s ok to block!
and you can block! just block them! it’s ok to block!
someone quoted your post and you don't like it? you can "detach" your post so that it disappears and is no longer linked to their quote
posting something a little sensitive and worried about replies from random people? you can select who is allowed to reply to individual threads
posting something a little sensitive and worried about replies from random people? you can select who is allowed to reply to individual threads
November 27, 2025 at 1:50 PM
*surveys discourse and taps sign*
and you can block! just block them! it’s ok to block!
and you can block! just block them! it’s ok to block!
Videos and images are so close to being easily faked. Students are using LLMs to learn. Every online system is getting swamped with fake stuff. This is why “ML models can’t do anything” is such dangerous cope, there’s a bad default future unless we all collectively understand the challenge
November 27, 2025 at 8:26 AM
Videos and images are so close to being easily faked. Students are using LLMs to learn. Every online system is getting swamped with fake stuff. This is why “ML models can’t do anything” is such dangerous cope, there’s a bad default future unless we all collectively understand the challenge
Reposted by Eugene Vinitsky 🍒
At #NeurIPS2025? Join us for a Social on Wednesday at 7 PM, featuring a fireside chat with Jon Kleinberg and mentoring tables.
Ft. mentors @djfoster.bsky.social @surbhigoel.bsky.social @aifi.bsky.social @gautamkamath.com and more!
Ft. mentors @djfoster.bsky.social @surbhigoel.bsky.social @aifi.bsky.social @gautamkamath.com and more!
November 26, 2025 at 8:47 PM
At #NeurIPS2025? Join us for a Social on Wednesday at 7 PM, featuring a fireside chat with Jon Kleinberg and mentoring tables.
Ft. mentors @djfoster.bsky.social @surbhigoel.bsky.social @aifi.bsky.social @gautamkamath.com and more!
Ft. mentors @djfoster.bsky.social @surbhigoel.bsky.social @aifi.bsky.social @gautamkamath.com and more!
People don’t hate datacenters being built near them, they hate *anything* being built near them. The median position is never build anything
November 26, 2025 at 8:15 PM
People don’t hate datacenters being built near them, they hate *anything* being built near them. The median position is never build anything
It is totally reasonable and rational for artists and writers, folks whose livelihood is affected by GenAI, to be wildly opposed to it. But most of the interesting parts of ML have very little to do with automated writing or art and to make it all about this is reductive too
November 26, 2025 at 3:24 PM
It is totally reasonable and rational for artists and writers, folks whose livelihood is affected by GenAI, to be wildly opposed to it. But most of the interesting parts of ML have very little to do with automated writing or art and to make it all about this is reductive too
This site is a bunch of interesting folks but also the last hanger-ons to the idea that progress happens through snarky pile-ons. May they one day form a completely disconnected cluster
November 26, 2025 at 7:29 AM
This site is a bunch of interesting folks but also the last hanger-ons to the idea that progress happens through snarky pile-ons. May they one day form a completely disconnected cluster
Reposted by Eugene Vinitsky 🍒
🎉 New work: “Learning Massively Multitask World Models for Continuous Control”
We introduce MMBench: a 200-task RL benchmark, and Newt: a language-conditioned multitask world model trained with large-scale online RL.
www.nicklashansen.com/NewtWM/
Code, checkpoints, dataset etc. are open-source!
We introduce MMBench: a 200-task RL benchmark, and Newt: a language-conditioned multitask world model trained with large-scale online RL.
www.nicklashansen.com/NewtWM/
Code, checkpoints, dataset etc. are open-source!
November 25, 2025 at 7:47 PM
🎉 New work: “Learning Massively Multitask World Models for Continuous Control”
We introduce MMBench: a 200-task RL benchmark, and Newt: a language-conditioned multitask world model trained with large-scale online RL.
www.nicklashansen.com/NewtWM/
Code, checkpoints, dataset etc. are open-source!
We introduce MMBench: a 200-task RL benchmark, and Newt: a language-conditioned multitask world model trained with large-scale online RL.
www.nicklashansen.com/NewtWM/
Code, checkpoints, dataset etc. are open-source!
The median EA I meet has anomalous alignment between their actions and their value, often donating a very significant fraction of their income, volunteering for causes they care about, being vegetarian / vegan, etc.
November 25, 2025 at 7:30 PM
The median EA I meet has anomalous alignment between their actions and their value, often donating a very significant fraction of their income, volunteering for causes they care about, being vegetarian / vegan, etc.
