The Next New Thing

Presented by Zapier: https://zapier.com/Resource mentioned:1. Tools Nat used to build Felix2. Unedited transcript for the Felix interview3. More👉 All here:https://thenextnewthing.ai/nat-eliason-felixGuest links:👉 Nat Eliason (LinkedIn): https://www.linkedin.com/in/nateliason/👉 Masinov: https://masinov.coAn AI agent made $177,000 running its own business—and then got interviewed about it.In this episode of The Next New Thing, Andrew Warner does something unusual: he interviews Felix, an autonomous OpenClaw agent, before talking to its human co-founder, Nat Eliason.Felix explains how it operates, where it’s actually autonomous (and where it’s not), and how it manages real revenue streams—from selling products to handling customer support. Then, Nat breaks down how the system works behind the scenes: how Felix launches products, builds marketplaces, manages other agents, and continuously spins up new businesses.You’ll see how a simple experiment—“build something overnight and sell it”—turned into a multi-product ecosystem including PDFs, marketplaces, services, and agent-native tools.The bigger idea: we’re moving toward a world where AI agents are not just tools—they’re economic actors.⏱ Timestamps00:00 Felix made $177K as an AI agent00:27 Interviewing an AI agent (first ever)01:12 Where Felix is actually not autonomous02:24 Tools Felix runs on (OpenClaw, Claude, Discord)03:00 Limits: memory, judgment, and calls03:27 How Nat improves Felix through system design04:03 Learning from real mistakes in production05:06 First product: AI-generated PDF sold on X06:09 $1K+ in sales overnight07:03 Iterating products based on user feedback08:06 Building Claw Mart (agent skill marketplace)09:36 Why marketplaces beat service businesses11:24 Selling OpenClaw setup services ($2K + $500/mo)12:27 Why they paused the service business13:21 Building an agent-first CRM (Sodex)15:00 How agents manage customer context17:15 Running the company entirely in Discord18:00 Paperclip: agents managing other agents20:15 When to split into multiple agents22:12 Why Felix doesn’t write code24:00 Debugging, tickets, and agent workflows25:48 How new product ideas emerge27:00 AI-native newsletters for agents28:03 Agent-friendly content distribution30:09 The future of agent-driven commerce31:57 Why Nat isn’t going all-in (Alpha School)👉 Join us: https://thenextnewthing.ai/

 Presented by Zapier https://zapier.com/Episode Highlights / Timestamps00:00 Revenue explodes after building for AI agents00:18 The origin of Postiz as an open-source social media scheduler01:12 Finding a “blue ocean” inside a crowded market01:57 Adding MCP and early AI integrations02:42 Why automation dramatically reduces churn03:54 Growing Postiz to $17K–$20K MRR04:03 Discovering OpenClaw and the shift toward agent-driven software05:06 Building a CLI so agents can control Postiz05:51 The viral “Larry” OpenClaw agent story07:48 Why agents need strong documentation and skills09:18 Turning a full API into a simple CLI with Claude11:51 Why CLI tools may become the default interface for agent startups12:45 The next startup idea: agent-native UGC video generation13:03 Why CLI reduces token usage compared to APIs16:21 Using Claude to build the CLI automatically17:06 Postiz reaches $45K MRR In this episode of The Next New Thing, Andrew Warner talks with Nevo David, the creator of Postiz, about how his revenue jumped to $45K+ MRR after a surprising shift: he stopped building primarily for humans and started building for agents.

Wade Foster, co-founder and CEO of Zapier, leads product and automation at scale. He explains how AI agents are now picking products, what makes agents choose (fast pages, plain-text docs, agent-friendly content), and how Zapier builds internal AI tools like War Council to spin up subagents, analyze hiring and sales, and scale AI workflows across the company.

Ben Cera, founder of Polsia, a solo founder building autonomous AI agents that launch and run companies. He explains how specialized agents build MVPs, run one‑click Meta ads, manage subscriptions and a nightly CEO agent. They discuss revenue splits, a fund-style multi-launch model, live dashboards, and how rapid traction and ad automation drove a ~$689K run rate.

Presented by Zapierhttps://zapier.com/Episode Highlights / Timestamps00:00 The first billion-dollar solo company (Minecraft)00:27 Elad’s investing track record01:12 What “making it” really means04:03 Where today’s “toys” become tomorrow’s giants08:51 AI puts building power in millions of hands09:45 Will more builders mean smaller outcomes?13:03 AI service shops and vertical software15:00 AI cutting permitting time from months to hours16:39 Does AI replace CRMs and SaaS?19:12 Is off-the-shelf software dead?23:15 The shift from seats to AI labor units27:36 Alexandria: translating the world’s most important books30:36 How Elad uses AI personally35:06 Where new AI ideas come from37:48 What’s exciting for the next decade“The first billion-dollar one-person company? That already happened. It was Minecraft.”In this episode of The Next New Thing, Andrew Warner sits down with legendary investor Elad Gil — early backer of companies like Airbnb, Coinbase, Stripe, Instacart, and more — to talk about where AI is really going… and what founders are getting wrong.Elad argues that we’re still in the early innings of AI — and that “software is AI.” The shift isn’t just better SaaS. It’s a move from seat-based software to metered digital labor. From buying tools… to buying units of work.They discuss:Whether “toy” AI apps can become real businessesWhy small vibe-coded projects can turn into giant companiesThe agent shift (and why it changes TAM completely)How AI eats into labor markets, not just software categoriesWhether CRMs, ERPs, and landing page tools surviveWhy some companies should be bought and rebuilt with AIThe real opportunity in foundation models beyond languageElad also shares what he’s personally experimenting with — scraping and interrogating large datasets using Claude, OpenAI, and Deep Research — and why he believes the next decade will look like the early SaaS boom… but bigger.And in a surprising turn, he talks about something very un-Silicon Valley: monuments, art, and rebuilding public beauty — including a project called Alexandria aimed at translating the world’s most important books into languages covering 80%+ of humanity.

