Simulated Intelligence

Understanding Artificial Intelligence: Of Minds and Machines

The Teaching Company / The Great Courses / Wondrium (seriously this company has an identity crisis as they keep changing names) has released a brand new course: “Understanding Artificial Intelligence: Of Minds and Machines”.

“AI” the topic that just won’t go away!

This is actually a very good non-technical course and I highly recommend it.

Once again, I have to emphasize the name should be “Simulated Intelligence”. Using the term Artificial Intelligence is misleading.

Usually “AI” is talked about as if it was one thing. “AI” is a type of software tool, different tools are developed to do different things (and they have virtually nothing to do with each other).

Another misleading topic is the enormous resources used for “AI”. This is virtually all for training the AI. Actual usage (asking it a question and getting a result) is fairly minor.

President Trump and others talk about “winning” the “AI Race”. This is misleading. It’s like winning the race to build the fastest computer. Tomorrow someone will build yet a faster computer. Likewise with simulated intelligence, we have a future of many, many different simulated intelligence implementations being developed. The world isn’t going to end if China develops a better simulated intelligence tomorrow.

“General Artificial Intelligence” is the Holy Grail that many hope to develop. This would be a simulated intelligence which isn’t simulated, but actually able to adapt and solve any problem, like a human would. Our flexibility is our human superpower. My belief is that no such thing will be created in the next century (if ever), but I’m sure we will make a lot of very good simulations of general intelligence! Although even that might take a couple more decades.

Ray Kurzweil predicts a “singularity” (aka the Terminator movies coming to real life) in 2045. Now Ray, is a pretty intelligent guy but I have to disagree with him on this point.

It’s fun science fiction to dream of killer robots and “AI” wiping out humanity. But at least for now, my computer isn’t going to take over anything. If it starts misbehaving I’ll simply unplug it.

For a singularity to occur, this assumes that the rogue AI has access to unlimited energy, unlimited rapidly advancing technology (faster computer chips) and it implies some physical manifestation (robots) capable of doing more than just thinking but actually taking actions. As long as we are unable to cram an entire data center into a robot, it would still rely on a connection to the “master brain” (which could of course be easily disconnected / turned off).

Should we contemplate these problems? Yes, of course we should, but there is no rush (other than the fact that politicians are idiots and will never understand this stuff).

One big change that may be coming in a few years is “practical quantum computers”. A major leap forward in computing power will have a dramatic impact on Simulated Intelligence (and it will also break and render Bitcoin worthless).

For my articles, I frequently use Grok to create “stock art”. But here is a reminder of what human art still produces: