A question with a similar title was asked here a few days ago, but they were actually asking for advice on side project ideas. But it got me wondering about the general question.<p>On the hardware side, I've heard a lot of conflicting predictions about optical computing and optical interconnects, both with a CPU and between the CPU and memory. I'm skeptical about the former. The latter has been 'just around the corner' for years now with no sign of becoming mainstream.<p>But what else?
I was very weakly involved in the optical computing area and watched not much happen over 40 yrs. My observations were existing CPUs just improved enough to render many of the optical computing promises unnecessary. So, I have always had skepticism many developing technologies that are over promoted. I think LLMs are great but are over promoted. I suspect there will be a lot of idle capacity that will serve as a platform for new projects. I'm thinking it may be worth looking into what will fill those voids: the excess capacity, the lowering of costs, the unemployed AI people looking for new projects.
Not my field, but I think DNA sequencing and analysis of new drugs on these platforms may in an area to consider.