Subscribe for startup ideas & tutorials on how to use AI to build & grow your startup. Hey, I’m Greg - CEO of Late Checkout and ex-advisor to Reddit, TikTok, and more. I post videos here from my podcast top 0.1% tech podcast called The Startup Ideas Podcast. You can also listen to The Startup Ideas Podcast on Apple Podcasts and Spotify. I keep it practical and never gatekeep the sauce. I share what I'm learning in real-time. Happy building
Product: The $999 AI Assessment Playbook: Sell, Run, and Upsell a Productized AI Audit Format: Digital download: the-999-ai-assessment-playbook-sell-run-and-upsell-a-product.pdf Price: $79.00 USD Description: Learn to land your first pay
In Hyper Agent's early adoption base, small shops have become surprisingly sophisticated very quickly and are running their operations in genuinely game-changing ways. The reason is simple: a 50,000-person company can't pivot on a dime — th
Airtable is a half-billion-dollar revenue business generating around a hundred million dollars in free cash flow per year, with over a billion dollars on the balance sheet. That financial strength allows Hyper Agent to be extremely generous
The design philosophy behind Hyper Agent is to serve the full range of users — from power users who want granular control over how the agent is set up, so they can squeeze out maximum bleeding-edge performance, all the way to people who jus
Greg Isenberg frames platform selection as an investment decision: when you pick a platform — whether it's OpenClaw, Hyper Agent, Codex, or something else — you are investing in an ecosystem. That framing matters because the switching costs
MembersAs users accumulate memories across many agents, managing those memories becomes its own problem. Hyper Agent recently shipped a defrag tool to address this. It auto-suggests related memories clustered by both keyword and embedding similari
MembersHowie pushes back on the idea that a visual, beginner-friendly agent builder means giving up control. In Hyper Agent you can still go deep: choose the exact model, write your own system prompt, select specific tools, and refine agent config
When Greg Isenberg asked why someone should invest in the Hyper Agent ecosystem, Howie pointed to the team's PLG pedigree. Airtable itself is a PLG product that scaled all the way up to Walmart-scale companies and fast-moving SMBs like Razo
Howie draws an explicit parallel to Apple and personal computing. Frontier agents — across every platform — are all going to get better, smarter, and faster. Hyper Agent isn't going to be the only powerful agent product. But just as Apple u
Howie's view is that the agent-builder landscape is fragmenting into two camps: easy prototyping tools that are fun to start with but don't scale, and heavy enterprise builder products with lots of configuration and control that are clunky
You don't need a perfectly formed idea to get started with Hyperagent. You can come in with something as broad as 'I'm kind of interested in building a solopreneur business' or 'I want to explore real estate' — or even pick one of Greg Isen
MembersHyperagent recently shipped a new onboarding flow that moves beyond a generic empty canvas. When you first land in, it suggests connecting your Gmail, Slack, Notion, and meeting tools like Granola. It then goes and reads through your recent
Users consistently report that Hyperagent is intuitive — you can usually just ask the agent to figure something out and it goes and does it, without needing to learn a lot of new configuration or UI. The harder challenge is actually identif
In a live walkthrough, Hyperagent was asked to build a custom skill integrating with the Twilio API — covering SMS, voice, and phone number provisioning with API key authentication. The agent researched Twilio's API documentation, planned t
One of the most powerful things Hyperagent can do is handle APIs it doesn't already have a connector for. You simply ask it to build a custom skill for a given API — say, Twilio — and it will go research the API documentation on its own, wr
Hyper Agent has a large number of connectors that work out of the box — you simply click a button inside the thread, tell it what tool you want to use, and it provides an OAuth link right there in the conversation. For example, you can say
MembersThere's a subset of emails that are low-stakes enough that you'd want an AI agent to not only draft a reply but, if it feels confident the situation isn't sensitive, just go ahead and respond automatically. Simple examples include inbound m
For X specifically, the winning strategy is simple: if you can get one good tweet out every single day, that's all it is. Content is a hits-driven business, and fewer high-quality hits are what actually matter. That means you can batch your
You can turn any thread or task in Hyper Agent into a recurring scheduled job simply by telling it in the thread: 'Can you turn this into a recurring daily email at 8 a.m.?' The agent will save the thread as an agent with a run schedule — e
