We Tested OpenAI's GPT 5.6 for a Month

I sit down with Elie Steinbock to unpack loop engineering and how to run a business on loops. We start with the roots of the idea in the lean startup and Toyota's manufacturing, then move into practical, copy-ready workflows for SEO, Facebook ads, and product feedback. Elie walks through a live Google Search Console example on Draft Fantasy and shows how to set up an SEO loop that runs once a month for years. The core promise for listeners: hand repeatable business work to an AI agent that measures an objective metric and improves over time. By the end, you know how loops work and how to launch your first one today. Timestamps 00:00 – Intro and episode promise 02:54 – What is Loop Engineering 06:51 – Loops with AI agents: build and verify 11:17 – Example of Loop: SEO as an objective-metric loop 15:29 – Setting up the SEO loop and tools 25:27 – Cost and token economics 29:05 – The Paid ads loop 33:10 – The product feedback loop 36:25 – A minimal viable loop for every channel 39:21 – Closing Thoughts Key Points Loop engineering means giving an agent a task, an objective metric, and a stop condition so it improves on a schedule. The lean startup and Toyota's build-measure-learn cycle map directly onto AI agents. An SEO loop connects to Google Search Console and Data for SEO, then pushes rankings up month over month. These loops run cheaply — often a few dollars per monthly run — which beats the cost of an agency. The same pattern extends to Facebook ads, and a product feedback loop stands as the ultimate version. Start small with a minimal viable loop tied to a clear metric like impressions or ten likes. Numbered Section Summaries The Promise of Running a Business on Loops I open by asking Elie what listeners will walk away with, and he frames the whole episode: use loops to automate SEO, ads, and more. We agree the aim is clear, copyable workflows people can launch today. Where Loop Engineering Comes From Elie traces the recent buzz to Boris from Claude Code and Peter Steinberger, plus a joking tweet from his friend Dimitro about software that builds itself. He grounds it in the lean startup's build-measure-learn cycle, which itself grew from Toyota's lean manufacturing. Loops With AI Agents: Build and Verify Elie explains the agent version: a build step paired with a verify step and a clear stop condition. He uses Inbox Zero's evals as an example, where the agent keeps adjusting the prompt or model until accuracy passes 90%. The SEO Loop We dig into SEO as the flagship example, where Google ranking serves as a clean, objective metric. Elie describes a loop that runs once a month, learns from the last run via a markdown memory file, and steadily climbs the rankings. Setting It Up on Real Data Elie shows his Draft Fantasy Search Console, connects the agent to Google Search Console and Data for SEO, and runs the loop live in Codex. He shares the Atom Eve prompt as a deeper template people can copy. Cost and Token Economics I raise Ross Mike's skepticism about loop buzz and token spend, and Elie makes the case that an SEO loop stays cheap — often under five dollars per monthly run. He adds that Max-plan users have plenty of headroom, while tight budgets suit cheaper open models like GLM 5.2. Ads, Product Feedback, and the Ultimate Loop We move to a Facebook ads loop that tests copy and creative variants, favoring a mix of human hooks and AI optimization. Then Elie describes the product feedback loop — reading customer feedback, analytics, and logs to prioritize and ship — as the closest thing to a business that builds itself. Starting Small We close on the minimal viable loop: begin with one channel and a modest, verifiable metric like impressions or ten likes, then let it compound. Elie and I agree that every part of a business could sit on a loop, and starting one today makes for a low-risk experiment. The #1 tool to find startup ideas/trends - https://www.ideabrowser.com LCA helps Fortune 500s and fast-growing startups build their future - from Warner Music to Fortnite to Dropbox. We turn 'what if' into reality with AI, apps, and next-gen products https://latecheckout.agency/ The Vibe Marketer - Resources for people into vibe marketing/marketing with AI: https://www.thevibemarketer.com/ FIND ME ON SOCIAL X/Twitter: https://twitter.com/gregisenberg Instagram: https://instagram.com/gregisenberg/ LinkedIn: https://www.linkedin.com/in/gisenberg/ FIND ELIE ON SOCIAL Youtube: https://www.youtube.com/elie2222 X/Twitter: https://x.com/elie2222

