AI has evolved in three waves (Handcrafted knowledge, Stochastic knowledge, and Adaptive knowledge). Current AI faces technical challenges driving the need for new algorithmic approaches. Generative AI is just a “stylizer”; i.e., it can stylize responses in the format of “conversation” (or any other style, like “old English”). The future of AI is based on AI’s ability to abstract and operate on data it has never seen before. The future of AI will need massively parallel gAI chips, those that go beyond GPUs. The next wave of AI will be hugely disruptive: decreased storage needs, decreased data management, elimination of traditional training tools (like those at the hyperscalars), and decreased need for GPUs.
