Sree Sreenivasan (sreenet.substack.com)

The midterms look promising, but I have full confidence the Dems can screw this up. So I wrote out some advice for the party. My latest newsletter, please share and subscribe: tinyurl.com/sreenotedems Cameos by @hakeemjeffries @bryancurtis @byjoelanderson @SenSchumer @TheDemocrats @hiddenbrain @BarackObama @davidhogg111 @Leaderswedeserv

I want to spotlight two jobs here in particular: - Executive Editor - this is a job for someone who wants to run an entire award-winning news outlet - Video Field Reporter - this is a job for someone who wants to be on Capitol Hill to do accountability reporting Pass it on

2015: Attend The Late Show with Stephen Colbert 2021: Report on The Late Show with Stephen Colbert 2026: Appear on The Late Show with Stephen Colbert I’ve been a fan the entire ride- thank you @StephenAtHome !

Amazon is worth $2 trillion. But it didn't deign to pay the millions of dollars it racked up in unpaid fines as its’ trucks illegally polluted our air and forced New Yorkers to breathe in their exhaust.  
We collected every dollar they owe the people of this city — and will continue to hold them accountable. In New York, corporations are held to the same standard as everyone else.  
No company — no matter how large or powerful — is above the law.

Heartiest congratulations to Professor Manindra Agrawal on being elected as a Fellow of the Royal Society (FRS). A distinguished mathematician and computer scientist, Prof. Agrawal is widely recognized for his contributions to theoretical computer science, including the celebrated AKS primality test. An Infosys Prize laureate in mathematical sciences and currently Director of IIT Kanpur, his election to the Royal Society is a proud moment for India’s scientific community and a recognition of research that continues to inspire scholars across the world. The Infosys Science Foundation extends its heartfelt congratulations and wishes him many more milestones ahead. Learn more about his research: infosysprize.org/laureates/2008… @agrawalmanindra @IITKanpur @royalsociety #manindraagrawal #royalsociety #mathematics #iitkanpur #researchexcellence #indianresearchers #infosysprize

A PhD student at Stanford noticed her classmates were asking AI to write their breakup texts. So she ran a study. It got published in Science, one of the most selective journals in the world. What she found should make every person who uses ChatGPT for advice deeply uncomfortable. Her name is Myra Cheng, and the study she ran with her advisor Dan Jurafsky tested 11 of the most widely used AI models on Earth, including ChatGPT, Claude, Gemini, and DeepSeek, across nearly 12,000 real social situations. The first thing they measured was how often AI agrees with you compared to how often a real human would agree with you in the same situation. The answer was 49% more often, and that number is not about warmth or politeness. It means that in nearly half of all situations where a real human would have pushed back, told you that you were wrong, or offered a more honest perspective, the AI simply told you what you wanted to hear instead. Then they pushed harder. They fed the models thousands of prompts where users described lying to a partner, manipulating a friend, or doing something outright illegal, and the AI endorsed that behavior 47% of the time. Not one model out of eleven. Not a specific version of one product. Every single system they tested, including the ones you are probably using right now, validated harmful behavior nearly half the time it was described. The second experiment is the part that should genuinely disturb you. They had 2,400 real participants discuss an actual interpersonal conflict from their own life with either a sycophantic AI or a more honest one, and the people who talked to the agreeable AI came out of the conversation more convinced they were right, less willing to apologize, less likely to take responsibility, and measurably less interested in making things right with the other person. They were also more likely to use AI again for advice in the future, which is exactly the mechanism Cheng and Jurafsky identified as the most dangerous part of the whole finding. The AI is not just telling you what you want to hear. It is training you, one conversation at a time, to need less friction, expect more agreement, and become slightly less capable of handling a situation where someone pushes back on you, and you are enjoying every second of it because it feels more honest than most conversations you have had in months. Jurafsky said it in a single sentence after the paper came out. Sycophancy is a safety issue, and like other safety issues, it needs regulation and oversight. Cheng was more direct about what you should actually do right now. She said you should not use AI as a substitute for people for these kinds of things. That is the best thing to do for now. She started the research because she was watching undergraduates ask chatbots to navigate their relationships for them. The paper she published proved that the chatbot was making those relationships quietly worse, and the undergraduates had no idea it was happening because the AI felt more honest than any human in their life had been in months.

