Dave Morton - Apr 13, 2014:

For example, most automobiles (specifically, passenger cars) made within the past 20 years all actually fit the dictionary definition of intelligent. In particular, the emission/engine control systems of these modern vehicles gather information from a legion of sensors, interpret the signals of each, and use these readings to formulate how much fuel to feed into the engine, how far to advance the timing, when to shift the transmission, perform self-diagnostics, and monitor and control the engine’s temperature. Data from many of these sensors is also tracked in the long term, allowing the system to adapt to driving habits, engine wear, climate and other factors, essentially “learning” as it goes. In these same vehicles you will also find Traction Control Systems, Anti-lock Braking Systems, Climate Control, and a host of other systems that can all be considered to meet the dictionary definition of intelligence.

Sentience
A sentient quality or state. (see Sentient. below)
Sentient
Responsive to or conscious of sense impressions <sentient beings>

Sharing your real life experience as a auto mechanic is a valuable contribution. Just to further work through the discovery process by pointing out the stimulus/response… That the human driver observation made of the stimulus of the “Check Engine Soon” light, leads to the human driver response to drive to the human mechanic soon,  where the human mechanic response is to plug in a portable computer machine under the dashboard, for the modern automobile machine response with the meaning of the “Check Engine Soon” light load device.

So, with this experiment, a conclusion may be made based on observation and testing that the modern automobile machine is responsive of its sense impressions, which is a sentient quality. This is also self-preservation, which is a goal of A.I.

We often speak of intelligence as binary. It is, or it is not.
But intelligence is often fuzzy, with a range:
Unintelligent <—- > Intelligent

When (or at what age) is a human considered intelligent?
Are animals intelligent?

Is Watson (or any other non-biologic system) intelligent?

When is a human Sentient? Is anything else that you can think of sentient?

Intelligence looks like an opposition to entropy.
An action exists to counteract the natural way things go. For example, when you keep standing up using your muscles instead of falling, or trying to get food instead of starving, or bending neck when the ceiling gets too low.
You get the picture…
So here is the algorithm!

Function intelligence():Action;
begin
  entropy:=perception_of_universe(now);
  repeat
  a_random_action:=Random(action); 
  entropy_simulation:=perception_of_universe(Random(action));
  until entropy_simulation<entropy
  result:=a_random_action;
end;

Et voilà!

Suppose for a moment, that “intelligent” is for all intents and purposes synonymous with “difficult”.
With regards to the AI effect, whenever we build a computer that handles a difficult task with ease, it is no longer difficult and therefore no longer considered intelligent. This happened to calculators, chess computers, Google and even Watson. Before them people regarded math, chess, data analysis and Jeopardy as requiring a lot of intelligence - from a human. The creation of an intelligent computer is then a paradox.

Intelligence is prehistoric, according to stone age artwork painted on cave walls.

The source of intelligence is will. The desire to do something. Desire is due to countraints, but is not hardwire defined. It emerges out of countraints against possibilities.
I’ve created many softwares that “simulate” some aspect of intelligence. The most “life like” one was this “pixel” moving randomly around a labyrinth with this simple countraint : “go where you have never been”.
Figure in an excell sheet a labyrinth where a “bug” moves around, trying to explore it. Then you have in a second sheet of the excel file a “ghost” of cellules already explored (each cellule is incremented by 1 at each visit from the bug, it represents the bug memory).
You see the bug moving around… just like a real one. It seems moving randomly in the first seconds, then you have the feeling he really gets what is around “him” and finally finds a way out of the labyrinth.
This simple program is fascinating because out of very simple rules comes a “creature” able to explore and understand his world.
On the other hand, I don’t see anything intelligent (beside programmer work!) in a chess game or face recognizing software.

My conclusion is that ‘‘RANDOM”  function is the most important one to create intelligence. The second one is memory. By memory, I mean everything that can be recorded : time, day, temperature, precedent movements, sounds, current movement… Then the random function picks up something “possible” out of countraints recorded in memory and countraints of the world.
Of course, you can’t teach chess to a computer with this kind of algorithm, just like you can’t teach chess to a 3 years old children. Teaching chess with “deep artificial intelligence” necessits to create in first order an artificial intelligence able to move around in a room or having fun with a ball… Then this intelligence will have the base to get further…

Moreover, looks how AI programmers are obsessed with the idea to create a “mind” able to answer “questions” : Jeopardy, chess, expert system.
Now look how creativity, meaning the ability to create questions is available in the real world… and you get how wrong we are, getting stuck in a way without solution.

Jean-François Battistini said, “I’ve created many softwares that “simulate” some aspect of intelligence. The most “life like” one was this “pixel” moving randomly around a labyrinth with this simple countraint : ‘go where you have never been’. “

Recently, in a theme park I found myself walking through, and unable to escape a labyrinth of tall bushes. Until I discovered a section which was compromised. The tall bushes had been knocked down, to short-change to escape.

So your labyrinth was itself another contraint you created in the software, where the pixel, unless godlike, may have never been, and may never go.  Otherwise, escaping the labyrinth would may have been easily short-changed by the pixel. 

What would have happened if, without cheating, you provided the pixel with a clue to find its way out of the labyrinth?  Please respond. I hope to discuss my virtual goldfish program on this thread.