nxthompson's banner

nxthompson

@nxthompson • 129,057 subscribers

CEO @theatlantic. Going out of business soon, since 1857. Author of “The Running Ground" and host of "The Most Interesting Thing in AI."

Shorts

Set a new American record for men 50+ in the 50k yesterday, running 3:10 at Mad City. When my calf cramped at mile 27, I just repeated in my head what my son had said after I bombed out in the NYCM: “bury that garbage.” Took it down to 5:55 pace and got the mark. 🤣

Set a new American record for men 50+ in the 50k yesterday, running 3:10 at Mad City. When my calf cramped at mile 27, I just repeated in my head what my son had said after I bombed out in the NYCM: “bury that garbage.” Took it down to 5:55 pace and got the mark. 🤣

199,926 views

Videos

Anya Rossi

sweetdream.ai

SweetDream.ai•Sponsored•Livecam

Watch Anya Live

Anya is streaming live right now! Join her private show and enjoy exclusive content.

Exclusive private shows

1.2k viewers online

Private Show

Join now for exclusive access

Free preview available • Premium content

Meta’s AI glasses are one of the company’s most successful products in years. They offer genuine utility, with a direct way for users to interact with its AI. But people have long been wary of smart glasses—especially when face recognition is involved. Here’s Boz on Meta's NameTag feature, which identifies faces and alerts the glasses wearer when it detects someone they know. Produced by Atlantic Re:think, The Atlantic's creative marketing studio.

Meta’s AI glasses are one of the company’s most successful products in years. They offer genuine utility, with a direct way for users to interact with its AI. But people have long been wary of smart glasses—especially when face recognition is involved. Here’s Boz on Meta's NameTag feature, which identifies faces and alerts the glasses wearer when it detects someone they know. Produced by Atlantic Re:think, The Atlantic's creative marketing studio.

259,195 views • 21 days ago

A new paper shows that AI does make your writing more homogenous—and it'll do the same even if you just ask it to edit your writing. Even weirder: it's not like it'll push you toward the voice of an average human, it'll push you toward a kind of AI-optimized voice that's subtly different from the way any people write.

A new paper shows that AI does make your writing more homogenous—and it'll do the same even if you just ask it to edit your writing. Even weirder: it's not like it'll push you toward the voice of an average human, it'll push you toward a kind of AI-optimized voice that's subtly different from the way any people write.

123,708 views • 4 months ago

I was a little surprised by Sam Altman's answer here. But if he’s right—and you can train a model completely on synthetic data—it would have all sorts of implications for how we develop and deploy this technology. Watch my full interview with Sam here: Produced by Atlantic Re:think, The Atlantic's creative marketing studio.

I was a little surprised by Sam Altman's answer here. But if he’s right—and you can train a model completely on synthetic data—it would have all sorts of implications for how we develop and deploy this technology. Watch my full interview with Sam here: Produced by Atlantic Re:think, The Atlantic's creative marketing studio.

66,386 views • 3 months ago

Nick Bostrom says there are a few reasons to treat AI models with respect: 1. It’s the right thing to do. 2. It helps build good habits. 3. The models might remember it when they become more powerful than us and spare the human race from total annihilation. He had some fascinating points. You can watch our full convo here: Produced by Atlantic Re:think, The Atlantic's creative marketing studio, in collaboration with PwC.

Nick Bostrom says there are a few reasons to treat AI models with respect: 1. It’s the right thing to do. 2. It helps build good habits. 3. The models might remember it when they become more powerful than us and spare the human race from total annihilation. He had some fascinating points. You can watch our full convo here: Produced by Atlantic Re:think, The Atlantic's creative marketing studio, in collaboration with PwC.

50,493 views • 2 months ago

As Sam Altman says here, no one knows what the media business will look like in the AI era. But it's increasingly clear it will involve agents–perhaps on a scale where vastly more of them interact with our content than humans. That could open up all sorts of possibilities with micropayments, licensing, bulk subscriptions, and other ideas no one has thought of yet. Watch my full conversation with Sam here: Produced by Atlantic Re:think, The Atlantic's creative marketing studio.

As Sam Altman says here, no one knows what the media business will look like in the AI era. But it's increasingly clear it will involve agents–perhaps on a scale where vastly more of them interact with our content than humans. That could open up all sorts of possibilities with micropayments, licensing, bulk subscriptions, and other ideas no one has thought of yet. Watch my full conversation with Sam here: Produced by Atlantic Re:think, The Atlantic's creative marketing studio.

