I go back and forth on this myself. To some extent, I expect it involves the whole 'thinking fast and slow' set of tradeoffs.

However, there are states of mind I've both read about and experienced, where it seems like you can fit a much more complete, rapid and specific understanding of a complex situation's interconnections in thought than you can get across in the bandwidth of speech. This is part of why we invented slide shows, haha. But in all seriousness, one of the more interesting outcomes here could be the ability to 'project' a complex thought/feeling experience in a way that speeds up the propagation of knowledge & understanding.

Something like this could revolutionize everything from education to mental health. It's super uncertain if we'll ever achieve that, but we could learn some useful things on the way. For me, the biggest questions are, can we significantly speed up the learning process, what types of learning work best or worst, and how can we use this to better understand the brain?

If you haven't seen it yet, here [https://waitbutwhy.com/2017/04/neuralink.html] is an extremely long, cartoon-filled, fanboy description of Neuralink's potential which somewhat discusses these ideas and a lot more.

Language is abstract of brain function. Neurons fire electrical signals in milliseconds some cases even faster, words come second as a result of the firing. I'd think that neuralink is trying to figure out right now what fires where and at what time in the brain to effectively capture and transfer it into bits imho.

That's not the only goal though, the more immediate goal is to restore mobility/communication/freedom for disabled people. Neuralink can be a huge success even if it doesn't bring about a cyberpunk future (which, honestly, do we even want that?)

I think tapping the brain directly could theoretically enable us not needing to "collect our thoughts" and work with unsupervised data.

I predict, the biggest obstacle will remain not having even the most basic understanding of the brain, with reductionist tech optimism remaining as prevalent. Ultimately I fear, understanding the brain will be as achievable as predicting the weather in silico. Complex chaotic, or evolutionary principles give birth to emergent states and we are stuck with a leaf's understanding of a tree, while spawning little complex blackboxes of our own creation, celebrated for unbeknownst slowly eating away our existential foundation (see The Economy).

I like the Kahneman "Thinking, Fast and Slow" systems concept and I would extend the metaphor using CPU (slow) and GPU (fast). I would agree that the brain is inherently slow, using the CPU metaphor it has an embarrassingly low clock rate. But each tick is supported by brilliant hardware optimization (GPU) so the data throughput and processing per tick (and watt) is remarkable. So yes I agree it is slow, but then the art of a BCI will be interpreting the state changes between these slow ticks that actually represent tremendous amounts of data.

It appears that internally its transmission rate is faster.

For example, try driving a car, vs having two people collaborate to drive a car, with one on the steering and the other on the brakes.

The two person setup is theoretically at an advantage because you have double the sensory input and double the 'brain power', yet the results are far worse.

This suggests speech is a limiting factor for performance for the latter setup.

> What if a brain is just inherently slow when transmitting?

I don't get that impression. That's the entire counterpoint. Maybe others can elaborate.

Language is the slow bus.