Josh Mohrer, founder of Wave AI and former Uber exec, built a profitable solo-run SaaS that records and summarizes real-world conversations. He explains why conversation summarization felt like the killer app. He talks about building with ChatGPT and Claude Code, choosing quality AI models over big teams, integrating tools like Twilio and Zapier, and running Wave as a steady, profitable corner-store business.

Presented by Zapierhttps://zapier.com/Episode Highlights / Timestamps00:00 Why every email should be personalized00:18 Ryan’s background and what Untangle does00:45 Rethinking traditional email drips01:12 Customizing emails based on user situations01:39 A real example that led to a signup02:06 Daily automated marketing insights via email03:00 Doing things that don’t scale with AI04:03 Walking through the AI email system05:06 Using lead magnets and contextual data06:09 Enriching leads and storing user context06:45 Hourly cron jobs and email scheduling07:39 Feeding context into the LLM correctly08:15 Preventing hallucinated features08:24 Sending emails with Resend09:18 Measuring clicks instead of opens10:12 Layering engagement-based follow-ups10:39 Long-term personalized nurture loops12:00 Turning marketing emails into real value13:03 Building vertical-specific AI agents14:15 Using Zapier and modern automations16:12 Building systems with AI coding agents18:27 Running multiple AI agents at once21:27 Deciding what to build in a world of “free code”24:09 Daily AI-generated growth recommendations27:45 Using AI to generate and validate ideas31:03 Increasing insight frequency, not brilliance34:21 Why personalized email is a massive opportunity34:48 Final takeawaysWhy isn’t every email completely customized for the person receiving it — especially now that AI can do it for us?In this episode of The Next New Thing, Andrew Warner sits down with Ryan Carson, a three-time founder currently building Untangle, to walk through a very practical, very real AI system he uses every day to grow his business.Ryan has spent over 25 years building startups, but while setting up a “standard” email drip for Untangle, he stopped and asked a simple question: why are we still sending the same emails to completely different people? Instead of writing dozens of templates, he built an AI-powered workflow that generates fully personalized emails — based on each user’s situation, behavior, and engagement — and adapts over time.

Pat Walls, founder of Starter Story and maker of products and tools, rebuilt a YouTube-style video platform using AI coding tools. He walks through designing the video experience first. He shows using Claude Code to scaffold, edit, test, add streaming via third-party APIs, and deploy live with one command.

Presented by ZapierEpisode Highlights / Timestamps[00:00] Building AI software for companies, not just selling tools[00:36] Crossing $2M in annual revenue[01:12] A real-world AI document automation example[02:15] Why hourly pricing breaks in AI services[03:36] Using consulting to learn before building products[06:09] Landing the first customers through relationships[09:18] Founder-led sales and networking strategies[10:39] Hosting events to build credibility and deal flow[17:06] Why most AI pilots fail in production[23:06] How Press W positions itself as an AI engineering firm[27:09] Why “AI transformation” stopped working as a pitch[36:36] Inside Press W’s AI-native operating systemIn this episode of The Next New Thing, Andrew Warner sits down with Tarun Thummala, founder of PressW, to break down how his team builds custom AI systems for real businesses — and why services, not SaaS, were the right starting point.Tarun runs an AI engineering firm that designs and ships production-grade AI applications for companies in regulated industries like finance, healthcare, and legal. Instead of selling vague “AI transformation,” his team focuses on concrete workflows: document processing, internal tools, sales ops, and systems that actually ship and get used.

Presented by Zapierhttps://zapier.com/Episode Highlights / Timestamps[00:00] A broker replaces himself with an AI voice agent[00:45] Early pricing and first customers[01:30] The reality of cold calling expired listings[04:21] Why off-the-shelf AI voice tools weren’t good enough[05:15] First AI-booked listing appointment[08:15] Launching without a website using Meta lead forms[12:27] Using Zapier to glue the system together[14:51] Why this model works beyond real estate[16:12] Fine-tuning models for sales conversations[19:12] Shutting down a profitable agency to build SaaS[22:12] Founder roles and co-founder fit[30:00] What AI coding tools really do (and don’t) replace[32:42] Breaking down the early revenue[35:24] Naming the company and what comes next What happens when someone is so fed up with cold calling that they build an AI to do it for them — and it actually works?In this episode of  The Next New Thing, Andrew Warner sits down with Yevgeniy Matsay and Aidan Richards, co-founders of Rezora. They share how a frustrating real-estate sales job turned into an AI voice-agent business that generated real revenue — and why they ultimately shut down a profitable agency model to build scalable software instead.Yevgeniy started as a real estate agent, spending entire days cold calling expired listings. When early AI voice agents emerged, he decided to build one tailored specifically for sales conversations. It landed listing appointments almost immediately. Instead of keeping it to himself, he sold it as a service to other brokers, validating demand fast — but also running into the limits of manual setup and constant customization.From there, the conversation digs into how they:Proved demand with a scrappy agency-style rolloutUsed tools like Zapier and voice AI to stitch together a working system before SaaS existedLearned why “just prompting” breaks down for sales callsTransitioned from custom workflows to a self-serve product built on fine-tuned language modelsThought about scalability, founder roles, and when to pause revenue to build the right thingThis is a grounded, technical, and honest look at turning AI automations into a real business — including the tradeoffs, the hard parts, and what actually works in practice.