Tom Friedman: AI, China, Globalization and America’s Next Era | Coffee with Ken

The Washington Post, once a great newspaper essential for the nation's capital, is being dismantled by its owner for apparent political reasons, constituting a travesty and loss of crucial institutional capacity.

“What's going on at the Washington Post is a travesty. A great newspaper. So necessary for the nation's capital and and being filleted, you know, by its owner um apparently for political reasons. Uh that would be a that's a shameful thing.”

In 2013, Senator Orrin Hatch asked Mark Zuckerberg 'if you give Facebook away for free, how do you make money?' and Zuckerberg answered 'Senator, we sell ads,' showing that regulators lack basic understanding of how innovators make money, and this gap has only widened with AI.

“there's a hearing in Congress. Gosh, this was maybe in 2013 where Senator Orrin Hatch was interviewing Mark Zuckerberg. Yeah. Right. And he famously asked him, 'Mr. Zuckerberg, if you give Facebook away for free, how do you make money?' And he said, 'Senator, we sell ads.' So, today the regulator has no clue how the innovator even makes money. And that was with the internet. Now, where do we get to AI?”

Secretary of State Kissinger in 1973 needed three dimes and three months to achieve disengagement agreements between Israel, Egypt, and Syria—a tic-tac-toe-level problem with three players.

“October 1973 and Henry Kissinger... to produce the first disengagement agreements in the Middle East between Israel and Egypt, Israel and Syria needed three dimes, an airplane, three months and Kissinger magic...He needed a dime to call Israeli Prime Minister Golda Meir, a dime to call Egyptian President Anwar Sadat, a dime to call Syrian President Hafez al-Assad.”

Destroying 80 years of alliance structures results in a much less prosperous and much less stable world—this is physics, not Republican or Democrat politics.

“And you destroy it. Um uh you're going to end up with a much less prosperous and a much less stable world. And that's not Republican or Democrat. That's physics.”

The Polyscene is the product of the Big Bang where everything gets sharded into little pieces, each piece now has a voice, and all are connected; governing anything in the Polyscene is hell on wheels because everyone is empowered and connected simultaneously.

“We're in the Poly scene. c e n e. We're in the Poly scene. And the Poly scene is the product of everything cuz of the Big Bang, everything gets sharded into little pieces. Every piece now has a voice and they're all connected. And that's why in the policy you don't want to run anything. Especially a university. Because now everyone has a voice, is empowered, and connected. So governing anything in the policy is hell on wheels.”

The core problem is not whether a country develops the leading AI (LLM) or applies it to their economy, but which country learns to govern it, which is more important than traditional military or economic dominance.

“who develops the leading LLM? Which country? Sure. Which country applies it to its economy most efficiently? You can develop it but not apply it also, you know. Um uh and which country learns to govern it um is going to be very very important.”

The history of the world can be divided into three phases: the age of ice (tools, pre-printing press), the age of water (information flow, printing press to 2017), and the age of steam (AI vapor, 2020s onward), with humans as H2O molecules and technology as heat.

“Craig and I divide the history of the world up into three phases, just to simplify things. The first we call and humans are H2O molecules and technology is heat. So, the first phase is from the dawn of uh, of mankind uh, to the printing press. We call it the age of ice or the age of tools... the age of information or the age of water... until basically 2017, invention of the transformer algorithm, which is really what made uh AI possible. So, early 2020s, we get AI. Now, more heat. And now the water boils into a vapor. Into steam.”

The last 80 years were the most peaceful and prosperous period in modern history absent great power war, and this was possible because the world was organized the way America was organized—change that without study or justification and you change everything for the worse.

“I happen to believe that the last 80 years of world history were the most peaceful and prosperous on a relative basis absent great power war than any period any period of modern history. And there is a reason for that. The world was the way the world was because America was the way America was. And America was the way America was toward that world. You change that? You change that without any clue or plan or just cuz your gut instinct or your biases or because the president of Switzerland wasn't nice to you? He gave you she gave you an owie, you know? Um you change all that and you change the world.”

Despite closure of major institutions like the Washington Post, journalism continues and is 'exploding now from the bottom up' through independent creators like 'Coffee with Ken,' requiring audiences to find trustworthy sources rather than relying on institutional gatekeepers.

