Share
Babbage from The Economist
Gary Marcus: a sceptical take on AI in 2025
•
From the release of AI agents to claims that artificial general intelligence has (finally!) been achieved, 2025 will probably be another blockbuster year for AI. That sense of continuous progress is not shared by everyone, however. Generative AI, based on large language models (LLMs), struggles with reasoning, reliability and truthfulness. While progress has been made in those domains, sceptics argue that the limitations of LLMs will fundamentally restrict the future of AI.
In this episode, Alok Jha, The Economist’s science and technology editor, interviews Gary Marcus, one of modern AI’s most energetic critics. They discuss what to expect in 2025 and why Gary is pushing for researchers to work on a much wider range of scientific ideas (in other words, beyond deep learning) to enable AI to reach its full potential.
Gary Marcus is a professor emeritus in cognitive science at New York University and the author of “Taming Silicon Valley”, a book advocating for a more responsible approach to the development of AI.
For more on this topic, check out our series on the science that built the AI revolution, as well as our episodes on AGI.
Transcripts of our podcasts are available via economist.com/podcasts.
Listen to what matters most, from global politics and business to science and technology—subscribe to Economist Podcasts+.
For more information about how to access Economist Podcasts+, please visit our FAQs page or watch our video explaining how to link your account.
More episodes
View all episodes
A tale of two outbreaks
37:49|A rare disease spreading on a cruise ship triggered memories of covid-19. But while people worried about hantavirus cases reaching their shores, a much larger outbreak of Ebola took hold in central Africa. The two outbreaks have come in the wake of a series of worrying developments for international public health agencies: America has left the World Health Organisation and the country has also gutted its spending on international aid. How should pathogenic threats be handled in this post-aid era? Guests and hosts:Slavea Chankova, The Economist’s health-care correspondentNicole Tischler of Fundación Ciencia & Vida, ChileTom Geisbert of the University of Texas Medical BranchHost: Alok Jha, The Economist’s science and technology editor Topics covered:Andes hantavirusBundibugyo EbolavirusPandemic preparednessTranscripts of our podcasts are available via economist.com/podcasts.Listen to what matters most, from global politics and business to science and technology—subscribe to Economist Podcasts+.For more information about how to access Economist Podcasts+, please visit our FAQs page or watch our video explaining how to link your account.
Maths enters its AI era
47:50|Not long ago, AI models struggled with basic arithmetic. Now, they’re solving complex mathematical conjectures. But researchers want to go further. They think AI could help them organise the mathematical literature into a common language—thereby finding connections between hitherto-disparate fields and, perhaps, even making completely new discoveries in one of the most fundamental areas of human inquiry.Guests and hosts:Anjani Trivedi, The Economist’s global business correspondentTerence Tao of the University of California, Los Angeles (UCLA)Patrick Shafto of America’s Defense Advanced Research Projects Agency (DARPA)Maryna Viazovska of L'École Polytechnique Fédérale de Lausanne (EPFL)Alok Jha, The Economist’s science and technology editor Topics covered:MathematicsLarge language modelsAutoformalisationTranscripts of our podcasts are available via economist.com/podcasts.Listen to what matters most, from global politics and business to science and technology—subscribe to Economist Podcasts+.For more information about how to access Economist Podcasts+, please visit our FAQs page or watch our video explaining how to link your account.
Michael Pollan on the mystery of consciousness
40:21|What is consciousness? It’s a question scientists have been struggling to answer for decades. Author Michael Pollan has been searching for scientific evidence that could unlock the mystery. His journey extends from humans to animals and even plants. But could consciousness ever exist in machines?Guests and hosts:Michael Pollan, author of “A World Appears”Host: Alok Jha, The Economist’s science and technology editor Topics covered:ConsciousnessSentienceNeural correlates of consciousnessTranscripts of our podcasts are available via economist.com/podcasts.Listen to what matters most, from global politics and business to science and technology—subscribe to Economist Podcasts+.For more information about how to access Economist Podcasts+, please visit our FAQs page or watch our video explaining how to link your account.
Why runners are getting faster
37:54|Sporting history was made at this year’s London Marathon. Sabastian Sawe became the first man to finish a competitive race in less than two hours. How are elite athletes getting so much faster? We investigate how runners are harnessing the science of physiology and the technology of shoes to increase their speed. Plus, how anybody can train for a marathon—or just get fit fast.Guests and hosts: Ainslie Johnstone, The Economist’s data and science correspondentTim Cross, senior science writer and author of the “Well Informed” newsletterHost: Alok Jha, The Economist’s science and technology editorTopics covered: London MarathonAdidas Adizero Adios Pro Evo 3Exercise hacksAudio and commentary taken from the BBC's live coverage of the London Marathon, April 26th 2026.Transcripts of our podcasts are available via economist.com/podcasts.Listen to what matters most, from global politics and business to science and technology—subscribe to Economist Podcasts+.For more information about how to access Economist Podcasts+, please visit our FAQs page or watch our video explaining how to link your account.
