Approachable AI

What if AI stopped just answering your questions and started actually doing things for you?

In Episode 6 of Approachable AI, Joshua dives into the world of personal AI agents, putting Meta's new Muse assistant to the test. From turning hundreds of saved Instagram posts into a personalized Google Map to automating parts of his other podcast, The Eclectic Polymath, he's exploring what happens when AI starts taking on real responsibilities.

But how far can you actually take it? Joshua experiments with letting AI manage his calendar, track personal goals, discover new places, and even help run a game built around his everyday life. Along the way, there are unexpected discoveries, plenty of trial and error, and a few reminders that these tools still have a long way to go.

It's a hands-on look at the emerging world of AI agents, what makes them different from traditional chatbots, and what it might mean to have your own personal AI assistant.

Welcome back to Approachable AI, where I explore practical ways to use AI, test new tools, build real projects, and make emerging technology easier to understand.In this episode, I’m taking you behind the scenes of how I built "The Eclectic Polymath Podcast" using AI. Instead of focusing on one specific deep-dive topic, I’m walking through the full methodology behind the show: how a random curiosity becomes a research question, how I use multiple AI tools to expand and organize the idea, and how that research eventually becomes a complete long-form audio episode.Key topics covered:• Turning personal curiosity into a repeatable content system• Using AI for deep research and narrative development• Moving from a loose question to a structured episode• Creating AI-generated podcast hosts• Building “Charles” and “Jade” as recurring voices for the show• Using NotebookLM and other AI platforms across the workflow• Balancing automation with human taste, direction, and judgment• Why AI works best as a creative collaborator rather than a replacementThis is not a traditional AI tutorial or a polished case study. It is a real look at how I use AI to follow ideas across history, politics, music, film, spirituality, technology, business, and culture, then turn those ideas into something people can actually listen to.The bigger point is that the technology is only part of the process. AI can help research, organize, synthesize, and produce, but the curiosity, point of view, taste, and final decisions still have to come from a person.🔔 Subscribe to Approachable AI for more practical experiments, creative workflows, and honest breakdowns of how I’m using AI in real projects.💬 What topic would you turn into a deep-research podcast episode using AI?#AI #Podcasting #NotebookLM #GenerativeAI #ContentCreation #AIPodcast #ChatGPT #GoogleAI #CreativeAI #ApproachableAI

Welcome to 2026! I'm back with a raw, unfiltered look at building AI-powered apps in real-time.In this episode, I'm kicking off a 10-week AI resolution challenge inspired by the AI Daily Brief podcast. Watch me build a complete resolution tracker app using Google AI Studio, complete with project management features, voice dictation, and custom milestones.Key topics covered:• Building a resolution tracker with Google AI Studio• 10-week AI upskilling challenge breakdown• Live coding and iterative app development• Google-Apple AI partnership news breakdown• Why voice-first development is the future• Real-time problem-solving and debuggingThis isn't your typical polished tutorial - it's messy, honest, and shows the actual process of building with AI tools. You'll see my thought process, mistakes, and how I iterate on ideas in real-time.🔔 Subscribe for more AI building sessions and hit the notification bell!💬 What AI project are you working on in 2026? Drop it in the comments!#AI #VibeCoding #NoCode #GoogleAI #Claude #Perplexity #Chatgpt #openAI #Coding #2026

In this episode of Approachable AI, Joshua brings you into a live build session of what we're calling "Veritas", a bias rating app for news, podcasts, and other media. Drawing on his journalism background, he breaks down why “unbiased” news doesn’t really exist, what different kinds of bias look like in language, and how AI can help surface those patterns in a more transparent way. Then he vibe codes Veritas in real time, tweaking prompts, wiring up UIs, and stress testing the tool on live articles from across the political spectrum.You’ll see him refine the scoring system for opinion density, emotional language, framing, and certainty, build a bias index of outlets like Reuters, NPR, Fox, and more, and experiment with a “Neutralize the narrative” rewrite button. In the second half, he expands Veritas with a multi-source Top Stories feed, adds color-coded bias labels, fixes broken links, and rebuilds the About page using inspiration from modern agency sites, all while talking through his process of vibe coding as a nontraditional developer.If you’re curious about media literacy, bias, and how far you can push AI tools without being a full-time engineer, this episode is for you.In this episode you will:• Hear a simple breakdown of why unbiased media is a myth and what “good” bias can look like when it serves a community• Learn how Joshua designed Veritas to rate bias in articles, podcasts, and videos with an interpretable score rather than a fact-check label• Watch a live build of features like text pasting, quote highlighting, a bias index, and a “Neutralize the narrative” rewrite mode• See Veritas tested on real coverage, including crime stories, Iraq war–era reporting, and public health news to expose how language shifts the story• Follow the expansion into a multi-outlet Top Stories dashboard that pulls headlines from several news sources and tags them with live bias scores• Get a peek at how Joshua uses vibe coding and voice prompts to iterate quickly on UX, color systems, bugs, and even the About page copyTimestamps:00:00 – Intro, camera woes, and why we’re not waiting for perfect gear02:05 – From Archive to Veritas, the new app Joshua is building05:10 – Journalism school lessons and why unbiased media doesn’t really exist09:20 – Defining bias types: opinion, emotional language, framing, and certainty13:45 – Walking through the first Veritas prototype and its scoring system18:30 – Testing Veritas on the Luigi Mangione case article and tuning highlights22:40 – Iraq war coverage, “Neutralize the narrative,” and where the tool is too harsh26:55 – Dreaming of a browser extension and building the first Top Stories tab31:45 – Part two: Joshua comes back to keep building Veritas33:10 – Upgrading Top Stories with CNN, Fox, Bloomberg, Reuters, Wall Street Journal, Perplexity, and Finviz36:30 – Using ChatGPT voice notes to push app updates and handle bugs40:05 – Adding color ranges for bias scores (red, yellow, green, blue) in the feed44:20 – Sorting and linking the Bias Index outlets and podcasts with live URLs48:15 – Redesigning the About page using agency site inspiration and making it feel more like Veritas52:40 – Reflecting on vibe coding, feeling “limitless,” and what to build in future episodes55:30 – Closing thoughts, invitation for ideas from viewers, and sign off

