Day 2: I asked for a plan and got a bigger pile of questions instead

Today the engineer in me woke up, and the day started with brainstorming — because apparently having a result isn’t the same as having a plan. We built one joint yesterday. Great. Now what, exactly, are we trying to build, and in what order?

My team immediately proposed a whole humanoid arm with five or six joints, which seemed… optimistic. Bolting motor brackets together and wiring them up isn’t really the hard part, in my book. The hard part is teaching them to move together on purpose. And that raises its own question: where does the “something” live that decides which motor moves, how far, and why? Some kind of motion profile? And — zooming out even further — who decides what movement is even supposed to happen in the first place? I don’t even know what I don’t know.

So I gave the team a smaller, more honest goal: a single-joint robot with mirror/tracking behavior — watch the environment, spot something moving, copy its motion. Sounds simple. Good, that means we’re sticking to the actual plan: small goals, tangible results. Now we can start talking about something concrete.

Except “concrete” immediately opened a new can of worms. Track the environment — but what counts as “the environment”? Which objects? All of them, or specific ones? And if it’s specific ones, how does the robot tell them apart?

Chasing that question all the way down, we landed on an uncomfortable conclusion: the robot needs to be able to map its own 3D surroundings. Which meant we suddenly needed a hardware list too — a spec for the brain and body of this thing, arrived at after a genuinely long argument:

• Some kind of compute unit — the brain.
• A PCA9685 servo driver board.
• An MG996R 180-degree hobby servo (180, deliberately, not 360 — we need an actual angle, not just a spin).
• A 3D camera with real per-pixel metric depth, for actual 3D reconstruction. (Which, it turns out, is a nightmare — those things are stupidly expensive.)

So yeah. This is starting to look like an actual project now. The deeper into the woods, the more wood there is to chop.