Well played, @macalestercollege.bsky.social, well played. The eligible teenager in my household was definitely won over.

For instructors out there: a very cool set of AI Course Policy Icons by Cornell's GenAI taskforce, to be used in combination for your syllabi or assignments. Inspired by @creativecommons.bsky.social and available w/ a CC license. Sample icons attached here. teaching.cornell.edu/generative-a...

"very upset and angry" is correct. The government is breaking another (implicit) contract. Science and society lose again.

At @hcde.uw.edu, we will be hiring an Assistant or Associate Teaching Professor this coming year! Timing wise, we are hiring for a Sept. 2026 start date, but could be sooner. We are looking forward to welcoming a new colleague! Learn more and apply at apply.interfolio.com/171881

They're founding a new organization - Stand Together for Higher Ed - which is organizing a national webinar on reclaiming the narrative on October 29. For more on the org and the webinar - www.standtogetherhighered.org

Excited for this event at Yale in a few weeks -- lucky to have a speaking slot in which I (in peak academia) will question* the very premise of the event by saying that we just don't have, or use, the right signals for our social and information systems. *"Problematize", I guess

If your community is thinking about Community Notes, we have some notes to guide your Community Notes thinking right here. #CommunityNotes

AI Doomerism is selling on the New York City subway. I guess the subway *is* unsafe

Pissing from the springboard, volume 644,473 www.nytimes.com/2025/09/20/u...

"They had a choice, and they chose tyranny." My colleague @jtlg.bsky.social is indeed a lawyer but he keeps it plain, short, and simple here. 3d.laboratorium.net/2025-09-20-d...

More illustrative version here -- two time ranges, and with "Cancel Netflix" as control, plus showing that the Disney backlash also trickled sideways to Hulu cancellations. #canceldisney+

Personalist regime tries to set up another mechanism of control. Silicon Valley firms will be subject to the whims of the mad king.

Exactly two months ago. When will we go to the streets? Maybe when Jon Stewart is out.

Great/important event at Cornell Tech Oct 9th and (remote) Oct 10th. Spread the word! "The Disability and Access in Tech and AI Summit is our annual conference to discuss issues around accessibility, including research, identity, and policy." tech.cornell.edu/cornell-tech...

Amy has been thinking about how to innovate (i.e. save) peer review for several years. This would be a huge step forward if adopted.

Inspired by Fred Rogers, @nature.com asked readers to help us 'find the helpers' amid the crisis in American science. What emerged were inspiring stories of researchers, legal experts, data hounds and government scientists who are fighting the dismantling of the US research enterprise. 🧪🧵1/6

We are hiring tenure track faculty this year: apply.interfolio.com/171020

Simpering cowards one and all. This is almost unbearable to watch.

Hard to find a more worthy organization and mission.

The numbers and data about the long-term exchange of the teen with the chatbot are quite shocking for anyone (like me) who only saw the headlines.

1. The One I Love is the perfect pop song. 2. This viral thread is optimally BSky. Hello 40+50 yo dads!

The numbers and data about the long-term exchange of the teen with the chatbot are quite shocking for anyone (like me) who only saw the headlines.

IHE / Generation Lab survey of 1047 college students reports that 85% used GenAI for coursework last year. But, for whatever reason, there is surprisingly little cynicism about the value of college. www.insidehighered.com/news/student...

In our 2023 paper, @mjakesch.bsky.social showed how AI systems can learn "to produce text perceived as more human than human" by human evaluators. In the thread below, a suggestion that models (GPT-5) are optimizing (maybe side-product?) to fool other AI models. www.pnas.org/doi/10.1073/...

The best people. tech.cornell.edu/news/cornell...

The Computation+Journalism Symposium will take place on December 11-12 at the University of Miami. Join us: cplusj2025.com The deadline to submit a talk, workshop, or panel is September 7th: cplusj2025.com/submissions/ We've already announced several speakers. More coming soon!

NYC watch: JPMorgan Chase Tower Opens to First Employees at 270 Park Avenue -- and I think celebrated with a LED rendition of... Galaga?! (Sorry for the low quality video, zoom from my window). About the building: newyorkyimby.com/2025/08/jpmo...

A clear-eyed view that is at once hopeful and depressing. "Contrary to much popular opinion, college is not in the information transfer business; we are in the identity formation business."

