You can hold one of two opinions about the future of SOC design:
Moore’s law has run its course due to ceilings on operating frequency and power dissipation. The game is over.
The future for SOCs is in massively parallel-processing systems.
Which of these two futures you foresee depends, in part, on how much systems-development work you think there is left to do. If you think the world will continue to be operated by a pointer on a screen that is manipulated by a mouse or by cheap infrared remote controls that have a few dozen rubbery buttons, then there really isn’t much left to develop. In actuality, several really tough problems remain to be conquered by electronic systems. An even dozen of these challenges include: ...