At this stage, generic computing is unlikely to get any more significant boost (under the assumption that generic computing is something not necessarily [highly

At this stage, generic computing is unlikely to get any more significant boost (under the assumption that generic computing is something not necessarily [highly] parallelizable); instead, we'll likely see specialized processors. Therefore, ultimately, the market/industry will decide what to accelerate - and the technology to support the acceleration.