When you get feedback on an AI-generated output — say a tweet that sounds too formal — you have several options right away. You can update the underlying skill based on that feedback, have the agent immediately regenerate a new draft that s
MembersWhen reviewing AI-generated tweet drafts, Greg Isenberg noted a common quality problem: the output felt too formal and stiff rather than sounding friend-to-friend, which is the tone that tends to perform well on social media. The fix isn't
MembersOne limitation of human-written tweets, as Greg Isenberg acknowledges, is that they're typically just hot takes with no supporting data. What becomes possible with a team of AI hyper-agents doing research in the background is that those tak
The biggest mistake people make when learning to build with AI agents is trying too sporadically — picking it up occasionally, then dropping it. Greg Isenberg recommends committing to a set period — 30, 60, or 90 days — with a literal calen
Greg Isenberg identifies two critical confidence milestones for new solopreneurs. The first is making your very first dollar from a stranger on the internet — no matter how small. That single transaction rewires your brain and proves the mo
MembersThe assumption most companies are operating under is that they're doing some things with AI — they have an AI center of excellence, maybe an AI feature — and that's sufficient. It's not. The wake-up call from investors to their portfolio co
A benchmark partner sent a wake-up call memo to portfolio companies urging them to radically rethink how they operate with AI — immediately. The memo ended with a parable: imagine two friends in roughly 2003, both selling knives door to doo
Getting great outputs from AI is like learning tennis — when you first start, you're bad at it, and you almost don't want to show up because you feel incompetent. But you go through the messy middle, you get better and better, and eventuall
A lot of people one-shot something with an AI agent, it's not quite as profound as what they hoped for, and they give up. But agents are powerful enough to do almost anything you want them to do — the issue is never whether they're capable.
The first version of a skill is never the final version. A dry run might produce an okay first effort — maybe 50% of the way there — but the whole point is that skills should be evergreen. Every time a skill is used, it can improve: either
MembersTo reliably mimic a specific creator's voice, an agent skill can be codified in a structured definition document. For Greg Isenberg's 'Contrarian AI Drafts' skill, this means spelling out explicit voice rules: the hook must land in the firs
Running an agent-first business is management 101, just applied to agents. If you're a CEO, you literally don't have time to look at every single thing every person in the company has done — and the same is true for agents. As you scale up,
MembersInstead of relying on yourself as the sole judge of every agent output, you can use a separate LLM as a judge that fires off and scores outputs along the dimensions you care about. You can then see the trend line of how an agent is scoring
Once you start running agents at scale, secondary capabilities like observability become really critical. It's not just about whether you can get one agent to do one thing — it's about how you oversee and run an entire business with many di
MembersHyper Agent has a feature called rubrics — essentially an eval rubric that lets you define what 'good' looks like for a specific type of task. For example, you could create a rubric called 'What makes great Greg Isenberg content?' and pin i
MembersAgents in Hyper Agent don't just accumulate new memories over time — they actively participate in their own improvement. The platform has an automatic self-improvement loop where an agent can suggest to you, 'Hey, maybe you should add this
Unlike Manus or Perplexity Computer — where the nominal experience is using one awesome agent directly — Hyper Agent is designed from day one with scalability and deployability in mind. Once you have an agent that works for you, you can dep
Against OpenClaude, the distinction is fairly straightforward. Hyper Agent is a more general-purpose agent platform that is turnkey, ready to go, safe and secure by default, and cloud-native. OpenClaude, by contrast, feels quite raw — when
Perplexity Computer and Manus are the closest competitive comparisons to Hyper Agent. Manus was truly groundbreaking when it first launched — it was the first real 'holy crap, yolo agent' that showed what autonomous agents could do, even be
MembersWhen you ask Hyperagent to create a skill — for example, one that posts in the style of a specific creator like Greg Isenberg — it doesn't just write a prompt that mimics the person. Instead, it draws up a plan: first research the creator's
