Machines Like Us

When Demis Hassabis co-founded DeepMind in 2010, his mission was simple: solve intelligence, then use it to solve everything else. What began as a scientific quest would eventually become a fierce global AI race. 

Today, DeepMind is owned by Google and is one of the world’s leading AI labs. Journalist and author Sebastian Mallaby spent years interviewing Hassabis, his collaborators, rivals and critics for his new book, The Infinity Machine. It tells the story of the remarkably small group of people who shaped modern AI.

When we recorded this conversation, Hassabis was still CEO of DeepMind, caught between the science he wanted to pursue and the commercial race Google needed him to run. Hours later, he announced he was stepping down as CEO to become Alphabet’s (Google’s parent company) chief scientist.

Hosted by Simplecast, an AdsWizz company. See pcm.adswizz.com for information about our collection and use of personal data for advertising.

You've probably seen one by now: a human-sized robot doing very human feats – landing a perfect backflip, racing through the streets of Beijing, or dancing in a crowd at a Knicks game. But behind these viral moments, there is serious money. Analysts predict the market for humanoid robots could be worth trillions in a matter of decades. 

Everyone from trillionaire Elon Musk to Chinese President Xi Jinping is betting humanoids are the future – that one day, they will live alongside us – in our homes, and even our military. The question is: can we trust them? 

Angelica Lim has spent her entire career trying to build trustworthy robots. A roboticist based at Simon Fraser University, Lim joins Machine’s Like Us to share lessons from building the world’s first mass-produced humanoid named Pepper, and why empathy in robotics is more important than ever.

Hosted by Simplecast, an AdsWizz company. See pcm.adswizz.com for information about our collection and use of personal data for advertising.

When Marc Miller became Minister of Canadian Identity and Culture, he inherited Ottawa’s oldest, unfinished file in tech policy: regulating social media. He took office at a global turning point. Australia had just banned social media for kids under 16, with Denmark, France, Turkey, and the U.K. quickly following suit. 

The proliferation of AI chatbots added urgency to his new file. Ignored in previous policy cycles, AI chatbots are now embedded in the lives of millions of Canadians. And in the wake of the tragedy at Tumbler Ridge, the risks of these unrelated systems are no longer theoretical.

In part two of our series on Canada's landmark new digital legislation, Minister Miller joins us to talk about his bill, The Safe Social Media Act, which restricts platforms for minors unless platforms can prove they’re safe and sets world-first rules for AI chatbots. It’s the federal government’s bid to hold Big Tech accountable. If you missed part one, you can check out our last episode with Minister Evan Solomon on the federal government’s plan to bring Canada’s privacy laws into the AI era.

Hosted by Simplecast, an AdsWizz company. See pcm.adswizz.com for information about our collection and use of personal data for advertising.

“AI for All.” That’s what Canada’s federal government is calling its national AI strategy. It makes the case for building AI with confidence: a homegrown supercomputer by 2031, 60 per cent adoption by 2034. But the new strategy offers little about how it will govern the technology. 

Ottawa attempted to answer those concerns with two landmark bills: the Safe Social Media Act, which regulates social media platforms and AI chatbots, and the Protecting Privacy and Consumer Data Act, which brings Canada’s privacy laws into the AI era. Together, these bills give the country something it’s never had: a single digital regulator with power to hold companies responsible. 

It’s the most ambitious tech policy move this country has seen. And it’s drawing fire from every direction. Critics say a superregulator gives too much power to the government, that privacy and data should be treated separately, and that this is part of a wider government effort to control speech.

To dig into the details, we sat down with Evan Solomon, Canada’s first Minister of Artificial Intelligence and the architect of this agenda. We talk about why data means something new in the age of AI, how this regulator would actually work, and whether it could have stopped something like Grok’s controversial nudification feature – a case the current Privacy Commissioner says he’s powerless to stop.

Mentioned:

Canada’s National Artificial Intelligence Strategy: AI for All

Backgrounder: Government of Canada introduces legislation to Protect Canadians’ Privacy in the Digital Age

Bill C-34, the Safe Social Media Act

Elon Musk's Grok AI tool under fire over digital undressing

Canada’s privacy commissioner expands investigation into ‘X’ over ‘Grok’

Yes, Musk's Grok chatbot is 'digitally undressing' women and children • FRANCE 24 English

Privacy Commissioner of Canada investigation into the Grok chatbot and sexualized deepfakes finds companies violated privacy law

Commissioner-initiated complaints concerning X Corp.’s and X.AI LLC’s compliance with PIPEDA

Canada has access to Anthropic’s powerful Mythos AI model, minister says

Hosted by Simplecast, an AdsWizz company. See pcm.adswizz.com for information about our collection and use of personal data for advertising.

Human beings have always loved to gamble. Archeological records suggest we’ve been doing it for the last 12,000 years, since the end of the last Ice Age. But for as long as we’ve been playing games of chance, we’ve worried about what they might be doing to us. For thousands of years, everyone from Aristotle to George Washington condemned gambling, an ancient anxiety that ran so deep it became something like a moral consensus.

And then that consensus evaporated. In the span of a decade, both Canada and the US legalized sports betting. Now anyone with a smartphone and a credit card can wager on basketball, hockey, or American cornhole.

But it turned out that was just the beginning. A few years later came “prediction markets” like Kalshi and Polymarket that let you bet on, well, just about anything: whether the US will invade Cuba, the odds of James Comey being sent to prison, and whether Jesus Christ will return before 2027. That last one, by the way, is currently sitting at 3 per cent on Polymarket.

If betting on missile strikes, military coups, and political prosecutions feels kind of gross, I’m with you. But James Surowiecki thinks we should give prediction markets a chance. Surowiecki is the author of The Wisdom of Crowds, a book he wrote more than 20 years ago, where he argued that large groups of ordinary people are actually better than experts at making predictions. It’s become something of a foundational text for these markets: the idea that they can crowdsource knowledge, aggregate what millions of people believe about the future, and use that signal to make better decisions. So I wanted to have James on to make the case for prediction markets, and to see if he could make me feel just a little less squeamish about a world where you can gamble on everything.

Mentioned

The Wisdom of Crowds, by James Surowiecki (Doubleday, 2004). 

Francis Galton, “Vox Populi,” Nature 75 (1907): 450–451 — the–ox-weighing experiment.

The 1986 Challenger disaster and Morton Thiokol’s stock: Maloney & Mulherin, “The complexity of price discovery in an efficient market,” Journal of Financial Economics (2003). 

Kalshi (prediction market platform).

Polymarket (prediction market platform).

The 2024 “French whale” (Théo), who used neighbour polls to bet roughly $85M on a Trump win — CBS–News / 60 Minutes.

The Polymarket trader’s well-timed bets on the June 2025 US strikes on Iran — CNN–

The market on the length of a Karoline Leavitt White House briefing

Coinbase CEO Brian Armstrong and the earnings-call “mention markets” — Tec–Crunch.

The market on Maduro’s removal and the ~$400K Venezuela payout — PBS–NewsHour.

The Zohran Mamdani NYC mayoral market — DL –ews.

The market on Bad Bunny’s first Super Bowl LX song — Pol–market.

DARPA’s Policy Analysis Market (the “terrorism futures” proposal, cancelled after backlash in 2003) — CNN–(2003).

The 1979 Iranian Revolution as a US intelligence failure — Nat–onal Security Archive, George Washington University.

Hosted by Simplecast, an AdsWizz company. See pcm.adswizz.com for information about our collection and use of personal data for advertising.