2 bits of data from new Economist/YouGov poll..... 13% support for cutting research funding to universities. 24% among Republicans. Polling on this continues to be catastrophic for the Rs, suggests Dems should be learning far harder into standing up for science and our universities. 1/

...in which Terry Gross hears about our work from @vauhinivara.bsky.social! (study led by @mjakesch.bsky.social, recent follow-up by @sterlingwc.bsky.social) www.npr.org/2025/07/23/g...

What kind of AI governance do we need? Our new piece in @science.org answers this: we need policy grounded in evidence and built to generate more of it. Evidence-based policymaking is not a slogan—it’s a design challenge for democratic governance in the age of AI www.science.org/doi/10.1126/... 🧵

This is too cool. Reminds me of what we did in the first years of geotagged Flickr (raise your hand if you remember) photos although we had to rely on text and tagging. No more!

The university presidents in Israel speak out against the atrocities in Gaza. The Israeli public has largely been shielded from the situation in Gaza by the cooperating media but cracks (like this one) are appearing. Netanyahu's criminal and shameful government must and will be defeated.

Can AI (GPT-4o-mini) #TrustSafety guardrails be easily circumvented via age-old human mechanisms of persuasion? Yes, in this extremely elegant experiment by @emollick.bsky.social‬ & team (incl. Cialdini aka Mr. Persuasion). Also: nicely presented on this page! gail.wharton.upenn.edu/research-and...

In targeting “ideological bias” and “social engineering agendas," the Trump administration’s AI Action Plan ultimately enforces them, writes Tech Policy Press fellow Eryk Salvaggio.

Months before ChatGPT, for my keynote talk at @recsys.bsky.social I suggested we may be soon falling off the Internet cliff. I used this Einstein(?) quote to capture the yet-unknown impact of AI on our information ecosystem and argue that the future of human communication may be face-to-face.

Free speech (but don't mention _these_ topics) and no bias (but the government will decide what's biased). Oh and these tools should reflect truth, and be trustworthy, and avoid social engineering, all of which I imagine the government will help us identify in a fair and robust manner.

Don't @ me, finally with justification

The Colbert Clock expired. Kimmel next or Meyers*? Strike Force Five vs totalitarianism. Shameful. (Or Jon Stewart).

I saw this "financial decision" coming. Who is next? www.nytimes.com/2025/07/17/b...

This @nytimes.com story is extremely well reported. But I don't understand why it potrays Netanyahu as enjoying a comeback. All the data I saw shows he made only minor gains in public opinion and polls in last month, certainly not the grand bounce depicted here. www.nytimes.com/2025/07/11/m...

This @nytimes.com story is extremely well reported. But I don't understand why it potrays Netanyahu as enjoying a comeback. All the data I saw shows he made only minor gains in public opinion and polls in last month, certainly not the grand bounce depicted here. www.nytimes.com/2025/07/11/m...

Forget El0n. I am curious who are the cowards working to develop this chatbot. He has no shame. Do they* have any? *Sadly it is possible that they do have shame but do not have a green card / employment options.

Though much remains uncertain (e.g. outcomes of lawsuits re: canceled grants), it is clear that U.S. gov funding for science — especially science that intersects w/ concerns about human flourishing — is going to be significantly diminished in coming years. Here are a few things we can do to help:

Excited to have this work out at ICML this year! Do LLMs make correlated errors? Yes, and those by the same company, and also more accurate/later generations are more correlated -- increasing algorithmic monoculture arxiv.org/abs/2506.07962

The skyline is red tonight. An urgent warning to all of us. Oh wait it's Canada Day

#CornellTech is honored to march in the 2025 NYC Pride Parade, celebrating love, identity & equality. 🏳️‍🌈💡 From Roosevelt Island to Manhattan, our message is clear: inclusion inspires innovation. Thanks to all who showed up with passion + pride! #NYCPride2025 #TechForAll

In a @newyorker.com feature, Professor @informor.bsky.social weighs in on how #AI tools may be reshaping culture. Naaman says these tools “work covertly, sometimes very powerfully, to change not only what you write but what you think.” @cornellbowers.bsky.social www.newyorker.com/culture/infi...

Do you teach (or aspire to teach) Trust & Safety in a higher ed institution? Please help me out by responding to this survey about the T&S Teaching Consortium: docs.google.com/forms/d/e/1F...