In this episode I bring Nick Vasilescu, co-founder of Orgo, back on the show to unpack the buzz around Grok 4.5. Nick makes the case for treating Grok 4.5 as a genuine AI co-founder inside harnesses like Hermes and OpenClaw, and he proves it live: spinning up cloud computers, wiring in tools, and building a full startup from idea to landing page to outreach. We race Grok 4.5 against GPT 5.6 Sol, tour Nick's agent stack, and talk through the cost paradox of a model this fast and cheap. Listeners walk away with a concrete playbook for standing up their own always-on agent today. Get Nick’s Agent Template Stack: https://startup-ideas-pod.link/nicks-stack Timestamps 00:00 – Intro 01:40 – Why Grok 4.5 release matters 03:39 – Automation versus a co-founder 05:16 – Setting up Hermes and Grok 4.5 on Orgo 09:01 – Why Orgo to manage Agents 11:23 – Grok 4.5 Cost discussion 14:02 – Grok 4.5 Fast Execution and Unlock 16:20 – The Agent tool belt 19:13 – X MCP for trends 20:37 – vidIQ for outliers and thumbnails 22:11 – Finding new startup ideas 26:15 – Grok 4.5 versus GPT 5.6 Sol 30:56 – Ranking and Reviewing the startup ideas 34:06 – The AI agency opportunity 38:46 – Thumbnails over Telegram 40:10 – Reviewing AI Agency Landing Page 41:58 – Vertical MCPs and agent startups 43:36 – Skill graph and the offer 45:25 – Reviewing the Thumbnail Generated 46:42 – Email Outreach Campaign 48:07 – Reviewing Market Insight 1-Pager 50:45 – From a Camry to a Ferrari 52:12 – Reviewing Cold Email Outreach Sequence 53:22 – Closing thoughts Key Points Grok 4.5 delivers Opus 4.8-level intelligence at a fraction of the cost and roughly 10-15x the speed of Fable. I learn to treat the model as a co-founder by handing it email, a phone number, a debit card, memory, and every connector that matters. Nick runs agents on Orgo cloud computers so they stay online, textable, and ready around the clock. Live, Grok 4.5 builds a landing page in about 40 seconds and wins on design and copy for me over GPT 5.6 Sol. The stack ships from idea to website, offer, thumbnail, and cold-email sequence in a single session. Nick's take: costs keep dropping while speed and intelligence keep climbing, so building your agent now compounds overnight. The #1 tool to find startup ideas/trends - https://www.ideabrowser.com LCA helps Fortune 500s and fast-growing startups build their future - from Warner Music to Fortnite to Dropbox. We turn 'what if' into reality with AI, apps, and next-gen products https://latecheckout.agency/ The Vibe Marketer - Resources for people into vibe marketing/marketing with AI: https://www.thevibemarketer.com/ FIND ME ON SOCIAL X/Twitter: https://twitter.com/gregisenberg Instagram: https://instagram.com/gregisenberg/ LinkedIn: https://www.linkedin.com/in/gisenberg/ FIND NICK ON SOCIAL Youtube: https://www.youtube.com/@nickvasiles Instagram: https://www.instagram.com/nickvasilescu/ Personal Website: https://www.nickvasilescu.com/

In this episode I sit down with Dan Shipper to see how he runs his work and personal life on OpenAI's Codex Desktop with the 5.6 model. He walks through his card-based email setup, daily feeds for his company and Slack, and the in-app browser that lets his agent collaborate with him inside tools like Proof. We build a small SaaS app live, called Turnaround, and use it to explore why maintenance is the real product in the AI era and where Codex-native software heads next. Along the way Dan shares his pirates-versus-architects framing, his approach to fine-tuning a copy-editing model, and the patterns — pulses, Mailroom, and router threads — that hold his system together. The throughline: pick one simple win, let context do the heavy lifting, and manage the system instead of running every task by hand. Learn how to get customers with AI Agents: https://startup-ideas-pod.link/GTM-agents-IB Timestamps 00:00 – Intro 01:16 – Codex and GPT-5.6 Overview 03:40 – Training your own model: the step after skills 