Where this shows up

The see → model → predict → plan stack isn't only for robotaxis in a few test cities. Pieces of it are already in millions of cars on the road today, and the same ideas are spreading well beyond passenger vehicles.

Here's where autonomous driving tech actually shows up.

The same idea, many jobs

Each is a different point on the autonomy spectrum, built from the same parts:

• ADAS (in cars you can buy now) runs the perception and planning stack for one task at a time — staying in lane, keeping distance, braking for danger.
• Robotaxis run the full loop, from the bird's-eye world model to predicting other agents.
• Trucks, delivery bots and warehouse robots reuse the same see-and-plan pipeline in easier, more constrained settings — which is often where autonomy pays off first.
• All of them improve through the data engine: more miles, more rare cases, better models.

Still supervised

Most of what ships today is driver assistance, not hands-off autonomy — the human is still responsible. Understanding how the stack sees and decides is exactly what tells you where to trust it and where to stay alert.

That's the course

You've followed the whole pipeline: cameras, perception, a bird's-eye model, predicting others, planning, and learning from the fleet.

If you enjoyed this, the other courses go deeper on the computer vision and neural networks that make a car able to see.