My thinking about AI has changed. I have long expected that sentience or self-awareness for artificial intelligences would come about when we could use a fundamentally new kind hardware. I theorized that as long as the basic computational model for machine language was binary, AI would not rise to the level of consciousness.
In fact, I thought that sentience would arise as an emergent phenomenon only when the hardware was complex enough to support something other than binary processing. That is to say that I did not expect sentience or self-awareness until we had achieved a new processing hardware model such as quantum computing.
I now think I underestimated the level of complexity we can achieve today. I am now beginning to suspect that the emergent phenomenon of sentience or self-awareness could arise in significant degrees (viewing sentience or self-awareness as a continuum rather than an all or nothing state) from simply applying enough data and enough processing power to existing technologies, without the need for advances such as quantum computing.
While I have readily dismissed the possibility of sentient AI in the absence of quantum computing or a paradigm-shifting processing equivalent, I now think we could almost be there. In a future post I will walk through my working definition of sentience/ self-awareness/ consciousness with respect to AI.
If you have any questions or would like to discuss generative AI, LLMs, prompt engineering, or similar topics, I invite you to contact me.