Sign in

Suresh Venkatasubramanian

@geomblog.bsky.social
4.7K followers 812 following 1.3K posts

Director, Center for Tech Responsibility@Brown. FAccT OG. AI Bill of Rights coauthor. Former tech advisor to President Biden @WHOSTP. He/him/his. Posts my own.

PostsRepliesMedia
Suresh Venkatasubramanian @geomblog.bsky.social · 44m
Now this is a red baseball cap I could work with
010
Reposted by Suresh Venkatasubramanian
EmilyTav @emilytav.bsky.social · 59m
The entire language of an “AI race” is flawed. A race has a finish line. Software doesn’t have a finish line… it’s never done.
1132
Reposted by Suresh Venkatasubramanian
Ami Fields-Meyer @fieldsmeyer.com · 2h
Four years since the AI Bill of Rights under @alondra.bsky.social , 24 hours since Trump announced the AI companies would do "tremendous self-policing." Me:
Red cap with "ALONRA WAS RIGHT ABOUT EVERYTHING"
0143
Suresh Venkatasubramanian @geomblog.bsky.social · 6h
I indeed went there and called SI "super silly". www.cnbc.com/video/2026/0...
cnbc.com
Trump's AI pact is voluntary and unenforceable: Former White House tech adviser
Brown University's Suresh Venkatasubramanian, a former Assistant Director at the White House Office of Science and Tech Policy says AI policy should focus on protecting people from current, tangible h...
192
Reposted by Suresh Venkatasubramanian
Melanie Mitchell @melaniemitchell.bsky.social · 30/09/2026
Get it while it's still cheap!
2344
Reposted by Suresh Venkatasubramanian
angela zhou @angelamczhou.bsky.social · 29/09/2026
Anyway, I feel super alive about tiny, tiny smol models
0101
Reposted by Suresh Venkatasubramanian
angela zhou @angelamczhou.bsky.social · 29/09/2026
open source alternatives to Jev are super fun because it's basically just variants of last-layer retraining with an updated pretrained representation that's better at language
112
Reposted by Suresh Venkatasubramanian
Data & Society @datasociety.bsky.social · 29/09/2026
Speaking to @politico.com, D&S board member @geomblog.bsky.social has some advice for lawmakers: "Stop worrying about superintelligence, and start making companies pay for putting out unreliable and error-prone systems that could and have harmed people." www.politico.com/newsletters/...
politico.com
5 Questions for Suresh Venkatasubramanian
0153
Suresh Venkatasubramanian @geomblog.bsky.social · 29/09/2026
I just want to say that a ban on recursion would make algorithms classes way more interesting. Khanna to introduce AI safety bill awaiting technology safeguards share.google/Pj9lsr5Jt3lJ...
share.google
Khanna to introduce AI safety bill with ban on 'recursive' technology until safeguards exist
The proposal joins a chorus of other measures being put forward in Congress as the body mulls how to regulate AI.
050
Reposted by Suresh Venkatasubramanian
David Brody @dbrody.bsky.social · 28/09/2026
Yes, and, the Markey/Clarke AI Civil Rights Act is several years in the making with over 80 endorsements from many major civil rights, consumer, labor, and civil liberties orgs. It is the consensus gold standard bill for the progressive plan to regulate AI.
021
Suresh Venkatasubramanian @geomblog.bsky.social · 28/09/2026
This is exactly right.
092
Reposted by Suresh Venkatasubramanian
Alondra Nelson @alondra.bsky.social · 27/09/2026
The idea that people want AI oversight to happen "all of a sudden" ignores years of research, warnings, and policy efforts. The delay hasn’t been a lack of awareness; it’s been because of years of industry lobbying against regulation, political resistance, and inaction.
37932245
Reposted by Suresh Venkatasubramanian
Mrs. Betty Bowers @mrsbettybowers.bsky.social · 26/09/2026
The major platforms—Facebook, YouTube, and X—are all refusing to show the trailer for the upcoming documentary that eviscerates the ridiculous pest Elon Musk. So I guess we have to do our part to make sure it is seen. www.youtube.com/watch?v=eY_8...
youtube.com
Musk | Official Trailer | Bleecker Street
YouTube video by Musk
23858683994
Reposted by Suresh Venkatasubramanian
dave.linkblog.org @dave.linkblog.org · 25/09/2026
US being left behind in EV charging speeds as China goes sub-5 min to 70%. arstechnica.com
033
Reposted by Suresh Venkatasubramanian
Emily M. Bender @emilymbender.bsky.social · 26/09/2026
