How many channels do you need in order to learn like in the Matrix?

It's hard to say. I resonated with at least a couple of things that were said. That Neuralink's biggest challenge is clearly materials science and that a very distant future iteration of this device could be biological instead of digital. (Then again, given the scientific progress maybe in the future there won't be any difference.)

I predict one of the medium-term iterations of this device could be a sensor mesh covering as much of the brain as possible.

I do remain a bit skeptical about its safety for the average human.

Still, I believe Neuralink's efforts are worth it. And at the end of the day, even if nothing large scale comes out of it, at least it has helped thousands of people with various sorts of brain or spinal cord problems.