Both increase my curiosity. Neither closes the decision. In the end, I’m trying to figure out whether the underlying conviction is real and whether it holds up when I do my own work. Sometimes it does. Sometimes it doesn’t.

Selective programs or well-known investors are different. I don’t know their internal debates. But I do know they see thousands of companies and choose very few. That tells me this team survived a competitive filter. That’s a statistical signal.

A personal introduction from someone I know carries one kind of weight. If I understand how that person thinks and how selective they are, I can calibrate their conviction. I’m hearing the opportunity through a lens I trust. That’s borrowed judgment, but it’s informed.

When I’m evaluating a company, I’m not starting with the category. I’m not asking whether SaaS is hot or whether robotics is having a moment. Early stage is too uncertain for macro enthusiasm to carry much weight.

I usually start with how this got to me.

But if someone is willing to not ride the dominant wave, it usually means they’ve thought harder about what they believe.

Conviction shows up in what you choose not to do. That restraint can be a signal of something investable.

I also see a lot of companies adding AI because it’s the easiest way to feign relevance.

So when I meet a team that isn’t centered on AI at all, I pay attention. That doesn’t mean it’s good. It doesn’t mean it’s fundable.

"When everyone zigs, zag." It’s a cliché, but that's for a reason. It's also solid advice.

Right now, everything is AI-infused. I mean everything. I see a lot of smart founders building great tools with it or for it.

None of this replaces data. But these instincts help me know when to lean in before the numbers are obvious, and when to step back, even if the story looks great on paper.

And I still pay attention to whether they simplify or complicate explanations. Hard problems are not an excuse for opaque answers. People who can take a complex system and make it legible are usually the ones who understand it.

Another is how they talk about customers off-script. It’s one thing to pitch a persona slide. It’s another to spontaneously tell stories about users, their problems, and how the product fits into their day. Founders who live in those details build products people actually use.

One is founder energy in follow-ups. A first meeting is always rehearsed. The second or third call, when you’re knee-deep in questions, shows you who’s driving. Are they energized by the conversation? Do they pull you forward? Or do they wait for you to chase them?

I love a good dashboard as much as the next investor. But over the years, a few non-quantitative signals have stayed with me. They’re not scientific. They’re just patterns you see when you’re around for long enough.

If you’re not prepared for that, both financially and psychologically, the strategy can feel broken well before it’s had time to work. Patience is a big part of the job.

It’s common to look at the portfolio, see more capital deployed than value reflected, and wonder whether something is off.
Most of the time, it isn’t. That's simply how it plays out. A small number of companies account for most returns, and they usually do so later than expected.

One of the harder parts of early-stage investing is living through the middle years of the portfolio. Losses tend to show up early, while the successes take much longer to reveal themselves. That’s the J-curve people talk about, and it’s not theoretical.

One of the harder parts of early-stage investing is living through the middle years of the portfolio.

Losses tend to show up early, while the successes take much longer to reveal themselves.
That’s the J-curve people talk about, and it’s not theoretical.

A great deck might get you in the room, but clarity in those deeper conversations is what builds trust over time. That’s something l I find myself paying the most attention to.

They can explain why a decision was made, what assumptions it depends on, and what would need to change if those assumptions turn out to be wrong. There’s a steadiness there that doesn’t come from polish.

Others are leaning on them to avoid sitting with the harder parts of the business. It’s not about having perfect answers. It’s about whether someone understands where the difficulties are.

The strongest founders I meet are comfortable naming the things they haven’t solved yet.

That’s when questions about customer acquisition, competitive pressure, and what happens when growth slows start to surface.

In those moments, the difference between founders becomes easier to see. Some are using tools to help organize and accelerate their thinking.

I've seen many beautifully constructed pitch decks. Clear story, tight visuals, and everything looks how you’d expect.

What usually matters more, though, shows up a few minutes into the conversation, once we start talking about how the business actually works.

It’s for people who are willing to stay when things are unclear, help through hard problems, and support through the unglamorous middle.

If you need liquidity and control, venture is wrong.
If you’re comfortable trading those for participation and upside, VC earns its returns.

Venture asks you to give up access so founders can take risks, and those risks don’t fit neat schedules. In return, you get exposure to outcomes that don’t exist elsewhere.

If you want to enter and exit on your own terms, public markets do that very well. Venture is different.

I still hear people talk about venture like it’s “flexible” capital.
It isn’t.

When you invest, you’re committing to a long, illiquid timeline, often ten years or more, where progress is dictated by company building, not investor preference.

It was a real, hands-on reminder that small changes in how you see things can have an outsized impact on results.

I also got to meet Randy Pobst. He even took my car out for a few laps at the end of the day. Completely different level.

Once I started looking way down the track, like, way down, everything settled down. My laps felt more consistent, and I stopped overcooking the end of the fast straights. Same car, same conditions. Just better focus.

I spent the weekend at Thunderhill Raceway Park with my Chevrolet Corvette Z06. Fred Stout was my instructor. The biggest thing he worked on with me was where I was looking. I thought I was looking far enough ahead. I wasn’t. He could see it immediately and kept calling it out.

I spent the weekend at Thunderhill Raceway Park with my Chevrolet Corvette Z06. Fred Stout was my instructor. The biggest thing he worked on with me was where I was looking. I thought I was looking far enough ahead. I wasn’t. He could see it immediately and kept calling it out.

I spent the weekend at Thunderhill Raceway Park with my Chevrolet Corvette Z06. Fred Stout was my instructor. The biggest thing he worked on with me was where I was looking. I thought I was looking far enough ahead. I wasn’t. He could see it immediately and kept calling it out.

I spent the weekend at Thunderhill Raceway Park with my Chevrolet Corvette Z06.

Fred Stout was my instructor. The biggest thing he worked on with me was where I was looking. I thought I was looking far enough ahead. I wasn’t. He could see it immediately and kept calling it out.

The investors I'm paying attention to are asking the straightforward yet complex question: what gets unlocked when the people using AI never think about it at all?

That’s the part of this cycle that still feels underexplored to me, and it’s probably where we're headed next.

That transition will quietly redefine which problems feel natural for AI to touch, and which ones still don’t. It will also reshuffle which companies actually matter, because distribution and design will suddenly outweigh clever prompting tricks.

Over time, they completely changed who could use a computer, what people expected, and where it fit into most work.

I suspect AI is headed down a similar path. The next shift will be when AI stops feeling like something you prompt and starts feeling like something you use.

What usually follows that phase isn’t a breakthrough in capability. It’s a better interface.

When graphical user interfaces first showed up in computing, they didn’t make computers smarter overnight. At first, they were mostly just a friendlier way to do the same things.