Chernobyl: a laboratory like no other
44:48|In the popular imagination, the Chernobyl nuclear-power plant sits in an abandoned, post-apocalyptic wasteland. In reality, the site of the world’s worst nuclear-energy disaster is a thriving experimentation ground for scientists studying the long-term impacts of radiation. This week, 40 years after the meltdown, we investigate what that laboratory has found so far. Guests and hosts: Jason Palmer, host of “The Intelligence” podcastOlena Pareniuk of the Institute for Safety Problems of Nuclear Power PlantsJim Smith of the University of PortsmouthAlok Jha, The Economist’s science and technology editorTopics covered: Chernobyl nuclear disasterNuclear safetyRadiation physicsTranscripts of our podcasts are available via economist.com/podcasts.Listen to what matters most, from global politics and business to science and technology—subscribe to Economist Podcasts+.For more information about how to access Economist Podcasts+, please visit our FAQs page or watch our video explaining how to link your account.
An antimatter road trip
39:08|The universe—and everything inside it—shouldn’t exist. The big bang should have produced equal amounts of matter and antimatter, and their mutual annihilation should have resulted in nothing more than a sprawling ball of energy. But, in fact, the universe is dominated by matter, while antimatter is vanishingly rare. To try to understand why, physicists have taken antimatter on a road trip—for the first time in history.Guests and hosts: Sam Wikeley, The Economist’s science correspondentApril Cridland of CERN’s antimatter factoryStefan Ulmer of HHU Düsseldorf and CERN’s BASE experimentHost: Alok Jha, The Economist’s science and technology editorTopics covered: AntimatterThe European Organisation for Nuclear Research (CERN)Particle physicsTranscripts of our podcasts are available via economist.com/podcasts.Listen to what matters most, from global politics and business to science and technology—subscribe to Economist Podcasts+.For more information about how to access Economist Podcasts+, please visit our FAQs page or watch our video explaining how to link your account.
Inside the mind of Demis Hassabis
41:53|The AI race is intensifying. It wasn’t supposed to be this way, according to Demis Hassabis, the boss of Google DeepMind. Journalist and biographer Sebastian Mallaby has had unprecedented access to Mr Hassabis over the past three years. In this episode, he tells us how the AI pioneer was swept into a commercial and geopolitical contest that he never expected. Can Mr Hassabis’s aim of safely building artificial superintelligence still be achieved?Guests and hosts:Sebastian Mallaby, author of “The Infinity Machine” Host: Alok Jha, The Economist’s science and technology editorTopics covered:Demis Hassabis and Google DeepMindAI safety and regulationChina’s open-source AI industryWatch The Economist’s recent interview with Demis Hassabis on “Inside Tech” here. Transcripts of our podcasts are available via economist.com/podcasts.Listen to what matters most, from global politics and business to science and technology—subscribe to Economist Podcasts+.For more information about how to access Economist Podcasts+, please visit our FAQs page or watch our video explaining how to link your account.
A vaccine for everything
43:31|Imagine if a single vaccine could protect against a vast array of pathogenic threats. Recent research suggests that a universal vaccine to fend off respiratory viruses, bacterial infections—even allergens—might become a reality. Such jabs would be vital for blunting the impact of the next pandemic. Guests and hosts:Ainslie Johnstone, The Economist’s data and science correspondent Mihai Netea of Radboud University NijmegenPamela Bjorkman of the California Institute of Technology Bali Pulendran of Stanford UniversityHost: Alok Jha, The Economist’s science and technology editorTopics covered:Broad-spectrum vaccinesImmunologymRNA technologyListen to Ainslie’s related reporting on dementia and infections by scrolling back to our March 2025 episode, “Going viral: could infections cause Alzheimer's?”Transcripts of our podcasts are available via economist.com/podcasts.Listen to what matters most, from global politics and business to science and technology—subscribe to Economist Podcasts+.For more information about how to access Economist Podcasts+, please visit our FAQs page or watch our video explaining how to link your account.
How AI is really being used in war
42:14|The conflict in Iran has been described as the “first AI war”. But how is artificial intelligence actually being deployed? We examine the conflicts in Iran and Ukraine, which raise some thorny questions about the mechanics and ethics of the future of war.Guests and hosts:Shashank Joshi, defence editor at The EconomistJames Lawson, a director at HelsingHost: Alok Jha, The Economist’s science and technology editorTopics covered:Military technologyAutonomous weaponsArtificial intelligenceTranscripts of our podcasts are available via economist.com/podcasts.Listen to what matters most, from global politics and business to science and technology—subscribe to Economist Podcasts+.For more information about how to access Economist Podcasts+, please visit our FAQs page or watch our video explaining how to link your account.