I’ve been thinking about a pattern that feels increasingly important as AI becomes more capable.
For a long time, progress was measured by how much intelligence a system could demonstrate. Better reasoning. Better outputs. Better answers.
But in practice, I’m not sure intelligence is where most of the user experience lives.
What I notice instead is the cost of accessing that intelligence.
Many people still spend significant mental energy figuring out how to ask the right question, structure the right workflow, provide the right context, or compensate for the system’s limitations. The intelligence exists, but reaching it requires work.
That creates an interesting inversion.
Instead of tools adapting to human behavior, humans adapt to the tool. They learn interaction rituals. They memorize patterns. They optimize themselves around the system.
The most effective designs seem to move in the opposite direction.
Their achievement is not making complexity disappear. The complexity is still there. It has simply been absorbed into the product itself.
Users no longer need to think about orchestration, prompting strategy, or process design because those burdens have been internalized by the system.
I increasingly see interaction design as the real frontier.
Not because intelligence has stopped improving, but because intelligence becomes far more valuable when people no longer have to spend cognitive effort accessing it.
The future may belong less to systems that demand expertise and more to systems that quietly eliminate the need for it.