When Coxon was suddenly ALL OVER the media, it felt to many like that was because his story deserved that much interest i.e. that the media was doing its job informing the public of an important threat. The wall-to-wall media coverage wasn't organic though, it was orchestrated: >>
piratewires.com
617981
Suresh Venkatasubramanian @geomblog.bsky.social · 25/09/2026
I answered 5 questions for @politico.com www.politico.com/newsletters/...
politico.com
5 Questions for Suresh Venkatasubramanian
0112
Reposted by Suresh Venkatasubramanian
Nathan Lambert @natolambert.bsky.social · 25/09/2026
Hearing AI regulation is in the air, especially with a misguided tone of “open dangerous, closed safe” so it felt timely to re-up this one. There are ways to make the AI industry safer without kneecapping transparency, education, and competition. www.interconnects.ai/p/banning-op...
interconnects.ai
Banning Open Source AI Would Be A Mistake
This post was originally an op-ed co-authored with Kevin Xu of Interconnected for a general, non-technical audience.
4308
Reposted by Suresh Venkatasubramanian
Tressie McMillan Cottom @tressiemcphd.bsky.social · 24/09/2026
Most people have no idea how much international students are the lifeblood of higher education. But they’re about to learn. Sigh.
322234585
Reposted by Suresh Venkatasubramanian
Timnit Gebru @timnitgebru.blacksky.app · 24/09/2026
Democracy Now! has now gone from constantly inviting Max Tegmark to platforming people from none other than MIRI. That's right MIRI. Y'all know how bad that is? Do you understand the level to which they are doubling down on the eugenicists?
11287100
Reposted by Suresh Venkatasubramanian
danah boyd @zephoria.bsky.social · 24/09/2026
Today is Pub Day for "Data Are Made, Not Found: A Story of Politics, Power, and the Civil Servants Who Saved the US Census"!!! ::bounce:: Please support me by buying a copy of the book or asking your library for a copy! press.uchicago.edu/ucp/books/bo...
1238795
Suresh Venkatasubramanian @geomblog.bsky.social · 23/09/2026
The group should have instead said that they used an agent swarm to hack the FBI. They might have instead gotten an invitation to speak at the UN Security Council.
010
Reposted by Suresh Venkatasubramanian
Serena Booth @reniebird.bsky.social · 23/09/2026
www.lesswrong.com/posts/HsijSh... @bradknox.bsky.social, Brian Christian, and I wrote a thing. We argue that an unexamined cause of the Open AI / Hugging Face Attack was the Exploit Gym evaluation metric. We post this on Less Wrong as a plea to practitioners to design better metrics in the future.
lesswrong.com
An unexamined cause of the OpenAI Hugging Face hacking incident: its binary performance metric — LessWrong
We argue that a main cause of the OpenAI Hugging Face incident was overlooked: the overly simple evaluation metric in ExploitGym was misaligned. Furt…
2259
Reposted by Suresh Venkatasubramanian
Ankit Panda @nktpnd.bsky.social · 22/09/2026
New in @foreignaffairs.com with @claroche.bsky.social: even if efforts to control the bomb may be a source of inspiration, they are not a ready-made template for AI governance. On the limits of the AI-nuclear weapons analogy. Gift link: www.foreignaffairs.com/guest-pass/r...
foreignaffairs.com
The Trouble With the Nuclear-AI Analogy
In humanity’s quest to constrain the dangers from fast-advancing artificial intelligence, many have looked to the most revolutionary technology that preceded it: nuclear weapons.
1127
Suresh Venkatasubramanian @geomblog.bsky.social · 22/09/2026
Personally, I use supercalifragilistic intelligence
081
Reposted by Suresh Venkatasubramanian
Bob Kopp @bobkopp.net · 22/09/2026
NEW PAPER: In a new paper, jointly published in @agu.org Earth's Future & @ametsoc.org BAMS, and co-led by Melissa Kenney, @costasamaras.com and me, 175 authors outline priorities to guide the future of scientific assessments of climate change in the United States. 🧵
doi.org
Research Priorities for Robust Climate Assessments in the United States
A network of climate assessments plays a critical role in synthesizing the state of science to support evidence-based decision-making To support future U.S. and sub-national climate assessments, ...
114782
Suresh Venkatasubramanian @geomblog.bsky.social · 21/09/2026
This is a really good and sober assessment of "sexy apocalypse" doomer nonsense.
1204
Reposted by Suresh Venkatasubramanian
Alondra Nelson @alondra.bsky.social · 20/09/2026
