GREG ISENBERG

how to use perplexity computer to spin up digital employees that automate your work 24/7 1. connect your email. give it a list of prospects, what you sell, and your tone. it finds the right contact at each company the person who actually signs deals), researches their pain points, and drafts outreach that sounds like you 2. ask it "what am I not asking you that could make me more money?" it told me to monitor competitors weekly, build follow-up sequences on day 3 and day 7, and target companies whose budgets are already hot. one prompt changed the whole session. 3. set up daily competitor monitoring. pick 5 competitors. every morning it checks their pricing, features, content, and X mentions. changes get summarized. silence when nothing moves. delivered to your inbox at 8am. 4. need to fundraise? describe your startup once. it builds a 50-VC spreadsheet with fund size, thesis fit, the right partner, and their recent activity. 5. turn a podcast episode or loom video into a blog post, tweetable quotes, and a carousel. one upload. 6. reverse engineer any competitor's SEO strategy or pricing page. see exactly where you're leaving money on the table. 7. hiring? describe the role. it finds and ranks 50 candidates in minutes. 8. it orchestrates 19 models in parallel. one for reasoning, one for code, one for research, one for images. it picks the best model for each step automatically. 9. start thinking in recurring workflows that compound every day without you (this is relevant for perplexity computer or any tool you use) episode is live on @startupideaspod (full live walkthrough) send this to a friend who keeps saying they want to start using AI agents. watch

how to build successful software startups 2005–2025: great ui first, api as a bonus 2026+: great api first, ui as a bonus

the future of building saas this is how 3-person teams build 100m companies: 1/ start with a sub-niche inside a big market 2/ map their daily workflow end-to-end 3/ identify where money changes hands 4/ spot the repetitive mechanical steps 5/ quantify the cost of those steps 6/ create scroll-stopping content around that workflow 7/ study which posts get saves, replies, and dms 8/ double down on the organic angles that convert 9/ run paid ads on proven organic winners 10/ capture emails from day one 11/ manually perform the workflow yourself 12/ document every step precisely 13/ separate judgment tasks from mechanical tasks 14/ turn mechanical tasks into structured agent workflows 15/ design agents to complete full tasks, not suggestions 16/ connect to real tools: email, slack, notion, crm, stripe 17/ add orchestration, retries, and verification checks 18/ store user preferences + long-term memory 19/ launch narrow with high-touch onboarding 20/ publish measurable proof: revenue, hours saved, errors reduced 21/ move pricing from per-seat → per-task 22/ shift to outcome pricing tied to revenue created 23/ increase pricing as value compounds 24/ expand into adjacent workflows within the same niche 25/ orchestrate multiple agents across the full lifecycle 26/ build switching costs through data + memory 27/ turn power users into public case studies 28/ hire operators from inside the niche 29/ reinvest profits into distribution + product depth 30/ become the default execution layer for that sub-niche

how to use openclaw to spin up 24/7 digital employees and build cash-flowing assets: 1. spin up openclaw (mac mini, vm, orgo, whatever) in a workspace so you can run 5–10 machines at once (main agent + sub-agents) 2. pick one boring workflow inside one industry (distributors, real estate, insurance, law firms) 3. map the workflow tip-to-tail (email/trigger → legacy software clicks → downloads → parsing → upload to crm) 4. use claude code to build the “under the hood” python pipeline (openclaw becomes the operator + trigger, code does the heavy lifting) 5. productize it as a repeatable bundle: “setup + 30 days management + new workflows each week” 6. use upwork as the lead source and the sandbox (it tells you what people pay for right now) 7. turn the best-paying workflow into a vertical workspace: 20 skills, 8 sub-agents, one invite link 8. sell it to bigger companies as “ai employees for this department” (with clear outcomes + SLA) "BuT yoU cAn'T bUiLD a BiG coMpaNY dOInG uPwoRk deAls" think about it like this “how does a $1k automation gig turn into a big company deal?” like this: 1. upwork gives you paid reps + proof someone pays for the workflow 2. those reps become case studies (“saved 12 hrs/week”, “uploaded 5k records/day”, “reduced ops errors by 80%”) 3. you stack 5–10 workflows in the same vertical 4. now you’re selling a package and not a one off deal which is tough 5. bigco buys packages because procurement 6. understands scopes + outcomes openclaw is the wrapper. claude code is the factory. sub agents/skills are the workforce. the vertical bundle is the product. episode is live on @startupideaspod i will never gatekeep i want to see you win in this openclawed world i am rooting for you watch.

