People don't buy products, they hire them to make progress in their life.

Think of your business like a race car: Growth Engine (self-sustaining loops), Turbo Boosts (one-off spikes), Lubricants (conversion/retention optimizations), and Fuel (capital, content, users). Get the engine right long-term — but Turbo Boosts win early.

Fast feedback loops are a growth flywheel. In Claude Code's early days, I'd fix every piece of user feedback within minutes — sometimes seconds. That speed makes people feel heard. Feeling heard makes people give more feedback. More feedback = better product.

People said they didn't want a dining room table in their new condo. I built one anyway — too small to actually eat at. Sales went up 22%. The table was the emotional bank account of their entire life.

Give engineers unlimited tokens at the start. Don't optimize early. A less capable model often uses MORE tokens to complete the same task — so the 'cheaper' model actually costs more. Use the best model and let people explore freely.

The biggest barrier to AI adoption isn't technology. It's organizational change. Vague AI mandates, procurement bottlenecks, and lack of guidance on high-impact use cases are killing adoption at most companies.

Three layers of AI safety: (1) alignment/mechanistic interpretability — watching what neurons are doing, (2) evals — a lab Petri dish with synthetic scenarios, and (3) behavior in the wild. As models get more capable, layer 3 becomes the most important.

I have never enjoyed coding as much as I do today, because I don't have to deal with all the minutia.

There are no new jobs. The context and outcome have existed for decades. What improves is just the technology's ability to deliver on them. Most jobs you can trace back 10 years and forward 20.

Southern New Hampshire University found 50-60 students attending fully online and paying full price when that wasn't even an option they built. They studied that anomaly and now have 200,000+ students.

AI won't replace PMs next year. But will the ratios between PMs and other teams shift? Yes. Will the required skills shift? Yes. Could they shift much faster than anyone anticipates? Probably.

Planning should take up no more than 10% of the execution time period. If you're planning for a quarter, your planning period should be slightly less than nine days.

Hypergrowth companies compress what takes others 5–10 years into months. That means you need to keep hiring leaders who've already seen the next stage of growth — because it'll arrive before you're ready.

Notion's ambassador program started with just 20 people in 2019 — the 20 most vocal users they could find on Twitter. They grew it slowly, inducting ~20 new people per month to preserve community quality.

At Ramp, our culture is velocity. It shapes every process and team ritual. It's how we develop our people. It's our solution to nearly every problem.

Starting a company has never been easier. Scaling one into a durable, high-impact organization has never been harder.

The Aladdin & Genie rule: If you tell the Genie 'make me taller,' he makes you 13 feet tall. AI doesn't know what you mean when you say 'you know what I mean.' Be specific. Provide references. Provide context.

If you can leave a system alone for a long time without it hurting your business, ask yourself: is it actually worth rewriting at all? Rewrite if you can't add what you need. Don't rewrite because the engineers find it annoying.

You want the shortest path from a crawl point.

Ask your leaders what they want before you build strategy. It's amazing how few people do this. Senior leaders welcome it — it gets their creative juices going. And it's not a sign of weakness. It's a sign of strength and humility.

We have more than 15,000 customers who have deployed agents on Azure. The number of agents actually running? Millions.

The Curiosity Loop: pick a specific question, curate the right people (subject matter experts + people who know you), make the ask lightweight, and close the loop by thanking them. It's user research for your life.

Conway's Law: you ship your org structure. How your teams are organized and how they relate to each other determines what you build. Changing outcomes without changing the structure of relationships is nearly impossible.

For the vast majority of people, they probably do not need a coach. If you're thinking about getting one, first ask: have you explored mentors, courses, communities, or a curiosity loop? A coach is often not the best option.

If you're an experienced IC who loves the work, don't let anyone pressure you into management. Use your expertise to negotiate strong comp — the data is on your side.

You don't have a choice. AI is going to disrupt every category in the most aggressive, violent ways. The question isn't whether to get in. The question is whether you're willing to pay the price to actually be competitive — or whether you're going to sprinkle some AI on top and call it a strategy.

Stop trying to force product-market fit. If it's extremely difficult to sell the marginal customer, you won't build a huge business. What you need is the customer that's surprisingly easy to sell — that's the signal.

Crazy Big Goals work because they make you feel completely inadequate. And that feeling makes you work hard enough to will them into existence. A reasonable goal you can hit easily doesn't produce the same obsession. Set goals that scare you.

The end of the busy work era is coming. You can't sit in a corner doing something that doesn't move the needle. You need to be thinking about end-to-end workflows and how you can bring more impact.

Hesitation is the most destructive thing a CEO can do. Both options are bad. That's the point. The only value you ever add as a leader is making a decision most people don't like. If everyone agrees, you added nothing. You need to run toward the pain.