Draw the Route

The Educator’s Role

As an educator, I keep asking myself what value I provide when a personalized course can be generated on demand. AI can build a curriculum in seconds. So what does the student still need from me?

My last post explored how AI is shifting the engineer’s work from writing code to specifying intent. Something similar is happening in education, but the relationship is different. On the one hand, you’re working with a machine. On the other, an emotional being.

What Stays Hard

What remains hard is intent. Someone has to know where they’re going before any of it can be written down. AI can execute on direction, but it cannot provide it.

The educator’s edge is student sympathy: genuine concern for the learner’s struggle to find a path through too much information. That means curating content into something absorbable, ordering concepts so effort lands where it matters, and catching drift before it compounds. That last part has grown more urgent. Students reach for AI first, and AI tends to agree. A flawed understanding gets confirmed, not corrected, and the student moves forward more confidently in the wrong direction.

A student who can’t supply their own direction will get fast output pointed at the wrong thing. That gap is where the educator’s value lies.

The Curator’s Route

○       ○                   ○
                       ○
· · · · ○ · ·                 ○          
  ○          · ○        ○
               · · · ○              ○
    ○                · · · · · ○
         ○     ○          ○    · ○ -> X

Imagine a sandbox viewed from above. Rocks are scattered across the surface, dozens of them, spread unevenly. A single line traces a path through the sand, winding from rock to rock, deliberate and unhurried. Not the shortest path. Not random. A chosen one.

That line is curation.

When anyone can learn anything, the question of what to learn and in what order is more important, not less. A learner dropped into an ocean of available knowledge doesn’t need more water. They need a route to shore.

Ask two educators to draw that route for the same learner, and you’ll get two different lines because no two people carry the same experience. AI-generated curricula tend to converge on the same generic path, which looks like reliability but is closer to a lack of perspective. The educator who zigs where another zags is drawing on experience the model doesn’t have.

The Speed Limit

Generating a course doesn’t mean anyone will learn from it, or that it’s even digestible. Knowledge has to be worked through slowly: tested, questioned, applied badly, then applied better. That process has a speed limit, and it belongs to the learner. Spaced repetition still works. Retrieval practice still works. These are not inefficiencies to be optimized away. They are the mechanism of learning itself.

Learning takes time no matter what, so the path through it matters. A good route doesn’t beat the speed limit. It just means none of that time is wasted on the wrong things.

The risk isn’t that AI makes learning worse. The risk is that the feeling of learning gets decoupled from actual understanding. Watching a well-explained AI response can feel like learning the same way reading a recipe feels like cooking. It’s a start, but at some point you have to make the food.

The reason to still learn isn’t to compete with AI. It’s to direct it. AI operates on the context a learner supplies, and without enough of their own, they can’t specify the problem clearly, evaluate the output, or catch it when it drifts confidently in the wrong direction. The learner who skips the work doesn’t just fall behind. They lose the ability to steer. A learner who can’t evaluate AI output is dependent on it, not empowered by it.

Conclusion

The educator is a curator, in the business of deciding what matters, what comes first, and what to leave out entirely. That is taste. I don’t generate taste. I build it by making those decisions over and over, getting some wrong, and learning to tell the difference.

So what does the student still need from me? I am the person who walked the path and remembers where the dead ends were. I looked at forty rocks in the sand and chose seven. That choice is my value. Not because the machine can’t draw connections, but because it doesn’t know which ones matter. Learning still has a speed limit, and every wrong turn is time the learner doesn’t get back. The route I draw for others matters more now than it ever did.