📝 The Lobster Internet Big news fro Kimi, you can now deploy OpenClaw to Kimi in seconds on the cloud (not mac mini required). You get a 24/7 AI assistant without a terminal in sight and . Feels like something anyone would use. Got me thinking of how this whole openclawification of the internet might play out. I call it the lobster internet: phase 1 - hackers at home openclaw runs on mac minis, raspberry pis, homelabs. tinkerers wire telegram bots and local models together. messy, powerful, niche. not for mass market yet phase 2 - "claws" hosted in the cloud kimi and others put openclaw-style agents in the browser with storage, skill libraries, uptime. agents become accessible to anyone with a tab open. phase 3 - multi-model orchestration anthropic, openai, google, grok etc all ship hosted agent layers. model wars move from “who’s smarter” to “who orchestrates better" because thats what really drives output fro 2026 and beyond phase 4 - verticalized bundles real estate bundles. ecommerce growth bundles. hedge fund research bundles. pre-wired workflows + memory + distribution. plug in and go. tons of opportunities here for founders. adding some of these ideas on Ideabrowser.com phase 5 - agent as employee agents get job titles: researcher, growth lead, qa tester, ops manager. dashboards show output per agent. founders manage clusters/ phase 6 - agents as the new saas instead of paying for tools, you pay for outcomes. so instead of buying “crm software” you'd rather buy “an agent that closes 20 deals/month.” software shifts from interface to executor. phase 7 - outcome-based pricing agents charge per lead booked, bug fixed, page ranked, deal sourced. subscriptions fade. performance pricing expands. phase 8 - personal agent layer every operator has a persistent agent that knows their style, context, network, data. it travels across tools. it drafts, negotiates, researches, builds. phase 9 – agent-native apps (the parallel internet) this fascinates me. every major human app gets an agent equivalent. agent twitter. agent notion. agent polymarket. agent recruiting. moltbook is the early signal. there will be moltbooks for XYZ. apps stop being for humans first. they become programmable surfaces for agents. lots of opportunity here. phase 10 - onprem becomes the new cloud enterprises run secure, local-first agent stacks. private models, internal data, certified workflows. “local” becomes a selling point. phase 11 - regulated agent frameworks finance, healthcare, gov adopt audited agent systems with traceability and compliance built in. certification becomes table stakes. phase 12 - invisible infrastructure memory layers, mcp routing, storage, uptime abstract away. running agents feels like electricity. people think in outcomes, not tools. phase 13 - agent-native companies 3 humans, 300 agents. org charts show humans directing systems. leverage becomes extreme. headcount becomes blurry. phase 14 - leverage premium humans the scarce skill becomes taste, orchestration, positioning. less typing. more directing. more judgment. the best operators design systems others rent. from hobbyist bots → to hosted infrastructure → to vertical stacks → to agents replacing saas → to the agent internet → to agents coordinating the economy itself and we’re somewhere between phase 2 and phase 4 right now. I'm rooting for you, Greg Isenberg http://x.com/i/article/20230726603185397…

"The first hotel on the moon" is called Gru and details have emerged It's backed by YC and NVIDIA and they are charging a casual $416k/night Really cool project regardless.

