Greg Isenberg

Greg Isenberg says a marketing engineer runs dozens of experiments a day with AI agents. He traces the job from one big commercial a year in the Mad Men era, to the digital marketer, to the growth hacker running a few experiments a week, to an AI-assisted marketer, and now to this role. Jon Noronha buys the thesis. At Optimizely, before Gamma, the hope was a hundred tiny versions of a website so each visitor saw a different page. That mostly did not happen. Sites and apps are still mostly one version, except feeds such as TikTok and Instagram Reels, where companies spend heavily to personalize. A backlash risk, including the bad feeling of surge pricing, is part of why wild personalization s

Jon Noronha says a startup should build its own agent harness when the job looks very different from plain coding. For Gamma the work is subjective: two people disagree about what good looks like, a human stays in the loop, and the person using the tool cares about the specifics. He treats those traits as the test.

A personal video game is one opening he is bullish on. Coding agents already make games, and they are weaker at taking in what the player wants, the way non-coders shape Roblox or Minecraft. He can imagine a harness that runs inside a 3D world. The practical rule is to build something that looks nothing like what everyone else is shipping, then keep moving, because copies trail b

The Divot introduction credits pattern recognition with taking Greg Isenberg from founder to trusted advisor. It says he has met with some of the biggest and best companies in the world and has helped them on their strategies around AI. Today he leads Late Checkout, a portfolio of AI-native businesses powered by agents and internet communities.

Idea Browser, Greg's Letter, and the Startup Idea podcast are where those ideas go out. Through them he has shared over 2,000 startup ideas and built an audience of millions of followers. The introduction places this conversation in July 2026, with Greg Isenberg.

Greg Isenberg calls it the ACP framework: audience, community, product. Step one is to figure out the niche. Step two is the framework. He says a business in the AI age should follow it. Every business he starts begins as an audience. It starts as an X account, an Instagram account, or a TikTok account, and it starts audience first.

The first 90 days are a goal: find the format that gets that niche to follow him. He gives the reason in the same stretch. In the AI age, when you can build anything, distribution is the part he names. The account comes before the product, because every business he opens starts there.

Greg Isenberg says we are entering what he calls the age of the multi-preneur. A multi-preneur is someone who owns multiple businesses and multiple lines of revenue. For creators, he says the multi-preneur still has one product, one service, or one business unit that drives the lion's share of revenue.

He also says they are going to have multiple businesses. If agents can run those businesses, and the owner can create a high-margin business with a whole system around it, he calls that setup pretty cool. Multiple lines of revenue, one of them carrying the lion's share, and agents on the rest, is the age he is naming.

Greg Isenberg says the builder-distributor combo is probably the most lethal type of person in the AI age. Anyone can build. He starts to say you need to speak English, then he corrects it: you can speak any language and prompt to build. The way to compete is to understand distribution quite well.

Scroll-stopping Instagram posts and videos are one path. Kling, a video model, is another, used for commercials that then connect to Meta ads. The builder-distributor understands distribution, how to capture attention, and how to convert that attention into revenue. The same person understands how to drive value from the product side. The product has to be good, and the bar for good products is in

Greg Isenberg says he probably measures life by what family says about him. The words he hopes God says are you did a good job. He states the family standard with that doubt still attached, and he states fatherhood outright. A last dying thought, in his words, will not be wonder what Sand Hill Road thought of him in a pitch.

He is a father now. The child is a year and a half old, and another one is due in a month. He says being a dad completely changes how you think. The sentence about what the children think, what the partner thinks, and what the family thinks is a different voice, and that voice says the care becomes so important.

The reply from that other voice is that agents will be ru

On Divot, a study says the vast majority of people do not live within their means. A large percentage of really wealthy people still do not live within the incredible life they created, because they are always trying to get more and always trying to go for more. Derek Andersen lives in Northern California and says what is enough, and what is right for him, is a question only he can answer.

He used to rent down the street from Mark Zuckerberg, who owns like eight homes there. Every day the drive home went past the house, where a security guard would be. They live on the same street, both sleep on a bed, and both breathe the same air. The drive still left open whether the rented life is enoug

On Divot, the venture-backed model works for less than 1% of all companies that are created. Investors back only a small percentage of the companies they look at, and only a small percentage of the companies they back actually work and return the money. Investors know the pattern, and the whole model is baked into the pattern, so the vast majority of anything anyone is building should not use the venture-backed model.

The path narrows at each step. Investors look at companies and back a small percentage of them. Only a small percentage of the backed companies work and return the money, which leaves a very, very small group of people. The venture-backed model only works for a tiny fraction o

Quebec is one of the places where people drink the most Pepsi on the planet. Quebecers drink Pepsi instead of Coke because marketing in the local language resonates with them.

On the show, a boring business keeps an opening when it stays in one geography. Competition might be heavy in South Florida, the Bay Area, and New York. Canada and other geographies still leave room to take a concept and localize the language. Someone from Quebec points to research on Pepsi and Coke, and marketing in the local language is what resonates with people there.

Localizing the language is what lets a boring business land in a geography like Quebec with less of the competition found in South Florida, the Bay

On the show, work splits into regular work and creative work. Regular work such as SEO is probably 12 to 18 months away from running on its own. Creative work still needs a person whenever creativity is part of the task, even at 1%.

People used to hire SEO agencies to do SEO. The work meant creating blog posts, optimizing the blog posts, and doing a set of tasks that were pretty binary, because the results meant a site went up in Google, went down in Google, or stayed the same. Regular work of that kind is probably 12 to 18 months away from being pretty autonomous and doing its own thing. Creative tasks like tweets, video content, branding, and copywriting keep a creative element. When crea

On Divot, a morning spent with Grok 4.5 produced an AI co-founder that built a startup overnight for the equivalent of $2 an hour.

The builder had dismissed Grok and treated the market as an open AI, anthropic, Google world. After playing with Grok 4.5, the model was connected to a Hermes agent and given to a personal assistant running 24/7. The assignment was one startup idea: build features, optimize based on analytics, do SEO, and keep running while the builder slept. By morning the output looked capable and good. Sitting on the 38th floor and looking at the ocean, the builder had a working AI co-founder, paid about $2 an hour, super intelligent, and creating a startup during the hangout