Uncomfortable Conversations with Josh Szeps

Among the first responses by Iranian state media to the US-Israeli war on Iran was a propaganda video. It flaunted row upon row of gleaming Iranian drones, safely lined up in an underground weapons cache, ready to strike Israel, Arab states and US bases.

Drones. Precision-Guided Munitions. A.I. war games. Autonomous Weapon Systems. At the Pentagon, at Anthropic, for Trump and in Iran, they're redefining warfare in real time.

When the Pentagon's A.I. partner, Anthropic, insisted its systems mustn't be used to spy on Americans or to build killer robots, President Trump baulked. On Friday, Trump directed every federal U.S. agency to stop working with Anthropic, and the Pentagon declared Anthropic to be a "supply-chain risk" - a designation normally reserved for companies in enemy nations, which would bar even private defence contractors from using Anthropic's A.I. Its competitor, OpenAI, stepped in and took the Pentagon contract instead.

As conflict spreads across the Middle East, how is artificial intelligence being used? How will these fights change in the near future? Can we control it? Toby Walsh thinks so. 

He's the Chief Scientist at the UNSW A.I. Institute and a leading voice in the global regulation of A.I. weapons. He studied theoretical physics and mathematics at Cambridge University, has a Ph.D. in artificial intelligence, and was the editor-in-chief of the Journal of Artificial Intelligence Research.

Toby stopped by the Uncomfortable Conversations studios on his way to the airport to fly to Geneva to participate in a United Nations conference about A.I. in warfare.

How do you learn things? Does it come easily or is it hard? Are you drawn to e-readers, or books? Typing, or handwriting? Duolingo, or moving to a Tuscan villa for six months to drink wine and debate epicureanism?

We’re all alert to the problem of smartphones, screentime, kids and social media. But what if the problem is bigger? What if our brains and bodies are fundamentally hardwired to learn in ways that technology cannot reproduce?

Today’s uncomfortable question is whether educational technology - Chromebooks and tablets and whizz-bang learning apps - are ruining education for students and grown-ups alike. That’s what the data says, according to the neuroscientist and educator Jared Cooney Horvath. He does education-related brain and behavioural research as the director of the Science of Learning Group and of NeuroEducation. He’s lectured and researched at Harvard University and been published in the New Yorker, the Economist, the Atlantic, the New York Times, Scientific American and New Scientist.

Jared’s new book is “The Digital Delusion: How Classroom Technology Harms Our Kids’ Learning -- And How To Help Them Thrive Again”. He joins Josh to wrestle with the problems of multitasking, rote learning, memorising, the transferability of skills, the fragility of Gen Z, and how you can learn better.

Military purges. DeepSeek AI. Social credit scores. Taiwan. 

Why has the Chinese leader purged five of the seven members of the military's top decision-making body (one of the remaining two members being himself)? 

Next month, the Chinese Communist Party will enact its fifteenth Five-Year Plan. It's a biggie, because 2027 is the centenary of the People's Liberation Army - a date when China has declared it will be ready for major military combat. 

Bill Bishop is the most-read China analyst in the world. A former media guy who wrote the New York Times Dealbook's China Insider column, he became an entrepreneur and was the media's go-to China analyst during Covid. He launched one of the first Substacks ever, a fascinating blog about China called Sinocism, which now has hundreds of thousands of subscribers. 

Bill and Josh step back to ask: Will China own the 21st century? Will it eat our lunch, or slowly deflate? Happy lunar new year, humans.

Imagine a world where vast solar arrays produced cheap, abundant energy for everybody... while also powering hyper-scaled superintelligent A.I. data centres, as well as factories that suck carbon from the atmosphere to make cheap natural gas from nothing but sunlight and air.

It's within our grasp. Solar is the cheapest form of energy. The only impediments are regulations that were written in the 1970s and our own ideological hang-ups and squabbles about energy. In fact, Europe could already be energy-independent from Russia if they'd decided to invest a fraction of their gigantic gas expenditures on solar arrays when Ukraine was invaded.

Casey Handmer is an Aussie in California who got his PhD in Theoretical AstroPhysics at Caltech. He's worked at Elon Musk's Hyperloop One and at NASA's Jet Propulsion Lab. Now he's the founder of Terraform Industries, which plans to bring cheap, carbon-neutral hydrocarbons to everyone on Earth, while displacing drilling as the cheapest energy source.

He joins Josh for an inspiring, freewheeling conversation about solar, nuclear, coal, Russia, China, defense, artificial intelligence... and why a truly extraordinary future of abundance is just around the corner, if we choose to take it.

How much will A.I. change your life in the next few years? In what ways? How will you navigate the upheavals? How will the society around you do so?

"Normies" are finally waking up, it seems, to the scale of what's happening. The dam broke this week in part thanks to an essay which went viral by the A.I. founder and investor Matt Shumer, titled ‘Something Big is Happening.’ It's a clarion call for everyone to pay attention after two A.I. giants released their latest models this month. The new tech leaves no doubt that things are going to get very weird, for all of us, fast.

The neuroscientist Professor Joel Pearson works at the coalface of how humans can be psychologically and practically resilient for the social, economic and political impact of A.I. He runs the Future Minds Lab at UNSW and is Deputy Director of Human Readiness at the UNSW A.I. Institute. Joel joins Josh live on Substack to explain the recent developments, paint a picture of what the near-term looks like, and guide us on how us humans can adapt to, and survive, the coming A.I tsunami.

Read Matt Shumer’s essay here:

https://x.com/mattshumer_/status/2021256989876109403