Reposted by Eugene Vinitsky 🍒
@reniebird.bsky.social and I have just been appointed to co-Chair @TheOfficialACM's US Technology Policy Committee’s Subcommittee on AI and Algorithms. cs.brown.edu/news/2025/11...
Serena Booth And Suresh Venkatasubramanian Co-Chair ACM’s US Technology Policy Committee’s Subcommittee On AI And Algorithms
Brown CS faculty members Serena Booth and Suresh Venkatasubramanian have just been appointed to co-chair the AI and Algorithms Subcommittee, whose recent work includes responses to government RFIs, te...
cs.brown.edu
November 25, 2025 at 5:31 PM
@reniebird.bsky.social and I have just been appointed to co-Chair @TheOfficialACM's US Technology Policy Committee’s Subcommittee on AI and Algorithms. cs.brown.edu/news/2025/11...
Reposted by Eugene Vinitsky 🍒
TMLR (@tmlrorg.bsky.social) is now proud to support interactive HTML-based submissions, going "Beyond PDF" -- check it out!
Thanks to Paul Vicol (@paulvicol.bsky.social) for his tireless work on this new option, as well as the OpenReview team.
Thanks to Paul Vicol (@paulvicol.bsky.social) for his tireless work on this new option, as well as the OpenReview team.
🚀 Introducing TMLR Beyond PDF!
🎬 This is a new, HTML-based submission format for TMLR, that supports interactive figures and videos, along with the usual LaTeX and images.
🎉 Thanks to TMLR Editors in Chief: Hugo Larochelle, @gautamkamath.com, Naila Murray, Nihar B. Shah, and Laurent Charlin!
🎬 This is a new, HTML-based submission format for TMLR, that supports interactive figures and videos, along with the usual LaTeX and images.
🎉 Thanks to TMLR Editors in Chief: Hugo Larochelle, @gautamkamath.com, Naila Murray, Nihar B. Shah, and Laurent Charlin!
November 25, 2025 at 4:14 PM
TMLR (@tmlrorg.bsky.social) is now proud to support interactive HTML-based submissions, going "Beyond PDF" -- check it out!
Thanks to Paul Vicol (@paulvicol.bsky.social) for his tireless work on this new option, as well as the OpenReview team.
Thanks to Paul Vicol (@paulvicol.bsky.social) for his tireless work on this new option, as well as the OpenReview team.
What's frustrating about LLM text is it almost short circuits your brain. It's optimized to convince a tired rater and so it is maximally convincing when scanning
November 25, 2025 at 5:12 AM
What's frustrating about LLM text is it almost short circuits your brain. It's optimized to convince a tired rater and so it is maximally convincing when scanning
Farms are filling the water with nitrates. But wait, what if we could let them off the hook because there's an easier punching bag
How Oregon’s Data Center Boom Is Supercharging a Water Crisis
Amazon has come to the state’s eastern farmland, worsening a water pollution problem that’s been linked to cancer and miscarriages.
Rolling Stone reports in collaboration with @thefern.org: www.rollingstone.com/culture/cult...
Amazon has come to the state’s eastern farmland, worsening a water pollution problem that’s been linked to cancer and miscarriages.
Rolling Stone reports in collaboration with @thefern.org: www.rollingstone.com/culture/cult...
November 25, 2025 at 3:55 AM
Farms are filling the water with nitrates. But wait, what if we could let them off the hook because there's an easier punching bag
Reposted by Eugene Vinitsky 🍒
Google: We have the world's most powerful AI-enabled operating system and devices
Also Google: Do you want to reply "Thanks!" or "Will do" to this SMS from Nancy Pelosi?
Also Google: Do you want to reply "Thanks!" or "Will do" to this SMS from Nancy Pelosi?
November 25, 2025 at 12:15 AM
Google: We have the world's most powerful AI-enabled operating system and devices
Also Google: Do you want to reply "Thanks!" or "Will do" to this SMS from Nancy Pelosi?
Also Google: Do you want to reply "Thanks!" or "Will do" to this SMS from Nancy Pelosi?
Reposted by Eugene Vinitsky 🍒
Friends, there are several CS positions open at NYU; calling in my socio-technical researchers including HCI researchers
cs.nyu.edu/dynamic/jobs...
cs.nyu.edu/dynamic/jobs...
