Will AI Do My Homework? (LLMs)
ChatGPT can write essays, but can it fix a broken robot? Why hands-on engineering is the most 'AI-proof' skill you can learn.
“Will AI Do My Homework?” (LLMs)
You type a prompt into ChatGPT. It writes your history essay. It writes your Python script. It even explains Quantum Mechanics. It feels like the end of school. Why learn anything if the AI knows everything?
But try this: “Hey ChatGPT, my robot is making a grinding noise and driving in circles. Fix it.” The AI will give you a list of possible reasons. “Check the gears.” “Check the wiring.” But it cannot walk over to your desk such up a screwdriver, and tighten the loose set screw on the output shaft.
The Limits of Software
We are entering an age where Digital Work is being commoditized. Writing code, writing marketing copy, and generating images are becoming cheap. But Physical Work—interacting with the unpredictable, messy, real world—is becoming more valuable.
- AI can generate the code for a PID loop.
- AI cannot tell you that the reason the PID loop isn’t working is because the encoder cable is plugged in backward.
- AI cannot feel that a bearing is gritty.
- AI cannot smell that a motor is burning.
Hallucinations in Hardware
In robotics, we use AI (Copilot/ChatGPT) to help us write code. But we have to audit it. Often, the AI “Hallucinates.” It calls a function that doesn’t exist.
- Software Bug: The app crashes. You restart it.
- Hardware Bug: The robot drives full speed into a wall and snaps a $200 custom part. The stakes in robotics are physical. You cannot rely on an AI that is “mostly correct.” You need to know the first principles yourself to catch its mistakes before they cause damage.
The “Human-in-the-Loop”
The future engineer isn’t a person who types code manually. The future engineer is a System Architect.
- You use AI to generate the boilerplate.
- You use your human judgment to integrate it into the physical machine.
- You use your hands to build the machine.
Robotics is the ultimate “Full Stack” capability. It combines the digital power of AI with the physical reality of hardware. As long as we live in a physical world, we will need humans who understand how to manipulate it.