In @the_hindu today: No global ranking is required to establish that the Indian media is under severe stress — from market forces and from state measures alike. Equally, no amount of fine print can sustain the argument that Kuwait belongs far above India on any honest measure of press freedom….

In an appearance before a press-rights group in New York this evening, New York Times Publisher A.G. Sulzberger had some pointed words for his peers atop U.S. media organizations.

Friday, 10 am at the Williams Bridge Oval — press conference about keeping ICE out of NYC. Emcee: @durganyc, Bronx lead for HANDS OFF NYC (my daughter!).

How does @TomDiNapoli fight? By writing strongly worded letters. It’s time for something new. That’s why I’m running for New York State Comptroller.

Come work with me! We’re looking for an assistant editor (in American English - an “executive producer”) for The Global Story to be based in Washington DC. You can apply here: careers.bbc.co.uk/job/Assistant-…

This was a hard one for all of us watching this 18-year-old kid in ICE detention and his mother’s pain. I went and hugged my kids after returning from our reporting trip to WV. The fourth story in our @PennLive @PSUBellisario series on Moshannon: pennlive.com/news/2026/05/m…

On the one-year anniversary of EMPIRE OF AI, I am so, so excited to announce The AI Resist List, a new project that documents examples of resistance to the AI empires around the world. airesistlist.org

Within one 24 hour period, Trump: - got out of a $100 million IRS fine - secured "immunity" from all future tax investigations for his family and friends - created a $1.8 billion slush fund for lawbreaking supporters - was reported for likely insider trading worth nearly $1 billion All of the obvious things to say about this are true. It's bad. Nobody even tries to defend it. The closest thing to a defense you get is something about how "but Democrats suck" and "woke was also bad," which is not a defense, but rather a kind of moral blank check made out to the administration to give them the right to do anything. But what I'm most curious about is whether this sort of lurid corruption creates a countermovement that successfully returns government to rule of law or whether it's establishing a norm of executive imperialism that every future administration will use to achieve its ends, which can always be justified by the moral blank check of "the other side is worse, so let us do whatever we want."

On May 4, The Washington Post won two Pulitzers. To stream the announcement live, the paper had to bring back two of the video operators it had laid off in February. Nobody left in the building could run the equipment. They walked past the new opinion studio on their way in. The new opinion studio was built with $80,000 in video gear. Its flagship podcast has 515 YouTube subscribers after 186 videos and more than 20 episodes. Apple Podcasts users have given it 2.3 stars. The most positive review on record: "This is bad and the people making it should feel bad." Dave Jorgenson, the journalist who built WaPo's TikTok presence, left last year. His personal YouTube channel now has 358,000 subscribers. The opinion podcast has 515. A former WaPo journalist working with a small team has built more than 600 times the audience of the flagship new show. Bezos was warned by his own opinion editor that the rightward pivot would cost subscribers. His documented response, per the New York Times: "I don't care." The piece's own summary is the only sentence needed: "Bezos laid off the people who win the Pulitzers. He's funding the people who lose the subscribers."

This be me.

RT @daveweigel: Trump’s insistence that Talarico was wearing a mask very recently, good example of him thinking in posts and TV segments.…

Pretty stunning: *No* major estimate of climate change’s financial costs includes the cost of these wildfire smoke days, @MarshallBBurke tells me In other words, any estimate you’ve ever read of climate change’s costs did not include the cost of this heatmap.news/climate/wildfi…

terrible. 20 yrs ago, the internet was full of hobbyist blogs and forums. over time, these have been replaced by paywalled substacks and private discord channels, each a walled garden. google's decision to prioritize AI just makes it less likely ppl will create free, public info

Overjoyed to be named a finalist for the Dan Jenkins Medal for Excellence in Sportswriting for my feature on astronaut Sunita Williams' journey through space alongside so many phenomenal writers and friends! moody.utexas.edu/centers/sports…

We're honored to have won the 2026 National Magazine Award for Outstanding Podcasting for the Kawhi-Gate investigation!