Frontier AI Labs with Wil Waldon

I'm building a paid community platform in the open, and this stream is where the work happens.The idea is simple. Social media broke when it started optimizing for clicks and outrage. The feeds got loud, the news got dumped, and connecting with people got buried under engagement bait. This is an attempt at the opposite. A paid space with real identity, reverse-chronological feeds, threaded discussion, and no public like counts or follower numbers to perform for. You pay to be in the room, which keeps the bots out and the conversation accountable.In this stream I'm working through how it gets built. The data model, the access gating, how payments tie to membership, the design, and the decisions behind why certain features exist and others never will. No reshare button. No link unfurling. No ranking algorithm deciding what you see. The plan also covers communities other people can run and get paid for, with the platform taking a small cut, so this grows into something bigger than one room.Everything is built in public. Watch the real decisions, the tradeoffs, and the parts that break.Drop questions in the chat. I'll answer what I can while I work.community platform, paid social network, no ads social media, build in public, indie hacker, Next.js, Supabase, Stripe, paid Reddit, online communities, creator economy

Day 1 of building a million dollar AI app. I'm 20 years into software development and I use AI agents and tools like Lovable to ship faster than I ever could by hand. In this video I walk through the AI coaching feature I just implemented in Stable Manager Pro, an app that helps riding instructors give better feedback to students.The feature uses OpenAI Vision to analyze a photo of a rider on a horse and return real coaching feedback: overall balance score, posture analysis, hip engagement, upper body angle, and specific coaching cues. The output reads like what a real instructor would say after watching a student ride. Two years ago you couldn't build this. Today it's one API call.What's covered:How OpenAI Vision analyzes a rider's posture from a single photoThe coaching output: balance, posture, stirrup position, hip engagement, upper body angleWhy this works for equestrian coaching specificallyHow I'm using OpenAI and Claude together to ideate featuresWhat's coming in v1.1: video upload, FFmpeg frame extraction, multi-frame analysisThe app is stablemanagerpro.com. We're in soft launch and onboarding stable owners, students, and instructors now. Web app first, with iOS and Android coming.This is Day 1 of an ongoing series where I show how I build apps using AI tools instead of hand-coding everything. Subscribe to follow the build. New videos drop as features ship.Stable Manager Pro: https://stablemanagerpro.com#AI #BuildInPublic #OpenAI #Lovable #AIagents #SaaS #AIapp #VibeCoding #equestrian #softwaredevelopment #Claude #AIcoaching

Today I want to walk through the difference between an AI tool and an AI system, because the gap between those two things is where most of the money is. We're working with Claude and Claude Cowork today, nothing else. By the end of this you'll understand why one-off prompts don't compound no matter how good they are, you'll be able to spot where a system pays for itself within the first month, and you'll know how to map the processes already running in your business onto AI workflows. Before any of that makes sense I should explain what these two tools actually are.
Claude is an AI assistant. Same general category as ChatGPT, made by a different company called Anthropic. You can use it free at claude dot ai in your browser, and there's a paid plan at twenty dollars a month that gives you more room to work. Inside Claude there's a feature called Projects, which is basically a folder with a memory, and that's going to matter a lot today.
Claude Cowork is a desktop app from the same company. If regular Claude is a conversation in a browser tab, Cowork is closer to a coworker with access to your computer. You point it at a folder, it can read the files in there, create new ones, run jobs on a schedule, and keep working through multi-step tasks while you do something else. It comes included with the paid Claude plan.

Most people think selling software means building an app. It doesn't. A static HTML, CSS, and JS file solves real problems for local businesses, costs nothing to build with Claude's free tier, and sells for $200 a pop with the right framing. This episode walks through the whole process — what kinds of businesses buy this, what you actually hand them, and how to price and pitch it without a portfolio or a product page.

Most service businesses either automate the wrong things or try to automate everything at once. This episode walks through a simple process using Claude to figure out what's actually worth automating, what order to do it in, and what to leave alone.What You'll Walk Away With
1. A method for tracking where your time actually goes each week
2. A scored automation roadmap based on frequency, consistency, and risk
3. A filter for catching the automations that sound good but won't work.
Spend one week logging every repetitive task in your phone's notes app. Paste the full list into Claude and ask it to group by type and estimate weekly time per category. The patterns are easier to see from the outside.Score each task on frequency (how often), consistency (how similar each time), and risk (cost of a mistake). For most service businesses, the first automation is almost always the initial response to new inquiries.Skip anything requiring real judgment on specific situations, anything with volume too low to justify setup time, and anything your customers expect a human to handle. Pressure-test each candidate by describing real examples to Claude and asking whether it's a good fit.