December 23, 2023

Unknown thousands of new companies start up continuously on complex technology which have no relationship to objective reality. The managers study business administration and economics and are clueless on the science and technology. So they believe anything is possible with technology. They compare the evolution of a light switch to the evolution of renewable energy, electric vehicles and self-driving.

Self-driving autos cannot properly evaluate flashing lights, because rapid change cannot be suitably evaluated by computers. To evaluate while moving slow would require about 20 frames per second by a camera. Moving rapidly, about 50 frames per second. During each 20 milliseconds, everything in a frame must be known, evaluated and compared to the previous 20 milliseconds, while it is all in a different position. Is a light the same one in the previous frame when it is in a different location?

Computers cannot determine what the same spot is when it is in a different location than before and environments change. Humans can determine what the same spot is while in motion, computers cannot.

To evaluate a blinking light which is on for 25 frames and off for 25 frames would require everything in the computer to be compared for differences 25 times for the light on and 25 times for the light off. That's terabytes evaluated and compared through a large number of algorithms every 20 milliseconds with no indication of why each frame is different, while light intensity, reflections, colors and borders change in each frame.

To evaluate a flashing light would require zoom as close-up focus for the camera. Close-up reduces the field of view. To observe a light differentially would reduce the field of view to a few feet. To observe the rest of the surroundings would require more cameras for every few feet. A vehicle would be nothing but a pile of cameras to do that, with each camera requiring terabytes of information being compared through numerous algorithms every 20 milliseconds.

But self-driving is supposed to be superior to normal vehicles and look the same, which means the sensors are stripped to almost nothing while doing the impossible. Why don't the engineers tell the administrators that it isn't possible? Because they don't want to lose their jobs. Try telling know-nothings that something is impossible and they will get someone else to do it.

Not the least reason is that the deciders are more interested in impressing the public than whether something actually works. They buy and sell on impressions, not results. And they assume the problems will be solved through future wizardry, which they rationalize with the slogan, "fake it until you make it."

Now they are investing in a new form of so-called artificial intelligence which they pretend will out-perform humans. They want to be the first and most advanced in developing the technology. All that type of technology can do is go through a complex data set and produce a quasi-average of what they find.

There are only two types of data sets being scanned: One is literature which has been compiled; and the other is images that are stored as data. Almost everything else that humans do is not compiled as a data set available to computers. How could anyone miss that elephant in the bathtub? It gets ignored in the fantasizing and futurism of the fakery.