The agriculture water issue is definitely real. Water rights and pricing are crazy
Thomas Dietterich
@tdietterich.bsky.social
Safe and robust AI/ML, computational sustainability. Former President AAAI and IMLS. Distinguished Professor Emeritus, Oregon State University. https://web.engr.oregonstate.edu/~tgd/
The Cornell Lab invites applications for Edward W. Rose Postdoctoral Fellowships. These competitive awards support innovative, independent research by early-career scholars of exceptional promise. Two to four awards are made in each cycle. academicjobsonline.org/ajo/jobs/32334
Cornell University, Lab of Ornithology
Job #AJO32334, WDR-00059934 Rose Postdoctoral Associate, Cornell University, Ithaca NY, Cornell Lab of Ornithology, Lab of Ornithology, Cornell University, Ithaca, New York, US
academicjobsonline.org
Today we add fresh arXiv HTML for 98.8% of articles with LaTeX sources, available immediately. Full summer reset. Rust rewrites rock!
I suspect that personal agents will not see a big uptake precisely because it is a form of programming that requires anticipating everything that could go wrong. And with agents, "everything" covers a lot.
Trump's corruption damages everything it touches. It is hard to believe that FIFA's World Cup could become more corrupt, but Trump always finds a way
Great post by Daphne Koller on linkedin: www.linkedin.com/posts/daphne...
#ai | Daphne Koller | 19 comments
Today is not only the 250th anniversary of the US, it's also my own anniversary as an immigrant: on July 4, 1989, I arrived with two suitcases and little else to start a PhD in Stanford University's C...
linkedin.com
Two businessmen meet. One says, I've got a truckload of vodka. The other says, I'll give you 300 dollars for it. They shake on the deal. Then they part ways — the first to try and find a truckload of vodka, the second to try and find $300. This is a joke about AI.
On energy, Trump and the Republicans have dragged the US back to the 20th century while the rest of the world creeps into the future www.energyinst.org/exploring-en...
I'm excited to share our latest work “Representation Learning Enables Scalable Multitask Deep Reinforcement Learning”. In this work, we revisit a fundamental question in reinforcement learning: What if representation learning was the key ingredient behind scalable RL? 1/🧵
Today #arXiv officially leaves Cornell to becomes an independent nonprofit org. Details: https://blog.arxiv.org/2026/06/30/arxivs-next-chapter/
**On July 1, 2026, arXiv will spin out from Cornell University, its home for the past 25 years, to become an independent nonprofit organization.** With this next phase in arXiv’s journey quickly approaching, you can read more about arXiv’s history and the decision to spin out from Cornell in this recent article in the Cornell Chronicle. arXiv has been discussing the opportunity to become an independent organization for a long time, and we’re excited for the new flexibility and opportunities that this change will bring. This step is an important part of arXiv’s story that will have a positive impact as we continue to grow and has required great effort from staff to transition smoothly. Thanks to this, in the coming months, arXiv authors, readers, and the community that relies on us will not notice much change at all. arXiv was built for scientists, by scientists, and while our home may be changing, our mission, vision, and values never will. arXiv will remain committed to being free to read and to submit to, and we are dedicated to supporting scientists across the globe with equitable access to new ideas and discoveries. As we spin out from Cornell University, arXiv staff, volunteers, and our supporters are working hard to ensure there is no interruption to the important service arXiv provides. arXiv is also committed to transparency. We are a community resource, and we want to make sure that our community is aware of any changes to arXiv that may affect them. We also want to make sure the community has a better understanding of how changes support arXiv’s long-term goals as an essential scholarly publishing infrastructure. arXiv has created a page dedicated to frequently asked questions and information about the arXiv’s establishment as an independent organization. This FAQ page will be updated promptly as new information is available. On this dedicated FAQ page, you can find out more about the new arXiv’s leadership, governance, and special projects. While the FAQ page will be the best place to find the most up-to-date information on arXiv’s spinout, major updates will also be announced through arXiv’s blog and social media as they occur. In addition to the FAQ page, arXiv will be releasing a series of blogs to cover recent arXiv milestones, new organizational updates, and upcoming changes, including: * Introduction to arXiv’s new Engineering Director and dev team projects * arXiv’s 3 million submission milestone * Recent changes to arXiv policy practice around AI articles These blogs will be up in the coming weeks, and as always, comments from our community are welcome.
blog.arxiv.org
write a paper that makes the key idea crystal clear. We are writing for people, and the writing should make it very easy for the reader to learn something. In a world like this, the bottleneck is not producing the proof but in digesting it, and we should review with that in mind.
Perhaps we need a "how to think" boot camp. An entire year of classes in which no computational tools are used. It would be like wilderness training. Build skills using paper, pencil, and a set of books. Build confidence in your own learning and reasoning abilities. 1/
Solving any design problem (any wicked problem) requires exposing yourself to the vulnerability of knowing nothing. You will never know less about the problem than when you first start solving it. Design problems are solved the same way that we learn to do design: by experiment. 5/
These are all great reasons for researchers to get started with an ORCID iD. 💫 Others include name disambiguation, name flexibility, and reduced administrative burden. Check out our page for researchers for more: info.orcid.org/researchers/ #ResearchSky
ORCID for Researchers
info.orcid.org
LifeSci Signal@lifescisignal.bsky.social · 2mo agoWhy every researcher should have an ORCID iD: @orcid.org 1. It's often required by journals, institutions and funders. 2. It makes your work more visible. 3. It connects your research activities in one place (less faff). 4. It follows you throughout your career. 5. It's free! #scicomm #openscience
“My mayor Muslim. My bagel’s Jewish. My Christian Dior. Knicks in four!” MD Hossain’s ode to his beloved basketball team in a viral 15-second clip is being embraced by people looking for joy in a fractious national moment.
He grabbed a mic to riff about the Knicks. His rhyme became an anthem of unity.
A fan blurted out an exuberant rhyme for his beloved basketball team. It has transcended sports and become a joyful anthem to unity.
washingtonpost.com
There is a little less than a week left to apply for the 5th CV4Ecology Short Course!! This program is a one-of-a-kind opportunity for ecologists to DIRECTLY build robust, applied AI expertise that is immediately usable and useful in their own research, for their own projects, on their own data.
CV4Ecology Workshop
cv4ecology.csail.mit.edu
Sara Beery@sarameghanbeery.bsky.social · 4mo agoWe are VERY excited to announce the call for applications has just been released for the 5th CV4Ecology Short Course!! This two-and-a-half-week intensive program trains ecologists and conservation practitioners to develop their own AI tools for their own data. Tell your friends! Apply now!
Excellent essay by @mazmhussain.bsky.social (and Anand Gopal). open.substack.com/pub/mazmhuss...
Conversation with Anand Gopal: Author of "Days of Love and Rage"
Anand Gopal is someone whose work I had been following for years prior to the publication of his gripping new history of the Syrian revolution, “Days of Love and Rage” which was released this year.
open.substack.com
Ted Chiang's role-playing argument is essentially my role-playng argument (as he kindly acknowledges), which in turn was much inspired by the "Simulators" essay by Janus (who probably wouldn't endorse his conclusion). www.nature.com/articles/s41...
Role play with large language models - Nature
By casting large-language-model-based dialogue-agent behaviour in terms of role play, it is possible to describe dialogue-agent behaviour such as (apparent) deception and (apparent) sel...
nature.com
Excellent and thought-provoking essay by Ted Chiang. [gift link] www.theatlantic.com/philosophy/2...
No, Artificial Intelligence Is Not Conscious
Taken to its logical conclusion, this line of thinking is absurd—and damning.
theatlantic.com
Possibly the best way to phish academics would be to just send them email that looks like a request to review an article and embed your malware in the "decline to review" link
Interesting post by Jessica Hullman. In a future where autonomous research is possible at scale, one can imagine companies building their own bodies of scientific knowledge and not sharing it. Unlike human scientists, AI scientists are not motivated to share their work.
AAndrew Gelman et al.@statmodeling.bsky.social · 3mo ago
What if scientists really were dispassionate observers, communicating ideas without irrational commitment? Look here, says AI. statmodeling.stat.columbia.edu/2026/05/28/w...
Excellent post on Waymo's handling of safety incidents. open.substack.com/pub/philkoop...
Robotaxi Safety Problems Are The New Normal
Standing down to manage risk is smart. But what needs to happen next?
open.substack.com
At @arxiv.bsky.social, we are receiving a new type of paper that I call an "I did this experiment" paper. These papers typically report some experiment with an LLM or LLM "agentic" workflow. They are the kind of experiments an "insider" engineer would run to optimize a system. 1/
#arXiv is banning researchers for one year whose manuscripts contain hallucinated references. My latest for @nature.com: www.nature.com/articles/d41... @briannosek.bsky.social, @tdietterich.bsky.social, @richardsever.bsky.social, @dermotlynott.bsky.social, @philipncohen.com, #WCRI2026
Researchers who use hallucinated references to face arXiv ban
The preprint server is the latest to impose stiff penalties on authors who contribute to AI ‘slop’ — but not everyone is convinced it’s the right approach.
nature.com
This article is clearly some flavor of rage bait, but (after I stopped sighing at the tone and the contradictions in the argument) it did make me think. www.timeshighereducation.com/campus/dear-... So as one of those lazy, technologically-impaired, boring professors, here's some thoughts. 🧵
Dear educators, Gen Z here. Could you please teach us like it’s 2026?
Learning has changed drastically in the past five years, but teaching practices haven’t caught up. From your students’ perspective, here’s how to adapt
timeshighereducation.com
We are implementing a similar policy at @arxiv.bsky.social. If there is incontrovertible evidence of LLM slop in a paper, this means the authors did not take the time to read the LLM output and we can't trust anything else in the paper. Penalty is 1 year ban from arXiv followed by...