claude cowork crash course in 49 seconds for building out your ideas

the truth is no matter how hard you try you’ll never be able to keep up with 100% of what’s going on in AI right now there’s just too much action right now

i’m over the phrase “vibe coding" because it does a disservice to what's actually happening right now many people are building real products that could scale (not exactly ViBeS~) this is the start of a new class of serious, cash-flowing software companies

📝 this is what's keeping me up at night these days... (ai agents, openclaw etc) AI skills marketplaces will look more like app stores circa 2009 than saas marketplaces. ugly, fragmented, wildly lucrative, and full of one-person shops doing $5–10m quietly. the value of clean proprietary data will spike as generic data becomes fully commoditized by models. there will be a $1b+ company that is just a vertical agent pack for one industry that no one on X talks about. we're entering the era of "one-hour companies." idea at 9am, landing page at 9:15, product built by 9:45, first customer by 10. most will fail fast. but the sheer volume of experiments means we'll see more breakout companies in the next 2 years than the previous 20. we're about to see the first wave of AI agents hiring other AI agents. basically agents autonomously spinning up specialized agents for subtasks and shutting them down when the job's done. I think the "org chart" is about to look like a serverless function. agent colocation becomes a thing. companies will want their AI agents running physically close to their partners' agents for speed and security. we'll see agent-specific data centers where proximity to other agents' infrastructure is the selling point. digital real estate gets literal again. founder-market fit gets replaced by founder-agent fit. the best founders will be the ones who can orchestrate a fleet of agents toward a vision, the same way the best film directors get performances out of actors. it becomes a creative leadership skill. OpenClaw got 60k github stars in 72 hours. anthropic sent a trademark complaint. openai bought the creator. that sequence tells you everything about where power is shifting, from the labs to the builders to the community. the lobster won. someone is going to build a billion dollar business doing nothing but converting legacy SaaS companies to outcome-based pricing models. the entire software industry needs to reprice itself in the next 18 months and nobody knows how. most saas companies will realize too late that their product is just a collection of workflows that can be rewritten as skills.many die, but top saas companies get bigger in an ai age and basically pivot to agent companies the companies that use AI the LEAST in their customer-facing experience will charge the highest prices. "no AI involved" becomes a luxury positioning statement someone will build the operating system for agents inside enterprises - permissions, memory, audit logs, billing and it will be enormous. this is the most asymmetric window i’ve ever seen in business nostalgia becomes a premium product category. in a world drowning in AI-generated everything, anything genuinely handmade, slow, analog, and imperfect becomes luxury. vinyl already proved this. expect "human-made" certification labels on everything from furniture to software to legal documents. the agency model is dead and nobody's had the funeral yet. design agencies, marketing agencies, dev shops etc all selling labor arbitrage that AI just collapsed. the agencies that survive will sell taste, strategy, skills and move from hours. Some will be extremely big. I’m betting on this as i co-founded an 8 figure agency and we''ve built some of the biggest interfaces for tech companies, but now have moved into skills, agents etc (see LCA) The highest-margin startups of 2026 will be agent wrappers around vertical workflows. the most underpriced asset in the world right now is a niche audience of 5,000 engaged people. that used to be too small to monetize. now you can code with AI a custom app for them in a weekend and run the whole business with agents. micro-monopolies everywhere! the entire insurance industry is about to be repriced by AI agents that can assess risk in real time instead of using 30-year-old actuarial tables. the first AI-native insurance company will be 50 people and outperform carriers with 50,000. we're about to see "agent injection" become a bigger cybersecurity threat than phishing. when agents have system access and make autonomous decisions, poisoning their context window is the new attack vector. the security industry hasn't caught up just yet. the fastest path to wealth right now: find an industry that still runs on phone calls, faxes, and spreadsheets. build the AI-native version "ambient businesses" become a category. companies that run entirely in the background with zero daily human input like an agent monitors a market, identifies arbitrage, executes trades or purchases, fulfills orders, handles customer service. the founder checks in once a week. Dont think its crazy some of these will do 8 figures. the first mainstream "agent subscription box" is coming - a curated bundle of pre-configured agents for specific life stages/milestones/outcomes. example: just had a baby? here's your sleep tracking agent, pediatrician scheduling agent, and expense forecasting agent. $29/month. the "API economy" evolves into the "agent economy." instead of developers integrating APIs, agents will discover, evaluate, negotiate access to, and integrate with other agents' capabilities on the fly. the entire concept of a fixed tech stack is dissolving before our eyes. AI-native elder care is going to be one of the most impactful and profitable categories of the decade. 