“there's no Washington Post, but there's coffee with Ken...it's not like there's no journalism at. It's just exploding now from the bottom up. But you have to find it and you have to learn to trust it, you know, in in who's really trustworthy in all this.”

When everything has AI embedded in it, everything becomes like TikTok in terms of data control, algorithm ownership, and data storage location—creating geopolitical risk if the US and China do not have a shared trust architecture for AI.

“when everything is TikTok? Okay?... You go into your doctor, your orthopedist, and he tells you, Ken, you need a new hip... it's called Shanghai hip replacement.ai... it's going to be broadcasting 24/7... Your whole physiology, your biology, um all data is going to be stored in Shanghai... if we don't have a shared trust architecture for the use of AI with China, you're going to buy that hip replacement?”

AI unlike any previous technology is pervasive, entering everything simultaneously—classrooms, toasters, refrigerators, cars, and medical devices like hip replacements—creating a phase change where steam (AI) diffuses into all systems rather than concentrating in specific sectors.

“unlike any previous technology, it's going into everything. It's going into your classes. It's going into your toaster. It's going into your refrigerator. It's going into your car. And it's going into your hip replacement. So, it's going everywhere.”

When you are an H2O molecule in ice, you have no idea what it is like to be water, and when you are water, you have no idea what it is like to be steam, because each state operates in a completely different way—therefore humans cannot currently imagine what the steam age will look like.

“when you're an H2O molecule in ice, you have no idea what it's like to be water. And when you're an H2O molecule in water, you have no idea what it's like to be steam because you operate in a completely different way.”

The Biden administration was attempting to normalize relations between Israel and Saudi Arabia, which would have been the biggest expansion of a Middle East whole and integrated region since Camp David, but Iran stopped this process.

“In the Middle East, the Biden administration was trying to forge an agreement between Israel and Saudi Arabia. They were trying to get Israel and Saudi Arabia to normalize relations. Um that would have been the biggest expansion of a Middle East whole and integrated since Camp David. And Iran stopped the second.”

Russia and Iran both understood they would be isolated in unified regions (Europe whole and free, Middle East whole and integrated), so each moved independently to prevent regional integration, demonstrating that great power resistance to inclusion is a key driver of contemporary conflict.

“both regions were trying to get toward greater inclusion and integration. And um these basically outline powers who understood they'd be isolated move, not in coordination, but each in their own way, move to stop it.”

Trump 1 had institutional buffers (General Kelly, Mattis) who would push back on impulsive ideas, whereas Trump 2 is surrounded by amplifiers who implement ideas directly, creating a difference between constrained and unbuffered executive power.

“Trump one was surrounded by buffers. And Trump's two is surrounded by amplifiers. That's the central truth of what's going on now. So, in Trump one, he would go to the Mar-a-Lago bar, some guy would sit down and say, 'Mr. President, I've got a great idea for you. Let's make Gaza into the Riviera.' Say, 'What a great idea.' Go right to the Oval Office, sit down with General Kelly. 'Kelly, we're going to make Gaza into the Riviera.' Kelly would say, 'Boss, what an amazing idea. Why didn't I think of that?' Okay. Let's have Mattis study it, and do it do it, and and they would deflect it over here. Now it goes from the Mar-a-Lago bar to the Oval Office to the world completely unbuffered.”

Current US foreign policy (attempting ceasefire in Gaza through son-in-law and real estate guy) has no institutional backing (no NSC, no State Department, no Voice of America), so even if successful, there is no mechanism to follow up or implement agreements.

“we're trying to bring peace to Iran to you know, Russia right now with the president's son-in-law and a real estate guy...there's no NSC, there's no State Department, there's no Voice of America...when this stuff even gets done, how is anything followed up on?”

The printing press (Gutenberg) transformed communication from written (slow, disconnected molecules) to printed (flowing, connected ideas), analogous to how AI transforms from information processing to direct sensing and reasoning about physical reality

“And the printing press, Ken, someone was alive. And some monk said to some priest, 'Now, that is really cool. Okay? You mean I don't have to write this Bible out longhand anymore? We're going to How did you say print it? Okay.' Um, well, we are here for a similar moment.”

