Dharmesh ShahCo-founder & CTO, HubSpot

We were two guys in Boston with a thesis and not much else. I was still in grad school, writing my blog OnStartups at night. The blog let us test ideas, learn SEO, and pull people in before we had a real product. Traffic, then email subscribers, then early customers. It all started from that.

Those companies were spending on things that did not really move the needle. I was just at home writing useful content and learning how Google worked. If you consistently publish pages that deserve to rank, and you structure them well, you slowly win. It looked like magic, but it was just focus and repetition.

SEO was about ranking in Google’s ten blue links. Now people ask ChatGPT or other AI tools and get an answer, not a list. AEO is answer engine optimization. Instead of only thinking about keywords and titles, you think in questions and structured answers so the model can trust and reuse your content inside its replies.

First, let the AI crawlers in. If you are blocking bots, make sure OpenAI’s search bot can see your site. Second, reframe key content as clear question and answer pages. Third, keep doing white-hat SEO so you show up in Bing and others. You are still building authority, just for a new distribution channel.

Most of us already know what to do. We do not lose because of lack of ideas. We lose because the intensity is not there. Intensity means more reps, more iterations, more at-bats than the next person. Not one heroic week of work, but years of showing up and pushing on the same core strategy.

Volume and consistency, for sure. I was not the best writer on the internet. But I wrote all the time. Some days I would publish multiple posts. Over months and years, that adds up. Every post is a small experiment. When you stack hundreds of those experiments, your odds of finding winners go way up.

Pick something you like enough that you can keep putting reps in when it is not fun. If you hate the domain, your intensity will fade. I tell people to choose a market they can be curious about for a decade. Margins matter, but if you give up after two years, they do not matter at all.

An efficient market is where every valuable transaction that should happen actually happens. Often buyers and sellers do not know each other, or they cannot trust each other, or they cannot price the deal. A good community plugs those gaps. You gather the right weirdos, reduce friction, and then participate in the value that gets unlocked.

They did not start with their product. They started with the algorithm and the audience. Then they reshaped the product to fit what the algorithm wants to show to bored humans. Most of us do the reverse. The lesson is simple. Respect distribution. Study what the system rewards, then design your format and hooks around that reality.

You need a belief that the universe rewards honest reps. Early on, eight people read my posts, then sometimes five. It is emotionally hard. I tried to detach from the day-to-day numbers and focus on iteration. Make the next article slightly better. Make the site slightly clearer. Over time, those small upgrades compound in surprising ways.

I want my notes, chats, and email to be one big, flexible memory layer that I control. If all my history only lives inside one app, I cannot combine it with other models or tools. My own interface lets me search across years of conversations, then pass that context into whichever AI is best for a given job.

Vibe coding is using AI to build software more by feeling and conversation than by traditional engineering. It is great for tools you build for yourself, for small internal apps, or for experiments. It is less great if you want to run a huge, long-lived production system. At some point you still need structure and tests.

Side projects are my way to explore ideas with low stakes. A domain is a cheap option on a possible future. Sometimes nothing happens. Sometimes a project like Wordle clone money pops up and makes real revenue. More importantly, each project sharpens my instincts and keeps me close to the actual act of building.

Induction says if something is true for step one and true for step n to n plus one, then it is true for all steps. I translate that into: start, then always find a next step. If you can keep making honest, slightly better iterations, the long-term result almost takes care of itself. The hard part is not stopping.

My main window into the world is a couple of screens. People appear as little talking heads. Code, products, and communities are like levels in a game. I treat business as a long game I get to play every day. That mindset removes some fear. You are just learning the game and trying different moves.

If money stops you from doing the thing you love, it was never about the craft. I like writing code, building products, and thinking about markets. That is my version of playing pickup basketball. Jordan had “for the love of the game” in his contract. I just gave myself the same clause for building companies.

I think we are still early. AEO will create a new class of experts and agencies. Agents will become real workers inside companies. Tools that manage knowledge, memory, and creativity will be huge. The opportunity is to combine old truths like iteration and intensity with these new capabilities. That mix will create the next generation of winners.