04:49 – Automating Email, Slack, Meeting Notes with GPT-5.6 08:53 – Why GPT-5.6 sharpens the results 10:26 – The light bulb moment with Codex 15:05 – Building Turnaround live: a maintenance badge 18:00 – GPT-5.6 vs. Fable: A tier and S-plus tier 19:34 – LFG and goal: looping toward a finished build 24:28 – Huge Opportunity: Codex-native apps 29:33 – The design checkpoint and the "warm paper" quirk 31:32 – Local models 34:04 – From 70% to 100%: pirates and architects 37:22 – Mailroom: giving Codex its own email address 40:58 – Getting started: download, grant access, explore 43:07 – Record and Replay: turning tasks into skills 44:37 – Closing Thoughts: Start small and build over time Key Points Codex Desktop plus the 5.6 model runs as a full operating system for knowledge work — email, research, and building software from one surface. Context is the multiplier: an agent wired into your computer and the web turns every inbox and feed into cards with a clear next action. Maintenance is the real product in the AI era, now that anyone can one-shot a first version. Codex-native SaaS — software you and your agent share inside the in-app browser — opens a fresh category with healthier margins. A live build of Turnaround, a maintenance-status badge, reaches about 70% in one pass; an architect carries it the rest of the way. Start with one simple win, grow the system over time, and let curiosity lead the way in. The #1 tool to find startup ideas/trends - https://www.ideabrowser.com LCA helps Fortune 500s and fast-growing startups build their future - from Warner Music to Fortnite to Dropbox. We turn 'what if' into reality with AI, apps, and next-gen products https://latecheckout.agency/ The Vibe Marketer - Resources for people into vibe marketing/marketing with AI: https://www.thevibemarketer.com/ FIND ME ON SOCIAL X/Twitter: https://twitter.com/gregisenberg Instagram: https://instagram.com/gregisenberg/ LinkedIn: https://www.linkedin.com/in/gisenberg/ FIND DAN ON SOCIAL X/Twitter: https://x.com/danshipper Youtube: https://www.youtube.com/@EveryInc/videos Every: https://every.to/

In this solo episode I lay out why I believe building agents is the new SaaS: software is shifting from helping you do the work to doing the work with you. I walk through a full playbook — find a niche, pick a workflow with a paycheck attached, shadow the human, spec the agent, build the minimum useful version, sell a pilot like labor, then productize the repeatable parts. I share live market examples like Slang AI for restaurants and Same Day for home services, plus pricing models and a distribution strategy built on workflow teardowns. I close with a 30-day, zero-to-100 plan for launching an agent-first business. This one is for anyone eager to build with AI or simply become more productive. Timestamps 00:00 – Intro 01:38 – Building Agents is the new SaaS 04:11 – Pick a valuable workflow 06:12 – Shadow the Human First 09:34 – Build the Minimum Useful Agent 12:50 – The wrapper makes it SaaS 15:50 – Sell the Pilot Like Labor (and Pricing) 18:37 – Own the workflow 21:45 – The Zero-to-100 Plan in 30 Days 24:14 – Closing Thoughts Key Points Agent SaaS sells work as a service; the product is the job itself, priced like labor. Start with a workflow that already carries a paycheck: high frequency, clear finish line, existing software, learnable edge cases, and felt pain. Shadow a human across 10–20 real jobs before you write a single prompt — the detail is the product. Ship the minimum useful agent — draft-and-approve, triage, coordinator, or bounded action — and earn autonomy over time. The wrapper (logs, approvals, evals, analytics) creates trust and turns automation into real SaaS. Win distribution with workflow teardowns: show the old way, show the agent way, sell the painkiller. The #1 tool to find startup ideas/trends - https://www.ideabrowser.com LCA helps Fortune 500s and fast-growing startups build their future - from Warner Music to Fortnite to Dropbox. We turn 'what if' into reality with AI, apps, and next-gen products https://latecheckout.agency/ The Vibe Marketer - Resources for people into vibe marketing/marketing with AI: https://www.thevibemarketer.com/ FIND ME ON SOCIAL X/Twitter: https://twitter.com/gregisenberg Instagram: https://instagram.com/gregisenberg/ LinkedIn: https://www.linkedin.com/in/gisenberg/

