The Job AI Was Supposed to Kill

In 2016, Geoffrey Hinton, one of the inventors of modern AI, predicted that AI would replace radiologists within years. Ten years later, we have a shortage of radiologists. AI has yet to revolutionize the field. What happened, and what does the story of AI and radiology tell us about AI and the future of work? Saurabh Jha is a radiologist at the University of Pennsylvania. He's written about how he uses AI in his practice, and why radiologists have been slow to adopt AI. Saurabh has also traveled to India and Nepal to understand the very different ways AI is being adopted for radiology in the developing world. In this episode, Saurabh explains:  Why AI still hasn't made radiologists faster or more productive The risks of novice radiologists using AI versus the benefits to experts Why a successful AI model would need to be nearly perfect  How AI is being used to spot tuberculosis in India Why diagnosing a normal scan is harder than diagnosing disease Connect with us: Follow Jacob Goldstein on LinkedIn, X and Instagram  Email us at [email protected]  Follow Pushkin on Instagram, LinkedIn or X  Listen to Jacob’s other show, Business History  To listen to the show early and ad free, sign up for Pushkin+ See omnystudio.com/listener for privacy information.

After spending decades as marginal players in the global auto industry, Chinese carmakers have become the world’s biggest exporters of cars. Chinese cars are some of the best cars on the planet, and they’re available almost everywhere – except in the US. Michael Dunne is the CEO of Dunne Insights, a company that consults on the future of cars. Michael has been working with the Chinese auto industry since 1990. I talked with Michael about how China did it. The problem we discussed was: How can the United States catch up with the Chinese auto industry? In this episode, Michael explains:  How China went from making terrible cars in 1990 to dominating global exports today Why China bet big on electric vehicles when nobody wanted them How Tesla's arrival in China revolutionized their EV industry  Why losing the car industry would be a national security problem for the US Why the future of a successful American auto industry is on the West Coast Connect with us: Follow Jacob Goldstein on LinkedIn, X and Instagram  Email us at [email protected]  Follow Pushkin on Instagram, LinkedIn or X  Listen to Jacob’s other show, Business History  To listen to the show early and ad free, sign up for Pushkin+ See omnystudio.com/listener for privacy information.

Phages are viruses that attack bacteria. People have been using phages to treat infections for over a hundred years, but in most cases they just don't work as well as antibiotics.  Paul Garofolo is the co-founder and CEO of Locus Biosciences. His problem is this: How do you genetically engineer phages to be more effective at curing disease?  In this episode, Paul explains:  Why wild phages never completely wipe out a bacterial infection How to engineer phages with CRISPR to kill bacteria more effectively The clinical trial testing phage therapy for chronic, recurrent UTIs Why persistent bacteria may contribute to autoimmune conditions like Crohn's disease How to use robots and AI to create phage cocktails from wastewater Connect with us: Follow Jacob Goldstein on LinkedIn, X and Instagram  Email us at [email protected]  Follow Pushkin on Instagram, LinkedIn or X  Listen to Jacob’s other show, Business History  To listen to the show early and ad free, sign up for Pushkin+ See omnystudio.com/listener for privacy information.

David Autor is a labor economist at MIT. He has written about how the arrival of personal computers increased inequality among American workers. But when it comes to AI, David thinks there could be a different outcome.  Read David’s article about AI in Noema Magazine.  In this episode, David explains:  How computers increased the wage gap between college and non-college educated workers Why AI’s ability to supplement expertise may open up expert fields to more people  How AI could make healthcare and legal services cheaper What fields will still experience rapid devaluation of skills due to AI  The “no regrets” policies that the US could enact to help ease the impact of AI   Connect with us: Follow Jacob Goldstein on LinkedIn, X and Instagram Email us at [email protected]  Follow Pushkin on Instagram, LinkedIn or X  Listen to Jacob’s other show, Business History  To listen to the show early and ad free, sign up for Pushkin+ See omnystudio.com/listener for privacy information.

John Mills is the CEO and co-founder of a nonprofit called Watch Duty. His problem is this: How do you build an app to warn people when they are in immediate danger from a natural disaster? Watch Duty has millions of users and played a key role in the Los Angeles fires of 2025. In the show, John talks about how he built his app with the help of an army of volunteers, and why the government hadn’t already built something like Watch Duty. In this episode, John explains:  Why one-quarter of Los Angeles residents downloaded Watch Duty during the 2025 fires Why the all-clear message matters almost as much as the initial warning How they are expanding into flood warnings How to decode firefighter’s radio chatter, and his favorite bit of jargon  Connect with us: Follow Jacob Goldstein on LinkedIn, X and Instagram Email us at [email protected]  Follow Pushkin on Instagram, LinkedIn or X  Listen to Jacob’s other show, Business History  To listen to the show early and ad free, sign up for Pushkin+ See omnystudio.com/listener for privacy information.