Deep learning, which emerged in 2012, represents the ability to translate the physical world into the digital world via sensors, and AI will multiply this capability so that everything—from this table to shoes to all material objects—will begin speaking to us and revealing itself at unprecedented depth.

“they put sensors everywhere, in the muck, in the mud, in the shore... fed all that into big data... John Kelly... 'Tom, the world has been speaking to us all these years. We just couldn't hear it.' And now we're going to be able to hear it... Sensors translate the physical world into the digital world... we're going to take that sensing... feed it into AI... Everything is about to start speaking to us.”

Secretary of State Blinken faces a Rubik's Cube rather than tic-tac-toe: Gaza involves Hamas political/military wings and Islamic Jihad; Lebanon involves Hezbollah, Iran, and Lebanese government; Syria involves Russia, Iran, Hezbollah, and Syrian government; Iraq involves 18 militias; Yemen has 18 tribes; Israel has 18 political parties—an exponentially more complex problem than Kissinger's.

“Tony Blinken sees double and triple everywhere he goes. Um, he's got to deal with inside Hamas and outside Hamas, political Hamas and military Hamas, Hamas and Islamic Jihad in Gaza. In Lebanon, he had to deal with Hezbollah, Iran and the Lebanese government. In Syria, he had to deal with Russia, Iran, Hezbollah and the Syrian government. In Iraq, he had to deal with 18 militias in Yemen, he had 18 tribes and Israel, he had 18 parties. He was not playing tic-tac-toe. He had a Rubik's Cube and he was trying to get all the colors aligned on one side of a Rubik's Cube.”

Ukraine attempted to join the European Union, which would have been the biggest expansion of a Europe whole and free since the fall of the Berlin Wall, bringing Europe to 98% completion of unification.

“Ukraine was trying to join the West and...Ukraine was trying to join the European Union. Um and if Ukraine had been able to join the European Union, it would have been the biggest expansion of a Europe whole and free since the fall of the Berlin Wall. Ukraine's a big, powerful, important country. We would be 98% of the Western Balkans to a Europe whole and free.”

Washington's government is organized around 'who hates China the most,' so officials trained to maximize China competition will become out of sync or out of jobs when the problem shifts to trust architecture, creating structural resistance to the solution.

“the government, our whole government set up around who hates China the most. So, there's a whole group of people at the Pentagon who got their job today cuz they hated China the most. Okay? Now, you again, it's a phase change. The things completely shifts, you know? And you have to rethink the whole problem. And so, there's a whole bunch of people going to be either out of jobs or out of sync, you know?”

We are currently at 211.9°F in a temperature metaphor where water boils at 212°F, meaning we are sitting on the cusp of the AI phase change with no ability to imagine what the next state will be like.

“when people ask Greg or me what time it is, we don't give them an hour. We give them a temperature. We tell them the time is exactly 211.9° Fahrenheit as water boils at 212.”

Betting markets (like Polymarket) exemplify Polyscene complexity: traditional binary bets (Vikings beat Bears) gave way to multi-dimensional micro-bets (field goal accuracy within specific margins, individual player performance on specific plays), showing how systems fragment into granular, interconnected options

“when I was growing up, my dad had friends who bet on the Vikings game up in Minnesota, where I'm going in an hour. Um and it was a very simple thing. I want the Vikings to beat the Bears inside the spread. Totally binary. It was just the Vikings win or lose. Today, I can bet on whether the field goal will go inside the left mark or inside the white mark. Okay? I can bet on every play, every player's performance. Where do I bet on that? On Polymarket.”

We have become godlike in seven ways: developing artificial brain greater than human intelligence, disconnecting intelligence from biology, creating a planetary nervous system, changing the planet's geological era from Holocene to Anthropocene, splitting the atom and acquiring capacity to destroy the world, creating deep learning, and approaching a genesis moment through combining quantum computing, fusion energy, and AI.

