It is debatable on the realms of technology and bandwifth realms on the question "is it solved?". Yes, in cities, stations and bases, I agree that the use and ease of terminals having easy access, physically or otherwise to the general neural net to be easy, and likely cheap due to such possible widespread adoption.
But the problem is between those points in space. True in stellar; as in within a solar system, trade lanes; stations; possible ships could act as relays. If allowed and not interrupted.
In the FL universe, we all know how simple it is to disrupt a lane, but does it stop with the 'power' used to hurl ships millions of miles? Does it also interrupt communications? If so, that could seriously be a problem. A theory I got is that there are small beacons, or transmitters either hidden, or so powerful they communicate with the gates, or even directly out of the system and to its destination directly. Slow, but more stable.
One problem is getting the information system to system. Even with gates, hurling objects was one problem, information could be another. A solution could be to 'piggyback' the information onto a travelling vessel heading to the information's destination, then relay it across the network on the other side. Problem, is it secure? Lane Hackers seem to have poisoned that method.
Solar information communications would be easier, and cheaper. For a true neural net in our sense, linking solar systems is a colossal problem in my mind. Keeping data secure, and fast to me is a near impossibility unless we cracked faster than light communications.
I'm unsure if its possible in Freelancer realistic wise, but in-game, we communicate to wherever we want instantly. But I'm doubtful its logical outside of the the in-game engine.
My extra 2 pennies.
EDIT: On the note of the 'brain-chip' style neural net. Planetary adoption is quite possible, quicker or immediate speeds.
In stations and solar systems to track people, trade information and such. Possible.
For intersteller connections. Unlikely, in my understanding of technology and some of its limits, it would be infeasible to 'link' intersteller networks in 'real-time' in what we understand neural nets to be. I'd perceive the data to come in pieces. Think of it as a YouTube video at dial-up speeds, you'd see the picture. Or pieces of it.