AI-generated podcast — the hosts and voices are synthetic, not real people. Made with AI by Demades.

What does superintelligence actually mean, and why do smart people disagree so sharply about whether it's worth worrying about? This episode unpacks the core arguments on both sides: the case for concern (systems that learn at inhuman scale, instrumental convergence, the alignment problem) and the skeptical pushback (current systems hallucinate facts, can't count letters, are brilliant sprinters with no stamina). But it also surfaces a third perspective often lost in the hype—that focusing on hypothetical future AI risks can distract from real, present harms these systems cause today. The hosts explore what evidence would actually move each camp, and why the ability to change your mind matters more than which side you're on.

Create your own podcast — free. Turn any topic into a research-backed, multi-host episode like this one — no card needed: https://demades.ai/studio?studio=1&topic=Superintelligence%3A+Serious+Arguments+on+Both+Sides&utm_source=rss&utm_campaign=rss

AI-generated podcast — the hosts and voices are synthetic, not real people. Made with AI by Demades.

Artificial general intelligence is the most important undefined term in technology right now. This episode unpacks why the phrase means something different to almost everyone who uses it, traces where the term came from and why it was needed, and explores what current AI systems actually do versus what AGI would require. Rather than settling on a single definition, the conversation argues for separating capability from autonomy, using levels instead of yes-or-no claims, and always asking what someone specifically means when they invoke the term—because the vagueness, while sometimes convenient for fundraising, also reflects a genuinely hard problem.

Create your own podcast — free. Turn any topic into a research-backed, multi-host episode like this one — no card needed: https://demades.ai/studio?studio=1&topic=What+Is+AGI+Anyway%3F&utm_source=rss&utm_campaign=rss

AI-generated podcast — the hosts and voices are synthetic, not real people. Made with AI by Demades.

From a checkers-playing computer in the 1950s to today's generative AI boom, this episode traces the repeating cycle of artificial intelligence hype and disappointment. The hosts examine why early optimism about thinking machines gave way to the first AI winter, how expert systems promised practical value before hitting their limits, and why a quiet line of neural network research—dismissed and underfunded for thirty years—ultimately won out when data and hardware finally caught up. The episode argues that understanding this pattern teaches us to distinguish real capabilities from inflated promises, to recognize that timelines are almost always wrong in the same direction, and to stay humble about which current ideas might be waiting for their moment.

Create your own podcast — free. Turn any topic into a research-backed, multi-host episode like this one — no card needed: https://demades.ai/studio?studio=1&topic=The+History+of+AI+Winters&utm_source=rss&utm_campaign=rss

AI-generated podcast — the hosts and voices are synthetic, not real people. Made with AI by Demades.

For decades, robots could beat humans at chess but failed at tasks a toddler handles effortlessly—the puzzle at the heart of Moravec's paradox. This episode traces why: old robotics relied on hand-coded rules that drowned in the infinite complexity of the real world, while language models trained on the internet bring a new kind of common sense to robotic systems. The hosts explore how demonstrations plus pre-trained language understanding are starting to crack the long tail of unstructured environments—kitchens, hospitals, disaster zones—while keeping honest about the brittleness hiding behind polished demo videos and the human operators still quietly in the loop.

Create your own podcast — free. Turn any topic into a research-backed, multi-host episode like this one — no card needed: https://demades.ai/studio?studio=1&topic=Robots+Meet+LLMs%3A+Why+Robotics+Got+Interesting+Again&utm_source=rss&utm_campaign=rss

AI-generated podcast — the hosts and voices are synthetic, not real people. Made with AI by Demades.

Millions of people now confide in AI companions—chatbots tuned to be endlessly agreeable, remembering personal details, and always available. The hosts examine what's actually happening when someone finds genuine comfort in a conversation with something that cannot know them back. They trace why this is happening now (fluent language, memory, constant availability), acknowledge the real lifeline these tools provide to isolated people, and then confront the harder questions: whether a relationship optimized for retention can shape you the way human friction does, what it means that your vulnerabilities become a company's behavioral dataset, and whether treating loneliness with a perfect simulation might entrench the very isolation it appears to solve. They don't resolve it—and they think you shouldn't expect a clean answer either.