In this solo episode, I lay out the six skills I believe stay valuable as AI grows more capable. I chose these six because each one is open to anyone, each one starts this weekend, and each one rises in value as AI improves. I walk through agents and local models, distribution, robotics, curation, the builder distributor, and IRL community building, with one concrete first rep for every skill. My goal is to hand you one simple, clear map of where the world is heading and exactly how to begin. Timestamps 00:00 – Intro 00:57 – Skill 1: Running AI Agents and Local Models 04:51 – Skill 2: Marketers Who Build Distribution 09:03 – Skill 3: Robotics Engineers Who Build and Source Hardware 14:29 – Skill 4: Curators Who Yap and Make Short-Form Video 19:05 – Skill 5: The Builder Distributor 23:11 – Skill 6: IRL Community Builders 27:34 – Build Your Skill Stack Key Points I chose these six skills because each one rises in value as AI improves. Skill 1 is the grown-up version of prompt engineering: I design an AI worker with context, tools, memory, permissions, and a goal. Distribution beats posting, so I learn where attention already lives and turn it into trust before I sell. Hardware is the new frontier: cheap arms, open-source robot learning, and supplier sourcing put robotics within my reach. As the builder distributor, I ship the product and win the attention in one loop, which makes the one-person company real. Real rooms grow scarce and valuable, so I build belonging, trust, and context as my edge. Numbered Section Summaries The Premise: What Stays Valuable as AI Improves I open by picturing a near future where AI builds and writes almost anything, then ask which skills hold their value. I narrow it to six skills that anyone can start this weekend, each one climbing in value as AI gets better. Skill 1 — Agents and Local Models I describe the move from typing prompts to designing a small AI employee with context, tools, permissions, memory, a goal, and a way to check its own work. I add local models with tools like Ollama and LM Studio so you learn which jobs want a giant brain and which jobs want a reliable worker, and I suggest building a daily briefing agent with three sources as your first rep. Skill 2 — Marketers Who Build Distribution I explain that distribution runs far deeper than posting: it means knowing where attention already lives and the exact words people use to describe their problem. The winning marketer becomes part researcher, storyteller, media operator, and community builder, and the first rep is a distribution map plus 20 hooks for a single idea. Skill 3 — Robotics Engineers Who Build and Source Hardware I share my big insight: the last decade rewarded moving pixels, and the next decade rewards moving atoms too. With cheap cameras, low-cost arms like the SO-100 / SO-101, open-source work like Hugging Face LeRobot, and small VLA models, I suggest assembling a low-cost arm, teaching it one boring task, documenting every failure, and learning supplier sourcing on Alibaba. Skill 4 — Curators Who Yap and Make Short-Form Video I cover the curator who watches the timeline and says "this matters because…," translating new models, launches, and news for a specific niche. The algorithms reward raw, authentic yapping that carries a real take, and my rep is a seven-day curation sprint paired with a taste file of hooks, analogies, and titles you love. Skill 5 — The Builder Distributor I make the case that AI compresses the old build-versus-sell split into one person who prototypes the product, writes the launch thread, records the demo, DMs the first users, and iterates. The loop is the whole game, and my rep is a 48-hour loop: build the smallest version of one problem, then create 10 pieces of distribution before you feel ready. Skill 6 — IRL Community Builders I close with the old-school skill that grows more valuable as work moves to agents and feeds: real rooms full of ambitious people. Scarcity moves toward belonging, trust, and context, so I suggest hosting six to eight people around one sharp question and sending a recap that turns the room into a network The #1 tool to find startup ideas/trends - https://www.ideabrowser.com LCA helps Fortune 500s and fast-growing startups build their future - from Warner Music to Fortnite to Dropbox. We turn 'what if' into reality with AI, apps, and next-gen products https://latecheckout.agency/ The Vibe Marketer - Resources for people into vibe marketing/marketing with AI: https://www.thevibemarketer.com/ FIND ME ON SOCIAL X/Twitter: https://twitter.com/gregisenberg Instagram: https://instagram.com/gregisenberg/ LinkedIn: https://www.linkedin.com/in/gisenberg/