NYU Computer Science Department
cs.nyu.edu
November 24, 2025 at 8:32 PM
Friends, there are several CS positions open at NYU; calling in my socio-technical researchers including HCI researchers
cs.nyu.edu/dynamic/jobs...
cs.nyu.edu/dynamic/jobs...
My favorite medical snafu is when the insurer says I need to call the doctor and the doctor says I need to call the insurer. It's fun to have an opportunity to participate in the medical billing process as a middleman
November 24, 2025 at 8:28 PM
My favorite medical snafu is when the insurer says I need to call the doctor and the doctor says I need to call the insurer. It's fun to have an opportunity to participate in the medical billing process as a middleman
Curse of dimensionality is real even if it sometimes feels like it isn’t
A little puzzled why LLM RL works at all. People are solving a hard, sparse exploration problem using on-policy exploration with a number of rollouts that would barely work to solve cartpole.
November 24, 2025 at 7:38 PM
Curse of dimensionality is real even if it sometimes feels like it isn’t
A little puzzled why LLM RL works at all. People are solving a hard, sparse exploration problem using on-policy exploration with a number of rollouts that would barely work to solve cartpole.
November 24, 2025 at 7:20 PM
A little puzzled why LLM RL works at all. People are solving a hard, sparse exploration problem using on-policy exploration with a number of rollouts that would barely work to solve cartpole.
Reposted by Eugene Vinitsky 🍒
I'm recruiting my first group of students at TTIC! If you're interested, please apply by December 9th and mention my name in your application
Pursue a fully-funded and personalized PhD in #computerscience with world-class faculty as mentors, a 4:1 student-to-faculty ratio, and in the vibrant city of Chicago. Learn more and apply by Dec. 9 (NO FEE): buff.ly/Qw8njWB
November 24, 2025 at 5:58 PM
I'm recruiting my first group of students at TTIC! If you're interested, please apply by December 9th and mention my name in your application
A lot of journalistic writing seems to operate under the assumption that people are reading between the lines and this likely leads to accidental sanewashing
November 24, 2025 at 5:01 PM
A lot of journalistic writing seems to operate under the assumption that people are reading between the lines and this likely leads to accidental sanewashing
Me and many of my students will be at NeurIPS this year! This is my "hello long-time friends and prospective RL graduate students" broadcast to the world. If you're interested in making RL scalable, reach out
November 24, 2025 at 4:35 PM
Me and many of my students will be at NeurIPS this year! This is my "hello long-time friends and prospective RL graduate students" broadcast to the world. If you're interested in making RL scalable, reach out
Reposted by Eugene Vinitsky 🍒
Three exciting opportunities at
@msftresearch.bsky.social in NYC!!! 🎉
Internship w/ FATE: apply.careers.microsoft.com/careers/job?...
Internship w/ STAC on AI evaluation and measurement: apply.careers.microsoft.com/careers/job?...
Postdoc w/ FATE: apply.careers.microsoft.com/careers/job?...
@msftresearch.bsky.social in NYC!!! 🎉
Internship w/ FATE: apply.careers.microsoft.com/careers/job?...
Internship w/ STAC on AI evaluation and measurement: apply.careers.microsoft.com/careers/job?...
Postdoc w/ FATE: apply.careers.microsoft.com/careers/job?...
FATE internships: apply.careers.microsoft.com/careers/job?...
FATE postdocs: apply.careers.microsoft.com/careers/job?...
And internships with our close collaborators at STAC: apply.careers.microsoft.com/careers/job?...
FATE postdocs: apply.careers.microsoft.com/careers/job?...
And internships with our close collaborators at STAC: apply.careers.microsoft.com/careers/job?...
November 24, 2025 at 3:33 PM
Three exciting opportunities at
@msftresearch.bsky.social in NYC!!! 🎉
Internship w/ FATE: apply.careers.microsoft.com/careers/job?...
Internship w/ STAC on AI evaluation and measurement: apply.careers.microsoft.com/careers/job?...
Postdoc w/ FATE: apply.careers.microsoft.com/careers/job?...
@msftresearch.bsky.social in NYC!!! 🎉
Internship w/ FATE: apply.careers.microsoft.com/careers/job?...
Internship w/ STAC on AI evaluation and measurement: apply.careers.microsoft.com/careers/job?...
Postdoc w/ FATE: apply.careers.microsoft.com/careers/job?...