I've read this bracing piece 2x this morning and know I will come back to it: "Najya’s teacher had most likely never performed the school’s mandatory reading assessments for Najya or any of her classmates but instead had assumed the flagging competency of these mostly low-income Black children."
49132
Reposted by Suresh Venkatasubramanian
Izabella Laba 🇨🇦 @ijlaba.bsky.social · 20/09/2026
If your dog bites a person, you don't get to say "oh whatever, the dog just did it on its own". You are responsible.
0146
Reposted by Suresh Venkatasubramanian
Alondra Nelson @alondra.bsky.social · 19/09/2026
44 is right.
1537
Reposted by Suresh Venkatasubramanian
Clément Canonne @ccanonne.github.io · 19/09/2026
"The proof [...] was obtained in a conversation with ChatGPT-6 Astra on Tuesday, September 15, 2026 at 1:02 AM PDT. We then prepared this manuscript for public release, with the intent of uploading it on the morning of Thursday, September 17, 2026" Shameless, and shameful. arxiv.org/abs/2609.207...
arxiv.org
The Strong Secretary Conjecture is True for Linear Matroids
We prove a $1/e$ guarantee for the matroid secretary problem on linear matroids, therefore settling the strong secretary conjecture in this class of matroids. The result holds both when the matroid is...
1459
Reposted by Suresh Venkatasubramanian
Andy Diggle @andydiggle.bsky.social · 18/09/2026
Hallucinating AI "almost started a war" by FALSELY claiming China was smuggling nuclear materials. "The US military swung into action with plans to intercept the vessel... Military planes were in the air"
edition.cnn.com
Exclusive: US military had close call after using AI for false intelligence report, sources say | CNN Politics
The episode shows the risks of using this new, relatively poorly understood technology in the middle of the Iran war
17190115
Reposted by Suresh Venkatasubramanian
Sarah T Roberts, PhD @ubiquity75.xyz · 18/09/2026
From Science: Free inquiry depends on freedom of movement Starting in May 2025, the US government implemented an immigration policy that targeted noncitizen researchers for visa denials, visa revocations, detention, and deportation based on the topics and findings of their work (1).
science.org
Free inquiry depends on freedom of movement
Starting in May 2025, the US government implemented an immigration policy that targeted noncitizen researchers for visa denials, visa revocations, detention, and deportation on the basis of the topics...
13321
Reposted by Suresh Venkatasubramanian
Critical AI https://read.dukeupress.edu/critical-ai/issue @criticalai-journal.bsky.social · 18/09/2026
We are not in the habit of linking to "the other place" but @banteg.xyz so far as I know has not posted this abso BRILLIANT send-up of EA logic on bluesky. Till then: x.com/banteg/statu...
x.com
012
Reposted by Suresh Venkatasubramanian
Gautam Kamath @gautamkamath.com · 16/09/2026
Nihar Shah did a heroic experiment for TMLR: he spent 20-25 hours over two weeks interviewing authors of seemingly low-quality submissions about their own papers. He confirmed what we all suspected: people submitting these papers have *no idea* what is going on in them.
5276110
Reposted by Suresh Venkatasubramanian
Ryan Calo @rcalo.bsky.social · 16/09/2026
Banning "superintelligent" 🤢 AI plays right into industry rhetoric and keeps the focus off immediate, real world harms such as privacy, bias, and labor. Literally the same name of the silly Bostrom book everyone was reading 12 years ago.
311733
Suresh Venkatasubramanian @geomblog.bsky.social · 16/09/2026
This is an eye opening experiment. And kudos to Nihar for doing this, and writing about it carefully.
040
Reposted by Suresh Venkatasubramanian
Gigi Sohn @gbsohn.bsky.social · 16/09/2026
Extremely thrilled to have @amac0.bsky.social joining me and my colleagues on the @eff.org Board! Alex's combination of industry, government and non-profit experience will be a huge benefit to the organization as it writes its next chapter. www.eff.org/deeplinks/20...
eff.org
EFF Welcomes Alexander Macgillivray to its Board of Directors
Macgillivray Brings Perspectives on AI, Policy-Making, Social Justice, and Trust and Safety to the Board of Directors
13512
Reposted by Suresh Venkatasubramanian
River_Tam @rivertam.bsky.social · 16/09/2026
"Whatever our goals are, we’ve operationalized them primarily through proving theorems. Almost all papers or PhD theses have a main theorem, and ostensibly a proof of it. But it should be clear that the goal of mathematics is not to prove theorems..." proofsandprompts.com/2026/09/14/a...