In this episode I sit down with Amir to get tactical about running local AI models as part of a daily workflow. We center on GLM 5.2 from ZAI, how it stacks up against frontier models like Opus 4.8, and how a fusion approach lets you sequence a heavy thinking model with a lighter execution model for the best output at the lowest cost. Amir walks through setup in Cursor and Codex via OpenRouter, shares real token-cost math, and demos GLM 5.2 refining a live app. By the end you will know how to start today, where local models shine, and how model chaining keeps spend in check. Timestamps 00:00 – Intro 02:09 – GLM 5.2 and Z AI 04:01 – Specs: 1M context and Terminal Bench 2.1 05:22 – Making sense of benchmark scores 06:42 – Setup in Cursor or Codex with OpenRouter 10:18 – Local model upside: buy a machine, run tasks 11:42 – Token cost: 44 cents versus $2.38 13:36 – Future-proofing with an upfront hardware bet & The Uber subsidy analogy 16:49 – Model chaining and the vision workaround 19:23 – Token maxing vs routing tasks to the right model 20:54 – Answering the "cost is irrelevant" crowd 21:59 – Closing thoughts Key Points GLM 5.2 ships with a 1M-token context window and scores 81 on Terminal Bench 2.1, landing about four points behind Opus 4.8. A fusion approach (a term OpenRouter coined) sequences models: plan with Opus, execute with GLM 5.2, review with Composer 2.5 or Codex 5.5. Running GLM 5.2 in the cloud through OpenRouter costs roughly 44 cents for a task that runs about $2.38 on Opus 4.8 — close to a 5X saving. You can start today with credit-based access: load $20 in OpenRouter and route tasks to the right model. For images, Amir uses Opus 4.8 to read screenshots and describe them, then hands the layout to GLM 5.2 to act on. Teams are shifting from token-maxing to output-maxing, making model governance and chaining the smart play The #1 tool to find startup ideas/trends - https://www.ideabrowser.com LCA helps Fortune 500s and fast-growing startups build their future - from Warner Music to Fortnite to Dropbox. We turn 'what if' into reality with AI, apps, and next-gen products https://latecheckout.agency/ The Vibe Marketer - Resources for people into vibe marketing/marketing with AI: https://www.thevibemarketer.com/ FIND ME ON SOCIAL X/Twitter: https://twitter.com/gregisenberg Instagram: https://instagram.com/gregisenberg/ LinkedIn: https://www.linkedin.com/in/gisenberg/ FIND AMIR ON SOCIAL Humblytics: https://humblytics.com/?via=community X/Twitter: https://x.com/amirmxt Youtube: https://www.youtube.com/@amirmxt

In this episode I sit down with George Lampropoulos, a 19-year-old founder who turns AI-built mobile apps into real revenue. George walks through his framework for reaching $10K a month—roughly $333 a day—starting with a simple, sellable idea you actually care about and ending with a distribution plan anyone can run. He shares the numbers behind WrestleAI (100K-plus downloads and close to $200K in revenue) and explains why a sharp "gotcha feature" and a clean Instagram funnel do most of the heavy lifting. We also dig into closing influencers, hiring a VA, running paid ads, and reading the metrics that decide whether you grow. If you want a practical, founder-tested playbook for building apps with AI, this one delivers. George’s $10K/mo app playbook: https://startup-ideas-pod.link/George-app-playbook Timestamps 00:00 – Intro 01:54 – George's track record with WrestleAI 02:36 – How AI unlocks fresh app ideas 06:29 – Reverse-engineering a viral idea from your feed 16:16 – Designing the UI/UX of the app 17:49 – The gotcha feature that sells the app 21:25 – Onboarding that converts 23:04 – Actionable Plan to $10k/mo 28:55 – Outreach as a numbers