Why great AI products are all about the data

Product management is still such a relatively undeveloped discipline. We're like 15 to 20 years into this, and so there's something about the current state of product management that isn't getting at the truly important things, the truly value-added things.

There is no natural force that pulls companies towards thinking about the end user's enjoyment and success early in their journey. If there is nobody in the organization whose true incentive is to measure their end user success, their enjoyment, their happiness, their retention, their engagement early on, it will not happen.

LLMs can only be as good as the data they are given and how recent that data is. They're ultimately like information shredders. They are limitless information eaters. You can never have enough information to give to an LLM to truly gain its value.

If you look at a piece of data and the result tells you something that your intuition tells you is insanely wrong, like they probably not right. First, believe your intuition and go and prove yourself right.

In everything always talk from the customer's perspective, from the market's perspective, from the competitor's perspective, the very small number of PMs do that. They get dragged into internal politics, they get dragged into scrum management or scrum execution or product delivery, and you just can't win that way.

I have to think of my career as a little bit like a bingo card. I've always been looking to fill in boxes I didn't have filled because I felt like that would make me a better professional.

I've rarely regretted going deep in something that isn't quite my job. The worst case scenario is I've learned something new that I will never use. But the very best case scenario is that when I least suspect it at some point in the future it will turn out to be the thing that matters.

People really underestimate where the value is created in these applications and they just kind of get it completely wrong. It's not the UI that matters and it's not the data model that matters, although those are both very useful. It's the years and years and years of evolution of the underlying workflows of the product to support the customers.