MembersSkills are described as the single most important concept in the frontier-agent world. The underlying models are already generally intelligent — think of it like hiring an Albert Einstein who is brilliant in the abstract but doesn't know re
Any agent you build in Hyperagent can be pushed into Slack with a single click. The agent configuration lets you choose a channel, set the trigger mode (mention-only or listen to the full channel), and toggle auto-responses. In 'listen to c
Once you've built multiple agents, Hyperagent gives you a Command Center dashboard showing an overview of your entire fleet. Even in a light first-week account the dashboard surfaces metrics like Active Operations, Total Runs, Cost, and Ave
Another practical example of Hyperagent's built-in tool use: you can give it a house and tell it to redesign the space using interior photos or exterior shots sourced directly from Zillow. According to the creator, it handles this task real
Hyperagent ships with powerful tools out of the box, including Google Maps integration. One example: you can instruct it to find real Street View imagery of billboard locations — it knows how to locate actual points of interest, pull the im
MembersBecause Hyper Agent has full coding capability, it doesn't stop at producing a research brief — it can go ahead and build a v1 of the product it just researched. In this demo, it created a working app called Blockpulse: a clean, interactive
MembersAs part of its research brief on the hyperlocal real estate market report idea, Hyper Agent didn't just analyze the market from the top down — it went and looked up Reddit to find actual real estate professionals saying they need exactly th
When Hyper Agent researched Greg Isenberg's startup idea — hyperlocal market reports for real estate agents generated from public data — it produced a business case that highlighted what makes the opportunity compelling: it's a medium-sized
Hyper Agent runs a full computer in the cloud — not on your own machine — with a full sandbox environment and complete coding capability. You can connect it to all your accounts: Slack, Granola, email, and more. It can send things on your b
Hyper Agent is built on the same design philosophy and obsession with great UX that the Airtable team applied to the no-code app category ten years ago, now brought to agents. Just as Airtable distilled the technical complexity of a Rails a
MembersHyper Agent is built around the idea that most other agent products out there — think Open Cloud and similar — are more like Linux: powerful but raw. Hyper Agent is the Mac version of that. It's a multi-window canvas workspace where an auto
There's a parallel worth noticing between humanoid robots and AI agents. Robots are converging on a humanoid form factor not because it's philosophically ideal, but because all of our existing infrastructure — homes, construction sites, fac
The future of work, in one image, is an Agent Command Center — a single operator overseeing a fleet of agents handling Customer Intel, Content Production, Competitive Research, Lead Enrichment, Inbox Triage, M&A Diligence, and more. A termi
The game theory for large company incumbents is brutally simple: the CEO's choice is either pay a massive check — say $100 million or more — to an AI consulting or transformation firm, risk wasting the money, and maybe get fired over it, or
MembersThere are two distinct angles for building a highly valuable business with AI as the wedge right now. The first is product-led growth (PLG): build something that actually works, let people use it freely, and watch adoption compound. Even pr
To grasp how steep the AI adoption curve really is, consider the aggregate revenue picture. OpenAI and Anthropic alone are estimated to have a combined revenue of $80 billion-plus right now, up from essentially zero a few years ago. No comp
A Gartner chart on enterprise AI agent adoption shows fewer than 5% of enterprise apps had embedded AI agents in 2025, projected to reach roughly 40% by end of 2026, and over 90% by 2033. For context, SaaS took about six years to reach comp
As a concrete illustration of AI-agent value, the speaker described a recent board memo he sent to his full board and major investor list. Much of the memo was researched and crafted by their AI agent (Hyper Agent), informed by his own inst
The mental model most people use for AI pricing is anchored to traditional subscription software — recoiling at a $20/month Netflix price hike, for instance. That framing is wrong for AI agents. The right question is: how much would it have
MembersA lot of people complain about the cost per token of frontier models — Opus 4.5 and GPT-5 are expensive, and even open-source alternatives aren't free. But this objection misses the point. In the old days of software, a lot of things were c