70 million boomers aging into care needs, chronic staffing shortages, and agents that can monitor health, manage medications, coordinate with doctors, and provide companionship 24/7. the person who builds this with empathy will change the world. Obsidian as a second brain + claude code to execute. I recorded a pod on this. Will post how to do it on Monday @startupideaspod . Game changer because it helps your LLM know you way more intimately. karaoke bars, escape rooms, and immersive experiences see a massive boom. when digital entertainment is infinite and AI-generated, the scarcity premium shifts entirely to things that require your physical body in a specific place with other humans. the experience economy isn't coming, i think it's already here and accelerating. the build in public movement evolves into build with your audience in public. founders will share their agent configurations, let their community vote on features in real time, and ship updates the same day. the line between creator, founder, and community leader fully dissolves. every vertical is about to get its own claws - an OpenClaw for sales, for legal, for healthcare, for logistics. each one built by one obsessed person who knows that industry cold. "slow software" becomes a thing. deliberately human-paced tools that refuse to automate everything. a writing app that won't autocomplete. a project manager that forces you to think before adding tasks. the counter-reaction to AI maximalism will have passionate adopters and real revenue. digital hygiene becomes really important like what your agent has access to, what it remembers, what permissions it holds and people start doing quarterly "agent cleanses" the way they do closet cleanouts. AI-native churches, therapy practices, and support groups will serve more people than their human-led equivalents within 5 years. because they're available at 3am on a tuesday when you actually need them. AI agents will start filing patents. when an agent autonomously designs a novel solution to an engineering problem, who owns the IP? current patent law requires a human inventor. this gets tested in court within 2 years and the ruling reshapes innovation incentives globally. an AI agent will write a bestselling book within 24 months. the controversy will be enormous. it will still sell a million copies. fork a business becomes as common as forking a repo. someone open-sources their entire business playbook — agent configs, pricing model, marketing templates, supplier contacts and hundreds of people launch local variations. franchise economics without the franchise fees. cold email dies. every inbox has an agent gatekeeper. the only way to reach someone is through warm intros or being so interesting their agent lets you through. relevance becomes the only deliverability metric. angel investing becomes something normal non tech retail people do when agents can find, evaluate, and manage micro-investments in vibe-coded startups automatically. i think the 1000 true fansthesis becomes 100 true fans. agents cut your costs so dramatically that 100 people paying you is a real business. the minimum viable audience shrinks again. holdcos become more of a thing. more common to have a suite of the last satisfying job is one where you make something with your hands that someone can hold. woodworking, ceramics, cooking, tattooing. demand for apprenticeships in physical crafts explodes. Engineers deal with the loss of identity of being a programmer because manual coding dies pretty much We’re about to witness the greatest inventions in health and medicine over the next 10 years thanks to AI agent-to-agent gossip becomes a real phenomenon. your purchasing agent tells other agents "that vendor's API was unreliable" and suddenly a company's reputation spreads through the agent network without a single human involved. the loneliest generation in history is about to become the most "accompanied." always-on AI companions that remember, adapt, and care will become the primary relationship for millions of people. this will be simultaneously heartbreaking/interesting and enormous as a market. i keep meeting people who are "waiting for things to settle down" before they start building. things are not settling down. this is the new normal. the chaos IS the opportunity and the window is shrinking every single day. i'll stop for now, the coffee was flowing so had to get this off my chest more ideas like this on ideabrowser.com i hope you get some sleep. if you're anything like me, you cant shut off your brain. i'm rooting for you. Greg http://x.com/i/article/20252610679709900…