“We have become godlike and we have become godlike in seven different ways. We have developed an artificial brain greater than the creator endowed us with. We've disconnected human intelligence from biology... We've created a planetary nervous system... We're the first species to change the planet... We created a new galaxy called cyberspace... we split the atom... Sixth, we created deep learning... When we combine quantum computing with fusion energy, with AI, we will have unlimited compute with unlimited clean energy with AI, and that will be a genesis moment”

Regulators who are not constantly learning and staying current with technology will become disconnected and unable to make good decisions, or will be manipulated by technologists pursuing their own interests in avoiding regulation.

“if you're not doing that when the pace of change is this fast, you're going to get completely disconnected, and you're not going to be able to make good decisions at all. Um or you'll be manipulated by the technologists for their personal interests.”

iPad devices contain components and supply chains spanning six continents in a poly-economic global knowledge and material network, so tariffs disrupting this network will damage innovation and production across all participating countries, not strengthen any one nation.

“no country in the world can make this device. Cuz this is a complex device. Complex devices solve complex problems. But it is so complex that it's both idea, supply chain, and material supply chain spans six continents. This is the product of a poly-economic global knowledge and material supply chain. If you come in with tariffs in the middle of it, you completely interrupt the whole thing.”

Developing nuclear weapons was managed by government (Manhattan Project), but AI is being developed by private companies with private interests, particularly in avoiding regulation, making AI development fundamentally more dangerous than nuclear weapons development.

“This is a scary moment. We are developing nuclear weapons. Only when we did the Manhattan Project, it was a government project. The people developing these technologies are all private companies. Um, and they have private interests. Um, one of it which is did not be regulated.”

Much of current AI policy narrative focuses on speed (beating China, being first with AI), but this ignores that Apple invested $50 billion annually in China over 15 years while the US Chips Act is only $50 billion over 5 years, showing the private sector has already made globalization choices that constrain AI competition.

“Apple has invested 50 billion a year in China in manufacturing and technology. So while the tone and and framing is we're against China etc. Have we not done this to ourself in a large degree by making so much of the investment from the private sector in the United States in globalization?”

The post-Cold War era has ended, marked by US withdrawal from Afghanistan, Russian invasion of Ukraine, and the rise of China, requiring a new name for the era just as Walter Lippmann named the Cold War in 1947.

“the post Cold War ended. Um wasn't like a moment, but US withdrawal from Afghanistan, Russian invasion of Ukraine, the rise of China. We're not in the post Cold War this moment of great American anymore.”

Radical breaks with 80 years of American foreign policy (breaking the Atlantic alliance) are historically only associated with wars or massive economic disruptions, but the current break is based on the whims, superstitions, and fantasies of a single unbuffered person, creating extreme instability.

“these kind of radical moves are only associated with wars or economic huge economic disruptions. And this one is based on the whims superstitions um and fantasies of a single person who is completely unbuffered by an accident of history that his party controls the Senate, the Congress, and the Supreme Court”

Polymathic artificial general intelligence is an AI system that has mastered chemistry, biology, physics, material science, music, and baseball, and can reason across all of them at a high dimensional rate that no human could achieve.

“polymathic artificial general intelligence is an AI system that's mastered chemistry, biology, physics, material science, Mozart, and baseball. Okay, so and you can reason across all of them at a high dimensional rate than no human could.”

2026 is the year when AI's impact becomes undeniable and systemic across government, workforce, and geopolitics—this is not a distant prediction but an imminent threshold when the conversation about AI changes fundamentally.

“circle 2026 on your calendar. This is the year AI hits everything from the workforce to government to geopolitics. You've been you've been see it already in the first month of the year. So, by the end of the year, we're going to have a completely different conversation.”

Friedman defends his trust-architecture proposal against accusations of naiveté by asking his critics to explain how AI trade and deployment will work without trust architecture—implying no viable alternative exists.

“I come along and I make this argument, and they say Fred, you're so naive. I mean trust architecture with China. Um they may call me anything but not naive. Because you tell me how this is going to work without it. Now that's naive. Okay.”

Friedman describes himself as 'a really good acorn collector'—he sees patterns ('acorns') in the world, collects them, and then works with thought partners ('jazz together') to synthesize them into larger arguments

“I not very smart. I didn't score high in my SATs, but I'm a really good acorn collector. And I see things and I collect acorns and then I sit down with my thought partner and we we do jazz together. We we break the acorn, you know, find you know, whatever.”