It's a good answer, not just clever. The issue he's addressing, clarified, is, "What if we had a source of infallible (moral) knowledge?"

Then, "ask that source, not me, you, or any other fallible human" is not just clever but correct.

However, infallible sources of knowledge badly violate everything we know about epistemology. So don't expect to actually find one. What interests me more is this issue:

If I had a magnet that made people act a certain way, and I thought that certain way was moral/good, then should I use it on people? The answer then is no, in the event that I disagree with someone about what's good, I shouldn't force them with magnets; it could be me who is mistaken not them. I should instead persuade them to use the magnet voluntarily.

What's good about a discussion with attempted persuasion is it could go either way. I could end up persuaded, or him. This is different than using the magnet which can't go either way, so if it's a mistake there's no possibility of it being corrected.

The question "What if there were a machine to make everyone moral?" implicitly assumed it would make people moral, whatever that is, as a premise, rather than moral according to someone's judgment. This reading may be debatable, but I think that's how he took it.