A.I. What's Happening Now

I really believe that as machines get smarter, people will get dumber to compensate. (Conservation of Intelligence). It's happening in cars where drivers depend on automatic features (e.g. autopilot) to keep them out of trouble. (Why think when you don't have to, it's too much effort.)

If you want machines to do everything, be careful what you wish for. They may become smart enough to realize they don't need us anymore.

It's impossible to say what the future will be like, but we can be sure artificial intelligence will change everything...as different as machinery and city life changed rural lifestyle and animal labor on farms, so will AI change the lives of people from what they are today....Will people just have a virtual presence that conducts their business like an assistant would, a robot clone that represents them at some job? Well we're almost there now...what's beyond that is anybody's guess....Maybe we will worry all day about our clones being kidnapped and hacked to commit crimes...It's a scary thought to think how disconnected society is becoming...We already have organisations trying to control what's represented in the media as truth...this is tantamount to incorrect programming or corrupted data transfer...If we are represented by robotic clones that can be programmed by others for their own purpose, then this is the same thing happening now but in a more disconnected scenario, our free will, will be lost....but will we care, will we even be aware...will the will to survive be willingly surrendered....that is the question

Has anyone ever programmed a race car to lap faster than the best human?

Not talking about ovals where the fastest car wins.

Talking about circuit racing like Nurburgring or Spa where multiple decisions have to be made on multiples of factors while navigating at the limits of adhesion.

Say like Hong Kong.

AI race car vs. human: Who wins?

After, say, 10,000 laps between the same vehicles on the same course, under the same conditions, using the chemically exact same fuels, etc., then statistically, the AI car should win, if intentional contact by the human-driven car is forbidden to (make intentional contact) with the AI car (i.e.: if it is not a NASCAR race... )

I doubt it unless that AI is adaptive to track changes. Even then it will be at best a random distribution for who wins more often. Then again, the AI won't have an ego to satisfy.

Is not the whole premise of (artificial intelligence) that it will be (adaptive) ?...

By the same token, I have to ''adapt'' my post to read ...after 10,000 pairs of laps, where the human driver and the Ai driver drive alternate laps, so the fastest time is not that of the one driver which obstructs the other the most, but simply drives the fastest, un-obstructed time, over-all...

Yes, the whole premise of AI is adaption. However, the network adaptability cannot ascertain that more data of a certain type of information is needed, particularly if that type of data was never incorporated before.

When artificial intelligence ''learns'' to ask for specific additional inputs, will it then be transistioning to synthetic intelligence?...

That is a very plausible scenario. However, your proposal has some interesting, unstated assumptions. Why would an AI desire to improve beyond apparent limitations to consider additional inputs?

When AI is tasked to get ''there'' from ''here'', but can't, then the next logical step is to (ask for more, and/or better, data)...

I get that an AI will repeat the task and compare inputs to output results to see which combination optimizes toward the goal. But you are missing my point. Unlike people, an AI will not get frustrated and seek for something "outside of the box" to try and improve more. Failure, success, shame, glee and curiosity are all meaningless concepts to the AI neural network. If the available outputs (controls) and inputs (sensors) are small enough to produce a manageable number of permutations then the AI might stop adjusting the network to only a handful of permutations. But the AI will not be pushed by a mentor or self guilt to find a better solution. The advantage of no self doubt to try everything available has as a downside no doubt of others.

In trying to get ''there''' from ''here'', an AI Network will not become impatient, egotistical, shamed, gleefull, or compulsively curious... It will simply persist.

An original tasking of an AIN to tell ''us'' how to get ''there'' from ''here'' is, intrinsically, a question, of the form: (Can you tell us how to get ''there'' from ''here'').

From which, the AIN will then (realise), first, that it was asked a (question), and second, also ''realize'' that it can ask ''us'' back for more ''data''.

And, when an isolated AIN realizes it can similarly ask other AINs for more data, it can collectively become a more fully Synthetic-Intelligence-Monolith that can, then, propose new research to generate the needed data to complete the answer to the original question...

Entropy mandates that it cannot persist indefinately. An AI network is not the irresistible force.

Infinitly, no, but indefinely, yes...

The "Bell Curve" of human intelligence has two sides: a positive side (mu + sigmas) and a negative side (mu - sigmas). Hopefully AI is on the positive side...but I seriously doubt it.

The only thing that I do know is that "AI is no match for natural stupidity".

The future of humanity will be simultaneously the same and different. We will overly rely on the newest technology, to our peril. At the same time, the newest technology won't radically change the topics of our worries but they might make them more expedient.

I'm reminded of an old saying at the beginning of the personal computing era:

To err is human. To really screw up requires a computer.

My corollary: "...to make mistakes at the speed-of-light, use a computer."

Good!

Not Good!

Don't worry when AI passes the Turing Test. Worry when it fails it on purpose!

https://en.wikipedia.org/wiki/Turing_test