Josh Tyrangiel is the author of “AI for Good: How Real People Are Using Artificial Intelligence to Fix Things That Matter.” On today’s show, Josh talks about some of these real people and the problems they’re working on: a professor trying to understand her nonverbal son, a general who delivered millions of Covid vaccines in 2021, and a hospital CEO trying to reduce the rate of deadly infections.  In this episode, Josh explains:  Where AI is helping to reduce hospital infections (and where it’s struggling) Why distributing the Covid vaccine was almost as hard as creating it Why convincing people to use AI is often the hardest problem to solve  Why we should use nuclear weapons as a model for regulating AI Connect with us: Follow Jacob Goldstein on LinkedIn, X and Instagram Email us at [email protected]  Follow Pushkin on Instagram, LinkedIn or X  Listen to Jacob’s other show, Business History  To listen to the show early and ad free, sign up for Pushkin+ See omnystudio.com/listener for privacy information.

Betsey Stevenson is a labor economist at the University of Michigan, and she was an economic adviser to President Obama. Betsey’s problem is this: How can we create a world where the benefits of AI are broadly shared? Betsey draws on history – including how the invention of household appliances created a crisis of meaning for American women – to understand how we should respond to the challenge of AI. And she suggests policies to help spread the wealth AI could bring. In this episode, Betsey explains:  How Engels’ Pause serves as a warning for workers How 20th century women adapted to automation How AI has changed life for college students The argument for taxing AI firms and distributing the proceeds to the public Connect with us: Follow Jacob Goldstein on LinkedIn, X and Instagram Email us at [email protected]  Follow Pushkin on Instagram, LinkedIn or X  Listen to Jacob’s other show, Business History  To listen to the show early and ad free, sign up for Pushkin+ See omnystudio.com/listener for privacy information.

Maps have gotten much better over the past few decades. But they're still mostly two dimensional, and they struggle to keep up with a world that is always changing. Brian McClendon is the Chief Technology Officer of Niantic Spatial, a spinout of the company that makes Pokémon Go. Brian's problem is this: How do you build a three-dimensional map of the world that both robots and humans can use to find their way? In this episode, Brian explains:  What he learned building Google Earth  How Pokémon Go players helped build the foundation for robot navigation What it will take to teach AI to understand the physical world. Why GPS is so bad in cities   Why trees are such a big problem for digital mapmakers Connect with us: Follow Jacob Goldstein on LinkedIn, X and Instagram Email us at [email protected]  Follow Pushkin on Instagram, LinkedIn or X  Listen to Jacob’s other show, Business History  To listen to the show early and ad free, sign up for Pushkin+ See omnystudio.com/listener for privacy information.

Dan Shipper is the co-founder and CEO of Every, a company that publishes newsletters about AI, develops AI-related software, and helps other companies use AI. Dan has two problems. One, how do you build a company where almost everybody has their own AI agent? And two, how do you use AI as a tool to improve your writing, rather than as a replacement for writing? In this episode, Dan explains:  Why managing agents will be a core skill of the future Why AI agents won’t replace you anytime soon  How writers learn faster than AI models  How to give AI the social skills to work in group settings  How AI can help you better understand yourself Check out Every at their website: https://every.to/ Connect with us: Follow Jacob Goldstein on LinkedIn, X and Instagram Email us at [email protected]  Follow Pushkin on Instagram, LinkedIn or X  Listen to Jacob’s other show, Business History  To listen to the show early and ad free, sign up for Pushkin+ See omnystudio.com/listener for privacy information.

The apples you can buy at the grocery store have gotten profoundly better over the past few decades. It’s a kind of everyday, hiding-in-plain sight innovation.  Kate Evans is an apple breeder and professor at Washington State University. Kate's problem is this: How do you invent a better apple? With her team, Kate has in fact invented a new kind of apple called the Sunflare. It’s arriving in stores in a few years. She also helped to invent the Cosmic Crisp, which is out already. In this episode, Kate explains:  Why grocery-store apples have gotten so much better Why some of the most delicious apples never make it into stores  Why it takes decades to invent a new kind of apple How much more we have to learn about apple genetics  How the apple industry is adapting to climate change  Connect with us: Follow Jacob Goldstein on LinkedIn, X and Instagram Email us at [email protected]  Follow Pushkin on Instagram, LinkedIn or X  Listen to Jacob’s other show, Business History  To listen to the show early and ad free, sign up for Pushkin+ See omnystudio.com/listener for privacy information.