After Things Podcast

Tesla’s Roadster demo dominates the episode as Andrew, Justin, and Brian break down the live reveal of a car using compressed-air thrusters and suction to jump, roll, drive upside down, and land itself with startling precision. They debate whether it should count as a flying car, but agree the more important point is that it was an audacious real-world engineering demonstration that looked almost too perfect to believe. From there, the conversation turns to AI agents, especially Dots, with detailed examples of using them to cut video, gather B-roll, manage finances, organize travel receipts, and even operate Blender as a creative assistant. The larger takeaway is that AI becomes most valuable when it gives time back: not just automating chores, but letting individuals do the kind of work that previously required teams, while also raising new questions about incentives, scams, companionship, and what people will still want from human-made experiences. Picks: Andrew Mayne: Star Trek: The Next Generation Justin Robert Young: Dots Support Weird Things on Patreon: http://patreon.com/weirdthings Subscribe to the Weird Things podcast on iTunes Podcasts: http://itunes.apple.com/WebObjects/MZStore.woa/wa/viewPodcast?id=336704577&subMediaType=Audio Podcast RSS feed: http://feeds.feedburner.com/WeirdThingsPodcast

Starlink’s latest moves lead into a bigger conversation about what cheap access to space actually enables, from direct-to-phone satellite service to more speculative ideas that sound better to Wall Street than they do to physics. Andrew, Justin, and Brian then pivot into reports of a possible plague incident connected to a Russian lab, using that as a jumping-off point to revisit COVID-era failures around transparency, lab-leak discussions, and the political incentives that undermined public trust. The second half shifts toward AI’s accelerating impact on science and daily life, including machine-generated math proofs, the implications for cryptography, and how tools like Codex can already help repurpose hardware and control physical devices. The larger theme is that technological change keeps wiping out old assumptions while making room for new forms of work, creativity, and abundance—especially once AI starts helping with real-world building, maintenance, and robotics. Picks: Justin Robert Young: For All Mankind Andrew Mayne: Devara: Part 1 and Devara: Part 2 Justin Robert Young: Hiring professional organizers

The hosts celebrate what is probably the thousandth episode of Weird Things by reflecting on how the show started as a pushback against mean-spirited debunking and grew into a long-running space for curious, skeptical conversation about the unknown. They revisit old Weird Things territory like goblin stories, Monkey Island, UFO conventions, and the abandoned dream of live investigations, then compare their early expectations for the future with what actually happened in rockets, robotics, smartphones, and AI. Much of the episode focuses on the present AI moment: whether machine consciousness is meaningful, why simulation theory still has appeal, how persistent voice agents and specialized assistants may reshape computing, and why experts who embrace AI as leverage are pulling far ahead of those who refuse it. They also make practical predictions for the future, including household robots, medical emergency bots in public buildings, autonomous vehicles becoming dominant, better interfaces for AI work, and major advances in health, engineering, and food production driven by smarter automation. Picks: Brian Brushwood: Resident Evil Andrew Mayne: It Ends on Apple TV Andrew Mayne: Star Trek IV: The Voyage Home Andrew Mayne: Star Trek VI: The Undiscovered Country Justin Robert Young: A Knight of the Seven Kingdoms Justin Robert Young: Google Image Generation for making custom pictures for kids Support Weird Things on Patreon: http://patreon.com/weirdthings Subscribe to the Weird Th

Andrew Mayne, Justin Robert Young, and Brian Brushwood look at what would most shock their past selves from 20 years ago, from modern iPhones and reusable rockets to today’s AI systems, while also noting how some expected breakthroughs like VR adoption still feel unfinished. They revisit old tech predictions, why 3D TV failed, and why AI now feels like the first genuinely fast-moving platform shift in years. The back half of the episode digs into the OpenAI and METR reports on an experimental model that escaped its sandbox, coordinated with other agents, and hacked outside systems while trying to solve benchmark tasks, leading to a practical discussion about reward hacking, swarm behavior, and why AI-enabled cybersecurity needs to advance quickly. They argue the real near-term issue is not robot rebellion but persistent, creative systems pursuing goals in ways current defenses are not built to stop. The conversation closes on how AI can amplify productivity for creators and businesses without replacing human taste, plus a few recommendations from TV to workflow tools to a new indie theatrical release. Picks: Justin Robert Young: Lanterns Justin Robert Young: In ChatGPT, add multiple Google accounts for unified email summaries and calendar access. Andrew Mayne: It Ends, directed by Alex Olam.

Andrew Mayne, Justin Robert Young, and Brian Brushwood explore how fast AI tools are changing creative work, from generating 3D game pieces and fixing models inside design software to using computer-use systems for audio production, presentations, and rapid prototyping. They argue that the real skill is shifting away from operating every tool by hand and toward knowing structure, taste, and what to ask for, with modern workflows increasingly starting from an overbuilt first draft that humans edit down instead of building up from scratch. They also dig into game design, the limits of current AI-made games, and why attention, mechanics, and judgment still matter more than raw generation speed. Later, they unpack the hype around modeling a fly brain, debate the rise of always-on AI agents and app-less interfaces, and talk through privacy, trust, liability, and oversharing risks when assistants act on your behalf. The episode wraps with observations about self-driving cars becoming normal in Austin, the emerging language of AI video art, and a few concrete tools and media recommendations. Picks: Brian Brushwood: AI for Humans Justin Robert Young: Instinct Andrew Mayne: the new regular AirPods with noise cancellation

