Back in the 1980s, we imagined that the future would be big arrays of screens enmeshed with dense bundles of wires; that technology would take over the world. In reality, as technology has taken over the world it has become increasingly invisible. The beautiful aspiration is not to be surrounded by machines, but for technology to disappear into nature.
Looking at a laptop, physically it is dominated by a screen, a keyboard and a trackpad. Modern science is already pretty comfortable extracting keyboard and mouse information directly from the brain; many groups have done it in many species. It is slow, finicky and requires bulky implants today, but I am confident that it will become fast, wireless, and miniaturized very soon.
Similarly, multiple groups have already shown the ability to inject visual information into the brain. The science here is admittedly less developed than that of decoding motor activity, but it is a much higher dimensional problem and we have been limited by tools, which is changing rapidly. In my mind, one of the most striking statements of the feasibility of a visual prosthetic is a paper from 1999 showing fairly believable reconstruction of vision from the thalamus of cats using only a handful of electrodes. There is certainly science to do here, but it will get done.
Once you can extract motor information and write visual information, all you need is a small block of compute with a battery and radios in your pocket. The computer effectively vanishes.
This is the last screen: the ultimate user interface that no one is designing for right now. When my eyes are open I should see the physical world with the possibility of overlaid information, and when my eyes are closed I should see the informational world, which can use completely different metaphors and interactions. When I want to rest it should be easy to turn it all off, but I also think our children will look at us baffled when we tell them that there used to be just nothing there when we closed our eyes.
Present phones are built around their screens and cameras, and so the demands on the neural interface support pocket-compute-block will be quite different. It almost certainly make more sense to start over rather than try and have the implants pair to phones as they exist today. That blank page is also an opportunity to do better on security and reliability, which are closely related. It is an opportunity to make sure that these platforms are truly personal and private and work for us, and that we are not the product.
I think there has been a sense over the last few years that personal computer revolution is over, that the progress is smartphones is approaching an asymptote and that the internet is pretty well understood now. Nothing could be further from the truth. As these technologies mature and reach increasingly broad audiences they are going to make the 90s look small.