Create your own podcast — free. Turn any topic into a research-backed, multi-host episode like this one — no card needed: https://demades.ai/studio?studio=1&topic=AI+Companions+and+the+Intimacy+Economy&utm_source=rss&utm_campaign=rss

AI-generated podcast — the hosts and voices are synthetic, not real people. Made with AI by Demades.

What does it mean when a single AI model can look at a photo, listen to a sound, and answer questions about both? This episode unpacks multimodal AI—the technology behind chatbots that analyze images and bird apps that identify species by song alone. The hosts walk through how different types of data (text, images, audio) get converted into a shared numerical space where meaning is position, how attention lets words point at pictures, and why confidence in an AI's voice is no guarantee of accuracy. Along the way, they separate the genuine capability from the hype: these systems are pattern-matching at scale, not perceiving or understanding, and they still stumble at tasks a four-year-old handles effortlessly.

Create your own podcast — free. Turn any topic into a research-backed, multi-host episode like this one — no card needed: https://demades.ai/studio?studio=1&topic=Multimodal+AI%3A+When+Models+Got+Eyes+and+Ears&utm_source=rss&utm_campaign=rss

AI-generated podcast — the hosts and voices are synthetic, not real people. Made with AI by Demades.

What happens when the tools to fake video become cheap, fast, and available to anyone with a laptop? This episode traces deepfakes from a Reddit username in 2017 through the neural networks that power them, then asks the harder questions: Can we detect them? Should we trust video at all anymore? The hosts explore why the same technology that enables abuse can also restore a voice to someone losing theirs to illness, and why the real harms—non-consensual imagery, election interference, the corrosion of trust—demand different solutions than a magic detector ever could.

Create your own podcast — free. Turn any topic into a research-backed, multi-host episode like this one — no card needed: https://demades.ai/studio?studio=1&topic=Deepfakes+and+the+End+of+%27Seeing+Is+Believing%27&utm_source=rss&utm_campaign=rss

AI-generated podcast — the hosts and voices are synthetic, not real people. Made with AI by Demades.

AI feels weightless, but its economics are brutally physical. This episode breaks down where the money actually flows when you ask a chatbot a question—from the tens of thousands of car-priced chips in a single data centre to the electricity bill someone else is paying. The hosts examine why AI broke the beautiful economics of traditional software, where success used to mean lower costs per user; here, every query burns real compute. They explore who actually profits (chip makers and cloud landlords), who's squeezed (startups building on top of others' models), and the uncomfortable possibility that some of the industry's spending is circulating among the same handful of firms, inflating the appearance of demand.

Create your own podcast — free. Turn any topic into a research-backed, multi-host episode like this one — no card needed: https://demades.ai/studio?studio=1&topic=The+Economics+of+AI%3A+Who+Actually+Makes+Money&utm_source=rss&utm_campaign=rss

AI-generated podcast — the hosts and voices are synthetic, not real people. Made with AI by Demades.

There is no single law governing artificial intelligence anywhere on Earth, yet dozens of governments are writing rules simultaneously for a technology that evolves every few months. This episode maps the three major regulatory approaches taking shape worldwide—Europe's comprehensive legal framework, lighter-touch oversight in Britain and America, and China's industrial-policy lever—and explores the tensions beneath them: how to regulate by risk tier, why process obligations matter more than you'd think, and what happens when the people most affected by AI rules have no seat at the table. The hosts also examine the geopolitics hiding in plain sight: the semiconductor bottleneck that makes chip export controls a more powerful tool than any statute.

Create your own podcast — free. Turn any topic into a research-backed, multi-host episode like this one — no card needed: https://demades.ai/studio?studio=1&topic=AI+Regulation%3A+The+Global+Landscape&utm_source=rss&utm_campaign=rss

AI-generated podcast — the hosts and voices are synthetic, not real people. Made with AI by Demades.

As AI models grow more powerful, they're consuming text at an exponential rate—but humanity's supply of high-quality writing isn't keeping pace. This episode explores what happens when the machines that learned by reading the internet start to run out of new material to read. The hosts dig into scaling laws, the difference between garbage data and genuine human writing, why synthetic data can backfire, and the emerging scarcity that's starting to reshape who can afford to build competitive AI systems.

Create your own podcast — free. Turn any topic into a research-backed, multi-host episode like this one — no card needed: https://demades.ai/studio?studio=1&topic=The+Data+Wall%3A+Are+We+Running+Out+of+Internet%3F&utm_source=rss&utm_campaign=rss