proofsandprompts.com
A beginning for mathematics
Daniel Litt, professor at the University of Toronto Three years ago, AI systems could not reliably add two numbers. A year ago, internal models at OpenAI and DeepMind received the equivalent of a g…
1103
Reposted by Suresh Venkatasubramanian
ACLU @aclu.org · 16/09/2026
Good. This bill contained provisions that would've endangered our rights for the sake of AI experimentation. As tech leaders warn of AI's risks, Congress must not remove the few protections we have to hold companies accountable.
politico.com
‘The bill is dead’: Crypto legislation collapses on Senate floor
Democrats said the legislation didn’t do enough to crack down on President Donald Trump’s potential conflicts of interest.
2498114
Reposted by Suresh Venkatasubramanian
Kate Starbird @katestarbird.bsky.social · 16/09/2026
So depressing. My PhD studies were funded by an NSF Graduate Research Fellowship. My research program was pretty much exclusively funded by the NSF between 2013 and 2020, and then 60% NSF until 2025. Future university researchers won’t have this opportunity.
212439
Reposted by Suresh Venkatasubramanian
Erinn O'Dear @erinnthered.bsky.social · 15/09/2026
So, there once was this Harry Potter fanfiction...
0449
Reposted by Suresh Venkatasubramanian
Ryan O'Donnell @booleananalysis.bsky.social · 14/09/2026
Outstanding progress towards the Unique Games Conjecture posted by Yumou Fei, Dor Minzer, and Shuo Wang: eccc.weizmann.ac.il/report/2026/... 👀
eccc.weizmann.ac.il
ECCC - TR26-179
3237
Reposted by Suresh Venkatasubramanian
Nutan Limaye @nutanlimaye.bsky.social · 14/09/2026
This is not where I expected TCS to be in 2026! The last line in the abstract is distressing. :( Is this how we need to work when we get a big result? As if someone is always chasing us?
0412
Suresh Venkatasubramanian @geomblog.bsky.social · 14/09/2026
I think the chance of what he's describing is closer to 15% than 10%. I'm stocking up on silver stakes.
020
Reposted by Suresh Venkatasubramanian
Seth Rockman @sethrockman.bsky.social · 14/09/2026
This is a big one: Brown University is searching for the next director of the Simmons Center for the Study of Slavery and Justice. Please help us identify excellent candidates to lead this signature program into the future. Please write me with ideas; please share in your networks; please apply.
networks.h-net.org
Brown University - Professor and Director of the Simmons Center for the Study of Slavery and Justice | H-Net
1. Academic Accomplishment. We seek a scholar with a strong scholarly record, including excellence in both research and teaching, appropriate for appointment at the level of full professor in a social science, humanities, or arts department at Brown.
2275234
Reposted by Suresh Venkatasubramanian
dame @dame.is · 14/09/2026
did someone say pace the frontier?
Pace The Frontier salsa jar with green lid. Glass bottle containing red chunky salsa with visible pepper and vegetable pieces. Label features "Pace" in red script with a yellow sun logo, "THE FRONTIER" text, and images of red peppers and jalapeños on green banner against white background.
1715
Reposted by Suresh Venkatasubramanian
Mike Masnick @masnick.com · 14/09/2026
One thing not getting as much attention re: Dario's essay. In it, he talks about the need for third party evaluators to review models for safety. But a few days before the essay, Anthropic... refused to submit its latest models to the UK's evaluator: www.ft.com/content/560e...
ft.com
Anthropic withheld latest AI model from UK testing agency
Decision to exclude Aisi has prompted fears inside British government of protectionist shift among tech groups
4384106
Reposted by Suresh Venkatasubramanian
Timnit Gebru @timnitgebru.blacksky.app · 14/09/2026
The only person in public service who has consistently made 100% sense on issues of "AI," which is why they hate her so much.
Law enforcers already have authority to charge companies and their CEOs for creating and releasing dangerous, unvetted, or defective products. We shouldn’t let discussions about new legal regimes distract from the fact that there’s no AI exemption from laws already on the books — a point 
@FTC
 emphasized repeatedly during my tenure.