MembersThe unit economics comparison between human agents and AI agents is dramatic. Cost per customer interaction drops from $3–$6 for a human to $0.25–$0.50 for an AI agent — an 88% reduction. First response time falls from over 6 hours to under
Using is believing. It is genuinely hard to fully grasp what frontier agents can do if you haven't gone hands-on and spent at least a full weekend with them — and that means more than a superficial, naive one-shot prompt like asking who wil
A trillion dollars is actually an underestimate. The real total addressable market for AI agents is essentially all white-collar labor — many tens of trillions of dollars, even in the western hemisphere alone. Once agents can do nearly anyt
Even in a basic synchronous chat interaction — not a full autonomous agent loop — you can give these frontier models the hardest, meatiest problems in domains like management consulting and receive genuinely expert-level answers. The model
Agents reached a genuine breakthrough roughly four to five months ago. Developers felt this most clearly with Opus 4.5, which set a new high-water mark: for the first time, the model could take on a task that would have taken a real human e
Sequoia has laid out a chart showing roughly a trillion dollars of addressable market within AI agents, spanning industries like consulting ($300B+), IT managed services ($100–200B), accounting ($50–80B), and freight ($100B+). The framing i
Howie Liu's new product, Hyperagent.com, is an AI agent builder that lets users create digital employees and build apps around different business ideas. To support the Startup Ideas podcast community, Howie is committing one million dollars
Sequoia has published a two-by-two matrix mapping agentic software opportunities across two axes: Copilot (human-assisted) versus Autopilot (outsourced), and Match (in-housed) versus Next Wave. The matrix maps this against Judgment versus I
MembersRather than using a traditional IDE, Howie Liu describes his own development workflow on Hyper Agent as running approximately 30 different Claude Code instances in parallel. Each instance is coupled to a browser and operates fully autonomou
MembersHowie Liu argues that the ~50% share of AI agent tool calls attributed to software engineering is actually an overestimate of how many developers are truly working at the frontier. Most engineers are still using AI in a 'gen one' mode — aut
Howie Liu's reaction to Sequoia's agent deployment chart is that it absolutely reflects the underpenetration of AI in industries that could already be disrupted or benefit with even today's AI capabilities. His view: if you took frontier ag
According to a Sequoia chart on AI agent deployment by domain, software engineering accounts for nearly 50% of all tool calls made by agents. The next largest categories are back-office automation at 9.1%, marketing and copywriting at 4.4%,
Howie Liu is the co-founder and CEO of Airtable. By the time of this conversation, Airtable had reached half a billion dollars in revenue, held a billion dollars in the bank, and was growing quarter after quarter — making Howie one of Greg
There is a clear startup idea sitting in plain sight: building company brains as a service. The concept is to come in and construct the foundational digital brain and content library that a company's entire marketing machine runs on. Greg I
MembersThe future of content marketing is having one really smart person serve as the foundational input for your entire content machine. That single person builds the digital brain — the library of ideas, frameworks, and voice — that the whole ma
The value of a content library or data asset isn't intrinsic — it depends entirely on what you can do with it. If a company buys a writer's entire catalog but lacks the operational infrastructure to scale it and squeeze all the juice out of
People instantly grasp the value of acquiring a bankrupt company's database. But the same logic applies to language, writing, and libraries of content. When you buy a catalog of writing, you are buying groupings of words, ideas, and crystal
MembersThe reason people are starting to say the company brain is the moat is because it compounds. If you hire someone who keeps adding to that company brain, and it grows more and more valuable over time, you can then use AI to slice and dice it
MembersOne of the highest-leverage marketing moves any large company could make is hiring a really thoughtful journalist or nonfiction writer in a relevant field, paying them a serious salary, and giving them one job: think deeply about the indust
Five pieces of content per day isn't an ambitious target anymore — it's the baseline expectation and just the infrastructure. If you're not posting at that frequency, you're not giving content a fair shot. The real question on top of that i
A huge portion of content struggles are self-inflicted. The key is getting out of your own way and hitting publish so you can discover what variables actually define you — your personality, your writing style, your voice. Generally speaking