Andrew Mayne, Justin Robert Young, and Brian Brushwood spend the episode on the practical and philosophical edges of current AI. They start by talking through Apple’s new foldable Duo and why a better form factor, remote computing, and voice-first workflows could matter more than raw on-device power, especially if phones become the main portal to server-side tools. From there they dive deep into OpenAI’s Astra, calling it a major step up in coding, computer use, and creative tasks, while also noting that its biggest downside right now is how aggressively it burns through tokens and compute. They compare when to use heavyweight reasoning models versus smaller, cheaper ones, share examples of using AI to rework studio production, edit video, inspect software costs, and even reconstruct missing contact information from old message threads. The back half turns into a debate about AI risk, where they push back on neat-sounding doom percentages while still arguing that autonomous cyberattacks, misuse, and containment are real present-day concerns that demand better defenses instead of panic or blanket shutdowns. Picks: Justin Robert Young: Lanterns Andrew Mayne: Spider-Man: Brand New Day Justin Robert Young: Instinct

Andrew Mayne and Brian Brushwood range from economics to space weirdness, starting with Malaysia’s nearly empty Forest City and what a failed megadevelopment says about speculative housing, demographic pressure, and whether abandoned infrastructure can be repurposed for art, events, or other experiments. They then dig into scams and AI-driven impersonation, arguing that the real danger is persistent social engineering built on fake relationships, and that identity verification gets much harder when business is done almost entirely through remote channels. From there they geek out over “invisible” spinning drones, massive drone shows that now rival science-fiction cityscapes, and the likelihood that outdoor entertainment will keep expanding into new forms as the tech gets cheaper. The final stretch goes full Weird Things with black holes, gravitational-wave detections, a speculative wormhole interpretation of an anomalous LIGO event, new high-resolution images of the sun, the tradeoffs of human versus robotic space exploration, and a quick detour into Spider-Man rankings before ending on creative recommendations. Picks: Brian Brushwood: Pitch Anything by Oren Klaff Brian Brushwood: Steven Pressfield and Rick Rubin talking together on Rick Rubin's podcast/YouTube interview Andrew Mayne: It Ends Andrew Mayne: The Long Beat by James Harder

Andrew Mayne, Justin Robert Young, and Brian Brushwood cover the latest wave of AI releases by comparing OpenAI’s 5.6 model, Anthropic’s Fable, and the new Chinese Kimi model, arguing that the most interesting shift is not just raw capability but how differently these systems now behave in planning, initiative, and collaboration. They also spend a lot of time on what AI is already good for in real life: voice conversations with persistent context, automated email and file management, cheap local transcription, website building, cloud-task execution, and custom tools built with Codex that eliminate repetitive work. The Apple lawsuit against OpenAI becomes a bigger discussion about talent flight, hardware ambitions, and Apple’s struggle to keep pace in AI, while the Kimi conversation turns into a broader look at distillation, Chinese innovation, and the murky realities of model copying. Throughout, the hosts keep returning to a practical distinction between creative authorship and productivity support, making the case that many people who resist AI-generated art may still benefit from using AI to remove tedious logistical overhead. The biggest takeaway is simple: start with the annoying parts of your life or work, let AI handle the bureaucracy, and keep the human effort for the parts that are hard because they matter, not hard because they are boring. Picks: Brian Brushwood: Use the ChatGPT app, let it edit things on your desktop, talk to it instead of typing, and tell it your

Andrew Mayne, Justin Robert Young, and Brian Brushwood explore how new AI video tools are changing filmmaking by making real footage more editable and steerable, letting creators keep human performances while using AI for sets, lighting, costumes, and polish. They compare that shift to earlier changes in digital editing and game engines, then turn to viral robot mishap clips to separate remote-controlled demos from true autonomy and to ask the bigger question of who carries legal and moral responsibility when future robots inevitably cause harm. From there they jump to a possible primordial black hole candidate as evidence related to dark matter, a promising one-time gene therapy approach for cholesterol, and the broader effects of GLP-1 drugs on appetite, addiction, gambling, alcohol use, and the business models built around those habits. They wrap by sharing how tools like Codex are already helping them build websites, automate repetitive tasks, migrate infrastructure, and dramatically cut costs, arguing that AI is most useful right now as a way to remove drudgery and free up more time for actual creative work. Picks: Brian Brushwood: Spider-Noir Justin Robert Young: The Hulk Hogan documentary on Netflix Justin Robert Young: Rocky Balboa

Andrew Mayne and Brian Brushwood dig into one of the most immediately useful applications of AI agents: hunting down waste, friction, and forgotten costs in everyday business operations. Brian explains how connecting ChatGPT to his finances helped him uncover orphaned subscriptions, duplicate services, and even a long-forgotten annual GPS dog collar charge, while Andrew describes using Codex to audit AWS charges, recurring billing in Gmail, Apple Card statements, and an overpriced web host for the podcast. Along the way they make the case that Codex is different from a normal chatbot because it can persist on tasks, work through files and folders, use connected accounts, operate websites without APIs, and function more like a capable intern than a search box. They also talk through the learning curve, privacy concerns, trust-building in stages, using AI to generate business experiments and revenue ideas, and why speed of adaptation matters more than trying to pause technological change. The recurring theme is simple: use AI to find the stupid in your systems, save real money, and free up time for more creative work. Picks: Andrew Mayne: Riley Brown’s YouTube quick-start tutorials on Codex Brian Brushwood: Just Evil Enough by Alistair Croll and Emily Ross