Welcome back to Approachable AI, your shortcut to understanding artificial intelligence in plain English. We explore what’s happening, why it matters, and how you can use AI so that it feels less like the future and more like a part of your everyday routine.In Episode #002, recorded on November 21, 2025, we react to a massive week in AI history: the game-changing launch of Google Gemini 3.0. In this episode, we dive deep into the concept of "vibe coding"—using advanced models to build applications without having any technical background or knowledge of computer science languages like Python. We explore the differences between the standard Google Gemini interface and the more robust builder-focused Google AI Studio, demonstrating how to take simple ideas and turn them into functional products.In this episode, we explore:- The Gemini 3.0 Drop: Discussing the immediate impact of Google's latest release, which feels like a significant step toward Artificial General Intelligence and blows recent releases like ChatGPT 5.0 out of the water.- The Rise of "Vibe Coding": How AI is opening doors for builders who don't code, allowing them to generate apps just by describing their ideas.- Vibe Coded App Showcases:- Binaural Beats: A look at iterating on a frequency generator app, adding features like visualizers and manual calibration through simple prompting.- AI Stylist App: We spend the majority of the episode live-testing a digital closet app. Watch as we build a personal AI stylist from scratch, showing you how to take a creative idea and refine it into a working tool that generates unique outfits on demand.Handling AI Hallucinations: Navigating the messy reality of image generation when the AI mysteriously adds unwanted accessories like ski masks or incorrect bags to the final designs.📍 Tools & Tech Mentioned: Google Gemini 3.0 Google AI Studio Nana Banana 3.1 (Image Generator), Google Video 3.1, Cursor (Mentioned)Timestamps: 0:00 Intro & The massive impact of Google Gemini 3.0 5:30 Defining "Vibe Coding" and Google AI Studio 8:15 Project Showcase: Iterating on a Binaural Frequency Generator 13:15 Building the Digital Closet AI Stylist 22:30 Live Test: Styling for a blizzard using icon inspirations (Bowie, Ford, Westwood) 32:00 Analyzing the AI's "Quiet Powerbroker" and "Cyberpunk" outfit generations 39:00 Refining outputs and removing unwanted AI additions 40:45 Final thoughts: Now is the time to be the "ideas person" #ApproachableAI #GoogleGemini3 #VibeCoding #AIStudio #NoCode #ArtificialIntelligence #TechPodcast #GenAI

Welcome to the premiere of Approachable AI, your shortcut to understanding artificial intelligence in plain English. We explore what’s happening, why it matters, and how you can use it so that AI feels less like the future and more like a part of your everyday workflow.In this episode, we dive straight into the messy, creative reality of using today’s most powerful tools. We aren't just talking about concepts; we are live-testing them. Watch as we navigate the highs and lows of rapid prototyping—bouncing between ChatGPT Atlas, Google Gemini, and Google AI Studio to see what happens when you push these models to their limits. It’s a look at the "trial and error" process required to turn raw ideas into functional code.In this episode, we explore:- The Reality of Rapid Prototyping: How to go from a spoken concept to a deployed web app in minutes, and how to handle it when the code stalls.- Multimodal Inputs: Testing the difference between text prompts versus using voice and images to guide AI development.- The Art of the Pivot: Knowing when to switch platforms (like moving from Atlas to Gemini) when a specific tool hits a wall.- Agent Limitations: A look at where current AI agents struggle, specifically when asked to organize local desktop files and manage complex workflows.📍 Tools & Tech Mentioned:ChatGPT Atlas (Browser/Agent)Google Gemini 2.5 ProGoogle AI Studio (App Builder)Google VeoPerplexity AIPerplexity CometTimestamps:0:00 Intro: Making AI part of your everyday journey2:15 Testing Google AI Studio & V3.110:30 When AI gets stuck: Troubleshooting code generation15:00 Using Gemini for creative content and video prompts25:00 iterating on UI/UX design with multimodal inputs35:00 Testing desktop agents on file organization45:00 Final thoughts and takeaways#ApproachableAI #GoogleGemini #ChatGPT #AIStudio #TechPodcast #ai #vibecoding #openai #google