People often assume that fear of writing is an everyday-person problem, and that highly successful founders or executives must find it easier. That's not the reality. Having ghostwritten for some of the most accomplished figures in business
Preferences like which adjectives you favor, what rhythm feels natural to you, which sentence structures resonate — these are things you can only discover by actually writing. If you haven't done the work to find out what you prefer, you ha
MembersAfter ghostwriting for over 300 founders, executives, and investors, the vast majority of the time the ghostwriter used essentially the same voice: authoritative, concise, with maybe a few vocabulary choices specific to the individual. Nine
When someone says they don't like their own voice but admires how someone like Greg Isenberg writes, the reality is that 99% of the time they can't articulate what makes their voice their voice, what they dislike about it, or even what spec
A client complained that an AI-written piece didn't sound like him because it lacked the citations and references he typically uses. Rather than rewriting the whole thing, the ghostwriter went back and added two or three sentences referenci
MembersIn ghostwriting, getting the voice wrong doesn't always mean rewriting the whole piece. One client sent back a draft furious, saying it didn't sound like him. Listening back to the original call revealed the issue: the draft leaned too hard
MembersWhen people say someone 'has a voice,' they treat it like some mysterious, unreplicable quality. In reality, voice is a bundled term made up of a handful of concrete data points. Word choice is one of the biggest levers: saying something is
The hope with AI in a content workflow is that the human is doing the writing, the thinking, the choosing. How do I see the world? What ideas do I agree with or disagree with? What are my personality details? Those decisions belong to the h
MembersEverything you create is, in some way, adding something to your digital brain — your accumulated knowledge, your personal language model of how you see the world. The question is how much it actually adds. If you spend 20 minutes repeating
The moment you embed a time-stamp into a piece of content — 'Substack just launched this new feature' or 'Here's how to start a Substack in 2026' — you've fundamentally changed its short, medium, and long-term value. That word, that date, m
A meme capitalizing on an Apple announcement gives you maybe 24 hours of distribution, and that's all the juice you're going to get out of it. You can't repost it a month from now, three months from now, or a year from now. In contrast, a d
MembersEverything you create sits on a spectrum between timely and timeless. Timely content has very high short-term value and very low long-term value — think a viral meme about a trending tech product that capitalizes for 24 hours, then falls of
MembersConsider two versions of the same short-form video opener. Version one: 'Here's how to start a newsletter on Substack.' That is commodity content. It's not worthless, but it's fundamentally interchangeable because many people can and probab
Video is no different from writing. Everyone uses teleprompters now. Short-form video content is increasingly scripted, and a script is writing — it's the same fundamental idea. Whether it's cutouts of long-form content or animation explain
One of the most common questions people ask is which is better: writing or video, podcast or long-form YouTube. The reality is the answer is yes — it's all of them, they all work. As an example, Greg Isenberg's business partner Dickie and h
There's a real irony when people say they want to write about everything. In doing that, you deprive yourself of ever going deep enough to arrive at something original. All you end up doing is typing and repeating commodity ideas all day lo
Original thinking doesn't mean writing a hundred-thousand-word book. A great example is Pythagoras saying 'the earth is round, not flat.' That's six words — fundamentally six words — that completely changed the course of humanity and that w
There's a faulty belief that if you were the one who wrote something, you own it and created something original. But writing a sentence like 'I woke up today and made coffee' isn't original — that's just an open-source grouping of words. No
There's a faulty belief that if your fingers touched the keys and typed a grouping of words, those words are yours. That's simply not true. If you write the sentence 'I woke up today and made coffee,' that is not attributable to you. It's a
If you're only creating commodity content — factual, useful, but interchangeable — with no personality component, you have to be honest with yourself: you are creating the lowest-value form of content in your topic. It's not worthless, but
Every piece here is something Greg Isenberg can talk you through, in their own words.