1. There is an extensive set of laws that govern dangerous and defective products. For example, releasing unvetted AI models or agents can violate consumer protection laws. Shipping flawed AI tools without implementing adequate measures to detect and stop rogue or defective AI agents can be an “unfair or deceptive” act or practice under the FTC Act (and analogous state laws). And some state AGs are already exploring holding AI firms and their CEOs criminally liable when their models participate in criminal activity.

2. Existing laws also prohibit “unfair methods of competition.” This covers instances where AI firms appropriate the competitively sensitive information of their customers, including through tracking their use of various tools. It can also cover instances where firms pursue dangerous behavior, aware that doing so may compel rivals to do the same.As the Supreme Court has noted: “A method of competition which casts upon one's competitors the burden of the loss of business unless they will descend to a practice which they are under a powerful moral compulsion not to adopt, even though it is not criminal, was thought to involve the kind of unfairness at which the [unfair methods of competition] statute was aimed."

3. The highly concentrated and interconnected structure of these markets could be creating major risks and conflicts of interest. We had started investigating these partnerships and cross-investments across the stack (and released a preliminarily overview of some findings: https://ftc.gov/news-events/news/press-releases/2025/01/ftc-issues-staff-report-ai-partnerships-investments-study). 

Both federal and state enforcers should be scrutinizing these opaque relationships and inter-dependencies. We are already seeing how these relationships could undermine accountability. For example, OpenAI could face liability given the Hugging Face incident, but Hugging Face being bought up by Nvidia means that we’re unlikely to see it file a lawsuit over this — given Nvidia’s strong incentive to see OpenAI continue full speed ahead.4. As AI tools dramatically change the landscape of cybersecurity risks and hacks, all businesses should be doubling down on having core security protections in place. Firms that fail to invest in adequate data security measures or fix known vulnerabilities can also be breaking the law. A recent analysis showed that around 1/3 of Fortune 100 companies do not even have a way to notify them about security issues. During my 
@FTC
 tenure, we sued firms for poor data security practices and held CEOs liable when they were personally responsible.

https://this.weekinsecurity.com/dozens-of-americas-largest-companies-have-no-simple-way-to-report-security-flaws/

https://ftc.gov/news-events/news/press-releases/2022/10/ftc-takes-action-against-drizly-its-ceo-james-cory-rellas-security-failures-exposed-data-25-million

5. As policymakers consider new legal regimes, we should be looking to lessons from prior efforts to govern major sectors, such as banking and other networks, platforms, and utilities. Tools like structural separations, nondiscrimination, and supervision could be key, and there’s a rich history of what works and what doesn’t. But we can and must pursue any new efforts alongside enforcing existing laws.
192029659
Reposted by Suresh Venkatasubramanian
Gautam Kamath @gautamkamath.com · 13/09/2026
In case you need a positive AI story right now: I have a baby monitor which automatically takes pictures and videos. Downloading is too slow: you have to manually click each one (infeasible for ~9k pics over 3 yrs). I had Codex write a script to download all of them in 45 mins.
1504