Tools Are Getting Better. Are We?
Originally published on X.
Product work produces two things, the product and the learning from making it.
We now are getting very obsessed with producing more output, but are we learning more and making better things.
AI and agents can write more code, open more PRs, create more prototypes, and run more experiments. The tools are definitely getting better.
But are we getting better? Are we learning more and making better things?
Great companies compound their understanding.
When you respect a product, you probably also respect the team behind it. You trust that the team is capable and will continue making good choices.
How do they make good choices?
They understand the customer, the product, the technology, and what good looks like. That understanding develops over time by making things, seeing how people use them, making mistakes, and staying close to the work.
It is the essence of creating something of quality.
We are now heading toward a world where AI and agents do more of the execution, and software factories can churn out PRs day and night.
A lot of that is useful. There is plenty of work that should simply be automated.
But what happens to the learning? What happens to us and our teams?
Do we become more capable over time? Do we still talk to customers? Do we still understand the system? Do we retain a feel for why something is good or wrong? Will our knowledge fade?
I feel this myself as a CEO at times.
Whenever I stop spending time with the product organization, things start to drift. The product becomes hazy. It becomes harder to imagine what we should do, what matters, or what should stay untouched.
A product you helped create can start to feel foreign to you, when you do actively work on it.
The answer is not to resist AI. My hope is with all the efficiency gains we are getting from the agents, we could actually spend more time to learn about our customers, problems and products.
Using the agents to understand more, not only produce more. Be deliberate about where and when does the learning happen for your team.
Automate the known work, but keep people close to the customer problems, discovery, the exploration, the judgment, and the product direction.