36,138 views • 2 months ago

This was a brilliant answer from Yuval Noah Harari on how AI might change the world. It's still far too early to tell, but this might become the most powerful technology ever created, so we should probably spend more time worrying about the implications of that and less time stressing over ROI every quarter. #Davos26 #WEF26

This was a brilliant answer from Yuval Noah Harari on how AI might change the world. It's still far too early to tell, but this might become the most powerful technology ever created, so we should probably spend more time worrying about the implications of that and less time stressing over ROI every quarter. #Davos26 #WEF26

52,671 views • 6 months ago

The most interesting thing in tech: Commerce bans foreign nationals—including Anthropic's own staff!—from using Fable. So the company shuts it down for all. Anthropic keeps warning AI is existentially dangerous but then keeps shipping. The government wants AI built here but incoherent regulations just derailed a leading product. It's not going well, folks.

The most interesting thing in tech: Commerce bans foreign nationals—including Anthropic's own staff!—from using Fable. So the company shuts it down for all. Anthropic keeps warning AI is existentially dangerous but then keeps shipping. The government wants AI built here but incoherent regulations just derailed a leading product. It's not going well, folks.

17,740 views • 1 month ago

This is one of the more interesting cases I’ve heard for why we should put AI data centers in space. Yes, it would benefit a few billionaires, but it could also lessen the environmental toll on Earth. And we’re going to need to cut AI's carbon footprint in as many ways as we can. Space architect Ariel Ekblaw and I went deep on the opportunities—and many challenges—of building AI infrastructure in orbit. Watch our conversation here: Produced by Atlantic Re:think, The Atlantic's creative marketing studio, in collaboration with PwC.

This is one of the more interesting cases I’ve heard for why we should put AI data centers in space. Yes, it would benefit a few billionaires, but it could also lessen the environmental toll on Earth. And we’re going to need to cut AI's carbon footprint in as many ways as we can. Space architect Ariel Ekblaw and I went deep on the opportunities—and many challenges—of building AI infrastructure in orbit. Watch our conversation here: Produced by Atlantic Re:think, The Atlantic's creative marketing studio, in collaboration with PwC.

25,785 views • 3 months ago

Sitting on some steps at a NYC high school waiting for my son to finish soccer practice just now …

Sitting on some steps at a NYC high school waiting for my son to finish soccer practice just now …

17,458 views • 3 months ago

The most interesting thing in tech: Might AI agents create a new privacy risk we’re not thinking enough about? Meredith Whittaker of Signal thinks they are. #WEF2026

The most interesting thing in tech: Might AI agents create a new privacy risk we’re not thinking enough about? Meredith Whittaker of Signal thinks they are. #WEF2026

28,793 views • 6 months ago

If you are going to give a commencement address in the next couple of weeks, do not mention AI or you might get loud boos. It happened in Florida last week; and it just happened to Eric Schmidt in Arizona. Many young people are decidedly not excited about AI and the future it promises to deliver them.

If you are going to give a commencement address in the next couple of weeks, do not mention AI or you might get loud boos. It happened in Florida last week; and it just happened to Eric Schmidt in Arizona. Many young people are decidedly not excited about AI and the future it promises to deliver them.

14,023 views • 2 months ago

One of the most interesting ideas I heard this year: In a world in which everyone uses A.I. for everything, the next great competitive edge might lie with people who keep their minds sharp by doing tasks themselves—even when they could be easily outsourced to A.I. I thoroughly enjoyed talking to Nita Farahany about the future of human cognition. Listen here: Produced by Atlantic Re:think, The Atlantic's creative marketing studio, in collaboration with PwC.

One of the most interesting ideas I heard this year: In a world in which everyone uses A.I. for everything, the next great competitive edge might lie with people who keep their minds sharp by doing tasks themselves—even when they could be easily outsourced to A.I. I thoroughly enjoyed talking to Nita Farahany about the future of human cognition. Listen here: Produced by Atlantic Re:think, The Atlantic's creative marketing studio, in collaboration with PwC.

25,449 views • 7 months ago

I did not know how much of a data advantage Google has over the other AI companies. Matthew Prince 🌥 says they can access 3X as much data as OpenAI and nearly 5X as much as Bing. This could make it hard for those models—and Grok—to beat Gemini in the long run.

I did not know how much of a data advantage Google has over the other AI companies. Matthew Prince 🌥 says they can access 3X as much data as OpenAI and nearly 5X as much as Bing. This could make it hard for those models—and Grok—to beat Gemini in the long run.

22,536 views • 6 months ago

I’m thrilled to announce The Most Interesting Thing in AI is back for another season. This moment in AI is the most exciting, as well as unsettling, that I’ve lived through. It has spurred lots of questions about where we are heading—and I’ve got an incredible list of guests to help me work through them. Episode one drops next Wednesday. We’ve officially made the switch to video; I hope you’ll subscribe and watch on YouTube: Produced by Atlantic Re:think, The Atlantic's creative marketing studio, in collaboration with PwC.

I’m thrilled to announce The Most Interesting Thing in AI is back for another season. This moment in AI is the most exciting, as well as unsettling, that I’ve lived through. It has spurred lots of questions about where we are heading—and I’ve got an incredible list of guests to help me work through them. Episode one drops next Wednesday. We’ve officially made the switch to video; I hope you’ll subscribe and watch on YouTube: Produced by Atlantic Re:think, The Atlantic's creative marketing studio, in collaboration with PwC.