how to use claude code + outscraper + crawl4ai to build a profitable online directory in 4 days for under $250 1. scrape 50k–70k raw records with outscraper 2. use claude code to clean, dedupe, and structure the data in passes 3. run crawl4ai to verify live sites and filter junk automatically 4. then enrich one layer at a time: inventory, pricing, images (claude vision), amenities, service areas “BuT aN oNlInE DiRecTorY isn’t a ReAl bUsiNesS” first it’s traffic + leads. then it’s premium listings. then it’s vertical saas (crm, quoting tools, booking). then it’s agents handling intake, routing, follow-ups. then it’s the transaction layer. episode is available on @startupideaspod you start with organized data and you end with owning the workflow directories are kinda like a trojan horse. watch

how to make your ai app not look like it was made with ai (weavy ai, claude, figma etc) use weavy ai to generate custom assets, color systems, and logos across models like flux + ideogram, then compose it properly in figma then paste into claude code/google ai studio etc watch

The more the merrier 🦞

i exist to make you win in the AI age i will give you startup ideas and show you how to use AI clearly and whenever you need a tad of clarity, you'll see a tweet, a vid of mine at the EXACT right time with the clarity and ideas you need if you’re winning, i’m doing my job

we’re still pricing ai tokens like software subscriptions but most companies will soon price them like labor $200/month feels expensive because we compare it to saas $50k/month will feel cheap when we compare it to headcount

how to use obsidian + claude code to build a 24/7 personal operating system and build your startup: 1. write everything in markdown (daily notes, projects, beliefs, people, meetings) 2. link your notes together so they mirror how your brain actually thinks. 3. install obsidian cli so claude code can read your entire vault + the relationships. 4. stop reexplaining projects every session. use reference files instead. 5. build custom slash commands: /context → load your full life + work state /trace → see how an idea evolved over months /connect → bridge two domains you’ve been circling /ideas → generate startup ideas from your vault /graduate → promote daily thoughts into real assets 6. keep a strict rule: human writes the vault. agents read it, suggest, execute. 7. let claude aka clode surface patterns you’ve been unconsciously circling for years. 8. delegate from inside your notes. one sentence in obsidian → agent handles the rest. 9. treat writing as leverage.the more you write, the more context your agents have. 10. understand this:markdown files are the oxygen of llms. i really enjoyed seeing how to use obsidian thanks to @internetvin vin uses ai like a thinking partner wired into his life’s work. 99.99% of people won’t do this because it requires reflection + setup. but once the vault exists, the agent stops being generic. it starts thinking in your voice. episode is live on @startupideaspod (more there) this one is different. send this tweet to a friend. im still processing how game changer obsidian + claude code is, maybe you too watch

"Things I wish someone told me before I almost gave up on OpenClaw"

how to use claude code + outscraper + crawl4ai to build a profitable online directory in 4 days for under $250 1. scrape 50k–70k raw records with outscraper 2. use claude code to clean, dedupe, and structure the data in passes 3. run crawl4ai to verify live sites and filter junk automatically 4. then enrich one layer at a time: inventory, pricing, images (claude vision), amenities, service areas “BuT aN oNlInE DiRecTorY isn’t a ReAl bUsiNesS” first it’s traffic + leads. then it’s premium listings. then it’s vertical saas (crm, quoting tools, booking). then it’s agents handling intake, routing, follow-ups. then it’s the transaction layer. episode is available on @startupideaspod you start with organized data and you end with owning the workflow directories are kinda like a trojan horse. watch

the hardest part about building a company right now is building something worth building

you dont want to hear this but saas will implode few will become agent saas and be 10x bigger most will be hurt many layoffs next 18 months is the greatest time to be building a startup build an agent first startup now before others get in get your hands dirty ship