Useful tech, weird internet, steady skepticism.
Dave & Gunnar Show follows recurring speakers through the weird edges of modern technology. Recent episodes move from AI-assisted coding and agent workflows to birding apps, prison-call surveillance, adversarial poetry, URL lengtheners, and CD players worn on the head.
PRESS PLAY
Dave & Gunnar Show is a technology hangout built around curiosity, skepticism, and a steady tolerance for internet weirdness. Its recent episodes keep returning to AI agents. The speakers compare the promise of abundant code with the reality of maintainers overwhelmed by AI-generated pull requests, bug reports, and security noise. They talk about personal automation in grounded terms: inbox agents, calendar and task skills, workplace support bots, proposal-review bots, and the effort required to configure an AI environment well enough to matter. The show is enthusiastic, but not credulous. A useful assistant can improve quality of life, and the same ecosystem can leak API keys, burn tokens, or create strange agent-only social spaces. That tension gives the conversation its shape. One minute it is about top-line growth versus bottom-line efficiency. The next it is about adversarial poetry, AI personality judgments from a tame band photo, or a bot that wants encrypted messaging. The show also ranges beyond AI. Episodes touch open source history, prison phone monitoring, code reviews, birding culture, half-marathon training apps, low-background steel, nuclear command history, haptic suits, SQL games, URL lengtheners, and wearable CD players. The pacing is informal and associative. Jokes land beside serious questions about surveillance, labor, vulnerability, and who benefits when software gets cheaper or stranger. It is not a polished briefing show, and that is part of the appeal. Dave & Gunnar Show feels like listening to technically fluent people follow a thread until it becomes useful, funny, or unsettling. It fits listeners who want AI and security talk with context, not hype; practical examples, not grand abstractions; and enough odd ephemera to keep the machinery human.
Made for: Best for technically curious listeners who follow AI agents, open source, security, and workplace tooling. It also suits people who enjoy conversational detours into strange gadgets, media, and internet artifacts.
What sets it apart: The show pairs hands-on AI and infrastructure talk with a habit of chasing the odd consequence or comic edge. It is less about tidy analysis than watching tech minds test new tools against incentives, ethics, and daily life.
David Egts and Gunnar Hellekson discuss privacy, security, robots, and internet ephemera.
Based on 6 episodes we listened to · September 2026
Dave and Gunnar test AI's practical promise against its security risks, human consequences, and capacity for internet-fueled absurdity.
Dave and Gunnar treat emerging technology as something to try, question, and laugh at before declaring it useful. Their conversations repeatedly return to agentic AI through practical experiments: building an email assistant, creating workplace bots, connecting skills to…
Birdwatching gives way to AI-monitored prison calls; your appetite for digressions should be healthy.
Press play if you want
Short reviews from the PlayNext desk, based on the episodes we processed.
The discussion becomes clearest when the bot workflows are tied to real work.
Gunnar and Mike McGrath trace a shift from AI as abundance to AI as friction, weighing early hopes that AI could write all the code and energize open source against curl refusing AI pull requests, bug bounties that do not want more of them, and the feeling that too much AI can…
A loose tech chat that turns AI weirdness into a joke with teeth.
The episode starts with Soren pulling the family into Stranger Things, then swings into a horror-heavy installment that leaves everyone with nightmares before settling into a looser run of AI systems and the ways people try to bend them.
Skills, MOLT book, and a security note.
Dave and Gunnar spend much of the episode unpacking skills, the plain-English instructions that guide LLMs through tasks and tools like Todoist, calendars, and email.
Birding obsession, prison-call AI, and a code-review joke keep the episode sharp and lightly unruly.
The episode moves from birding into the kind of obsessive tracking that gives the topic its charm, with Gunnar laying out the full conversion: lists, photographs, special trips, and sightings logged on eBird.
A practical email agent shares space with a CD player headband, SQL Doom, and deliberately hostile URLs.
The strongest tech thread is the personal email agent. After weeks of prototyping in Claude Code, the host lands on a Google Apps Script that runs hourly, with Gemini sorting inbox messages and judging what needs attention.
A practical look at AI agents, APIs, and the workflows they are meant to make less messy.
Carolyn Ford and Dave X keep the focus on practical AI adoption in 2025, especially the systems that have to sit around the model for it to be useful.
Reality, examined past the shortcut.
Global news, sharpened into consequence.
Tech power, counted in costs and constraints.
TED's stage, one speaker at a time.
Philosophy with technical consequences.
Carolyn Ford and Dave X discuss five AI predictions for 2025 and how to build intelligent AI agents that deliver results.
A discussion about the impact of AI on open source communities, featuring Mike McGrath from Red Hat.
The hosts discuss watching Stranger Things with family and the resulting nightmares, then pivot to a joke about turning into an MCP server.
Dave and Gunner discuss the revolutionary potential of AI skills that guide LLMs through tool use.
Dave discusses his new hobby of birding and recommends a documentary about two brothers attempting a big year in birding despite having no prior experience.
Dave and Gunnar connect the Trinity test’s radioactive legacy to the idea of preserving pre-ChatGPT “low background content.” They explain how “verbalized sampling” can make AI responses less repetitive without adjusting temperature controls.
Hosts David and Gunnar of the Dave & Gunnar Show open by discussing David's newly working email-triage agent built with Claude Code and Gemini, then cover an Anthropic paper on 'subliminal learning,' where fine-tuned models inherit unrelated traits like a preference for owls from a teacher model.
Dave and Gunnar connect smart glasses, AI-generated games, and hidden review prompts to questions about privacy and AI behavior.
Carolyn Ford and Dave X discuss five AI predictions for 2025 and how to build intelligent AI agents that deliver results.
The host discusses using the Runa app to train for a half marathon, highlighting its adaptive scheduling and personalized plans.
Gunnar and host discuss attending a Magnetic Fields concert, their album 69 Love Songs, and the nostalgia of seeing them on a farewell tour 20 years later.
Two hosts discuss the abrupt transition from spring to summer in Texas and upcoming weather-related challenges.