10,963 views • 3 months ago

This week on my podcast, The Most Interesting Thing in AI, I had the pleasure of talking with Audrey Tang about two big topics: How can AI help democracy? Audrey has helped get Taiwan to use the tools of modern technology to strengthen democracy even as they have frayed it around the world. What is the secret, and could it work in the U.S? How can we build humanistic and pluralistic values into the most powerful AI? Audrey has a great theory about the two races in AI---one vertical, and one horizontal---and how, with the tools of open-source, it's possible that we can align all future models with our highest ambitions. It’s a brilliant framework for thinking about the future of AI. I hope you'll listen to our conversation: Produced by Atlantic Re:think with PwC.

This week on my podcast, The Most Interesting Thing in AI, I had the pleasure of talking with Audrey Tang about two big topics: How can AI help democracy? Audrey has helped get Taiwan to use the tools of modern technology to strengthen democracy even as they have frayed it around the world. What is the secret, and could it work in the U.S? How can we build humanistic and pluralistic values into the most powerful AI? Audrey has a great theory about the two races in AI---one vertical, and one horizontal---and how, with the tools of open-source, it's possible that we can align all future models with our highest ambitions. It’s a brilliant framework for thinking about the future of AI. I hope you'll listen to our conversation: Produced by Atlantic Re:think with PwC.

28,514 views • 1 year ago

The A.I. companies took our content in secret and used it to build businesses worth billions of dollars. Then they said it’s impossible to compensate us for the use of our work. Bill Gross recognized this as one of the great injustices of our time and decided to do something about it. He left his job, started ProRata.ai, figured out how to reverse engineer the outputs of an LLM, and devised a system for sharing revenue. It could finally establish a fair exchange of value between A.I. companies and creators. Bill and I dove deep on payment and attribution in the age of A.I. Listen to our full chat here: Produced by Atlantic Re:think, The Atlantic's creative marketing studio, in collaboration with PwC.

The A.I. companies took our content in secret and used it to build businesses worth billions of dollars. Then they said it’s impossible to compensate us for the use of our work. Bill Gross recognized this as one of the great injustices of our time and decided to do something about it. He left his job, started ProRata.ai, figured out how to reverse engineer the outputs of an LLM, and devised a system for sharing revenue. It could finally establish a fair exchange of value between A.I. companies and creators. Bill and I dove deep on payment and attribution in the age of A.I. Listen to our full chat here: Produced by Atlantic Re:think, The Atlantic's creative marketing studio, in collaboration with PwC.

15,653 views • 8 months ago

The most interesting thing in tech: Cloudflare has built a new tool that allows publishers to identify, track, and block any AI bots that are scraping their sites without permission. Here Gitesh Gohel and I explain the spreadsheet we've built to decide who to block --- and then we start blocking. It feels like an important moment where the balance of power is shifting back to creators and away from companies who have scraped with impunity and without any fair exchange of value.

The most interesting thing in tech: Cloudflare has built a new tool that allows publishers to identify, track, and block any AI bots that are scraping their sites without permission. Here Gitesh Gohel and I explain the spreadsheet we've built to decide who to block --- and then we start blocking. It feels like an important moment where the balance of power is shifting back to creators and away from companies who have scraped with impunity and without any fair exchange of value.

20,871 views • 1 year ago

My read on how A.I. companies view revenue sharing with publishers: - Some just want our businesses to die. - Some don’t want the hassle of dealing with us. - Some want to pay us, but they don’t want to be the only ones. We need to find a way to get all of them on board. It wouldn’t just be the right thing to do—it would be good business. YouTube did this years ago. The platform is now worth half a trillion dollars. I had a riveting chat with Bill Gross about how to fix A.I.’s fairness problem. Listen to our full convo here: Produced by Atlantic Re:think, The Atlantic's creative marketing studio, in collaboration with PwC.

My read on how A.I. companies view revenue sharing with publishers: - Some just want our businesses to die. - Some don’t want the hassle of dealing with us. - Some want to pay us, but they don’t want to be the only ones. We need to find a way to get all of them on board. It wouldn’t just be the right thing to do—it would be good business. YouTube did this years ago. The platform is now worth half a trillion dollars. I had a riveting chat with Bill Gross about how to fix A.I.’s fairness problem. Listen to our full convo here: Produced by Atlantic Re:think, The Atlantic's creative marketing studio, in collaboration with PwC.

15,420 views • 8 months ago

The most interesting thing in tech: A brief conversation with AI pioneer Kai-Fu Lee Lee about the tech dynamics between the US and China. Why, for one, is China so much more focused on open source AI?

The most interesting thing in tech: A brief conversation with AI pioneer Kai-Fu Lee Lee about the tech dynamics between the US and China. Why, for one, is China so much more focused on open source AI?

22,069 views • 1 year ago