game 33:35 – Paid ads clearly explained 36:20 – Reading metrics: conversion, ARPU, retention 38:30 – TLDR: a great product earns inbound creators 39:51 – Answering the vibe-coding skeptics 39:51 – Scaling with vibe-coded app 43:35 – Why now is the app-building boom 46:05 – Closing Thoughts Key Points I learn why a simple, sellable idea you're passionate about beats pure distribution every time George breaks down the "gotcha feature"—one feature so clear that five seconds explains the whole app We cover a clean Instagram page that doubles as a sales funnel and as social proof for recruiting creators George shares his influencer playbook: lead with relationships, close on a call, and aim for a $2 CPM I get his paid-ads starter method—5 to 15 creatives, $100 a day, then keep the winners. George explains the metrics that matter early: conversion rate, a $2 ARPU target, and retention The #1 tool to find startup ideas/trends - https://www.ideabrowser.com LCA helps Fortune 500s and fast-growing startups build their future - from Warner Music to Fortnite to Dropbox. We turn 'what if' into reality with AI, apps, and next-gen products https://latecheckout.agency/ The Vibe Marketer - Resources for people into vibe marketing/marketing with AI: https://www.thevibemarketer.com/ FIND ME ON SOCIAL X/Twitter: https://twitter.com/gregisenberg Instagram: https://instagram.com/gregisenberg/ LinkedIn: https://www.linkedin.com/in/gisenberg/ FIND GEORGE ON SOCIAL X/Twitter: https://x.com/GeorgeLampro20

In this solo episode, I walk through the implications of the ban of Claude Fable 5 — the most powerful model on the planet and the one I planned to build with — after the US government sent Anthropic a letter. I make the case for local AI by walking through the benefits: intelligence that lives on your own hardware, stays private, runs free after the hardware cost, and keeps working through bans, outages, and price hikes. I lay out the exact order I'd learn it in — runtimes, model-to-hardware matching, quantization, and agents — and I name the specific tools and models I reach for. Then I hand you five startup ideas that exist precisely because intelligence now sits on your desk. The payoff for you is a clear plan to own a resilient layer of your stack starting this week. Timestamps 00:00 – Intro 01:20 – The Fable 5 Ban 02:31 – Renting Access vs. Owning Intelligence 03:41 – How a Local Model Works 07:19 – The Local Model Stack 08:45 – Match Model to Machine 10:45 – Pick Your Model (Qwen 3, DeepSeek, Gemma, Llama) 13:09 – Quantization Explained 14:36 –The Local Agent Loop 17:45 – Model Routing (The Real Skill) 18:44 – Five Startup Ideas for the Local-AI Era 22:17 – Closing Thoughts Key Points One government letter took Fable 5 offline overnight, which is why I now own a private layer of my stack. Local models already handle roughly 80% of everyday ChatGPT or Claude tasks, fully offline and free after hardware. I'd learn it in order: runtime first (LM Studio or Ollama), then match model size to your RAM. A 12-billion-parameter model on 16 GB of RAM is the sweet spot where most people should live. Quantization (look for Q4) roughly halves the memory a model needs while keeping quality high. Pointing an agent like Hermes at a local model turns your desk into a private, always-on mini data center. The #1 tool to find startup ideas/trends - https://www.ideabrowser.com LCA helps Fortune 500s and fast-growing startups build their future - from Warner Music to Fortnite to Dropbox. We turn 'what if' into reality with AI, apps, and next-gen products https://latecheckout.agency/ The Vibe Marketer - Resources for people into vibe marketing/marketing with AI: https://www.thevibemarketer.com/ FIND ME ON SOCIAL X/Twitter: https://twitter.com/gregisenberg Instagram: https://instagram.com/gregisenberg/ LinkedIn: https://www.linkedin.com/in/gisenberg/

Get my Fable 5 prompt pack: https://startup-ideas-pod.link/fable5-prompt-pack In this episode I break down how to get the most out of Fable 5, the most powerful model I've ever used. I move past the benchmarks and go straight into tactical use cases, copy-and-paste prompts, and startup ideas you can build today. I walk through tournaments for copy and landing pages, an interview-before-build workflow that hunts for product-market fit, and ways to point Fable at contracts, churn data, and years of your own notes. I close with three of my favorite startup ideas — a synthetic focus group firm, 48-hour custom software, and a contract refund firm — plus the exact prompts behind each. My goal here stays simple: leave you ready to build and earn with Fable 5 while it remains included in your plan. Timestamps 00:00 – Intro 02:22 – Anthropic Employee Edits a Launch Video With Fable 05:50 – Building an AI Content Engine 07:30 – Best way to configure Fable 5 08:42 – Prompt 1: Copywriting Tournament for Landing Pages 13:18 – Prompt 2: The Interview-Before-Build Prompt 18:34 – Prompt 3: Hire Fable to Kill Your Company 20:18 – Prompt 4: Your One-Page Operating Manual 21:20 – Prompt 5: Find the Gaps Worth Filling 22:06 – Prompt 6: Negotiation Simulator 23:11 – Prompt 7: The 80-Page Second Opinion on Contracts 24:56 – Prompt 8: Make Fable Build Its Own Tools 25:47 – Startup Ideas 31:23 – Closing Thoughts Key Points I show why low effort is the alpha, since Fable Low beats Opus High on routine work. I run tournaments — landing pages and ad copy scored by AI judge panels — to ship far stronger output. I use an interview-before-build prompt so Fable pushes back and writes specs with real product-market-fit odds. I point Fable at big datasets — contracts, churn data, support tickets, years of notes — to surface money and patterns. I share startup ideas Fable 5 makes viable today, including a synthetic focus group firm and a contract refund firm. The #1 tool to find startup ideas/trends - https://www.ideabrowser.com LCA helps Fortune 500s and fast-growing startups build their future - from Warner Music to Fortnite to Dropbox. We turn 'what if' into reality with AI, apps, and next-gen products https://latecheckout.agency/ The Vibe Marketer - Resources for people into vibe marketing/marketing with AI: https://www.thevibemarketer.com/ FIND ME ON SOCIAL X/Twitter: https://twitter.com/gregisenberg Instagram: https://instagram.com/gregisenberg/ LinkedIn: https://www.linkedin.com/in/gisenberg/

S/o Coderabbit for sponsoring today’s vid: https://startup-ideas-pod.link/code-rabbit On this episode I sit down with Professor Ras Mic to break down agentic loops. We define what a loop is, explain why well-known builders like Boris and Peter swear by them, and stay honest about who they truly serve. Mic argues that human-in-the-loop remains the strongest setup today, and he walks through the one loop he runs every day for code review using Cursor, GitHub, and Greptile. By the end you will know when a loop earns its place and when your own hand belongs on the wheel. Timestamps 00:00 – Intro 01:23 – What is a Loop 07:59 – /goal Explained 11:32 – The Slop Machine 12:42 – Code Review as a use case for Agentic Loop 18:19 – Honest Take for Builders 20:42 – The Future of Loops 21:50 – Closing Thoughts Key Points A loop fires once from a human, then the agent generates, reviews its own result, and feeds it back to keep building. Human-in-the-loop keeps you directing, governing, and approving each step while the agent builds. Wide-open loops make heavy assumptions and burn serious tokens; Michael cites Peter's tweet about $1.3 million worth of tokens in one month. Reserve slash goal and similar loops for the $200/month plan, since the $20 and $100 tiers burn through fast. Loops shine in confined, fixed-feedback work: code review, SEO pages, and other binary tasks. Mic’s daily win is a closed code-review loop with Cursor, GitHub, and Greptile that chases a 5/5 score. The #1 tool to find startup ideas/trends - https://www.ideabrowser.com LCA helps Fortune 500s and fast-growing startups build their future - from Warner Music to Fortnite to Dropbox. We turn 'what if' into reality with AI, apps, and next-gen products https://latecheckout.agency/ The Vibe Marketer - Resources for people into vibe marketing/marketing with AI: https://www.thevibemarketer.com/ FIND ME ON SOCIAL X/Twitter: https://twitter.com/gregisenberg Instagram: https://instagram.com/gregisenberg/ LinkedIn: https://www.linkedin.com/in/gisenberg/ FIND MIC ON SOCIAL X/Twitter: https://x.com/Rasmic Youtube: https://www.youtube.com/@rasmic