I, for one, do not welcome our new AI overlords.
As a techie, I have loosely followed the capabilities of self-driving cars. Decades ago, I did some work in computer vision; but, at that time, the computers simply did not have the horsepower to do realtime scene analysis. Boy have times changed:
****Link seems to have died about six hours after I posted it.****
This realtime video was created by a neural net chip designed by Tesla. (Who knew that Tesla was into cutting edge chip design?) Apparently, the chip is needed because the car has eight cameras:
Tesla has now deployed a new unified camera network that handles all 8 cameras.
Processed resolution of 3 front cameras and back camera: 1280×960 (full camera resolution)
Processed resolution of pillar and repeater cameras: 640×480 (1/2×1/2 of camera’s true resolution)
all cameras: 3 color channels, 2 frames (2 frames also has very interesting implications) (was 640×416, 2 color channels, 1 frame, only main and narrow in V8)
Yet, all that processing power is delivering a really impoverished view of the world. Watch the incredibly dull video of what the computer actually "understands" about the scene:
So, from the technical POV, I am simultaneously impressed by the ability to steer in traffic; but I am relieved that the machine has a view of the world that barely rises above the level of "stick figure". It is not some Cylon robot out to kill us all - yet.
And that is where the title of this OP comes in. The number of Teslas that will be bought at $50,000 a pop or higher is a drop in the auto bucket. But, the uses of these vision systems for other purposes is a concern. We are already aware of the Digital Panopticon of surveillance cameras, with face recognition AI at a level comparable to the car autopilot. (Not to mention license plate readers and automatic highway toll cameras.) That is, recognition good enough to automatically preprocess a lot of info and deliver the results to a human for final judgement. In the case of surveillance, that judgment can ruin a person's life. To think that such "stick figure" reasoning is being rolled out across the board is alarming.
For a while now, I have been convinced that the end game of TPTB is to arrange for 90% of the population to kill each other (without starting a nuclear war) while they hide out in NZ, etc. The handwriting about overpopulation and ecological damage has been on the wall for decades, yet the Quiverfull movement and other childbirth extremists continue to be funded and publicized by rich assholes. The rich are quite aware that they have sent the old First World manufacturing jobs to China or even poorer, more draconian Fourth World shitholes. Now they are busy automating checkouts and designing shelving robots and Amazon warehouse robots.
None of this automation is so that 10 billion consumers will have their cheap Chinese crap delivered in a timely fashion - because those consumers will have no jobs and no money. No. This automation is putting in place an infrastructure that will allow the top 10% to survive without the bottom 90%. Its sorta like the Scottish enclosures, except they don't even need the displaced proletariat for brutal factory work. The checkout people and the truck drivers (see the autopilot above) will be truly superfluous people. And what does one do with superfluous people? The only solution that comes to mind is "the final solution".
We have already declared Moslems to be the new Untermenschen. Neo-nazi and crypto-nazi governments are in power in Eastern and Central Europe; and they would have no qualms about mass incarceration of "the other". The US is full of potential concentration camp guards who would love to abuse and murder "the other". We already have a huge, privatized prison guard population trained up.
The US has already murdered tens of millions in wars in the Third World. Bombed them, starved them (or both as in Yemen), poisoned them with chemical and radiological agents. That policy can continue in places the US military can control - like Africa and, soon, South America. The real heavy lift for TPTB will be to "normalize" the mass incarceration and murder of large populations in the first world.
The first step to mass murder is the demonization of the target population. The introduction of AI that destroys jobs and makes people "lazy moochers" is going to go a long way towards creating an atmosphere of desperation, scapegoating, and fighting over scraps. That atmosphere will be manipulated by TPTB toward the final solution. TPTB can play the two party game: the "left" party will demonize "deplorable" whites. The "right" party will demonize people of color, feminists, and intellectuals (just like the Nazis did). As long as the mass murder can be cloaked in domestic politics, no one is going to start slinging nukes at each other.
And, like the proverbial nail that loses a battle, the Tesla autopilot is the beginning of a chain of causality that will end up decimating the population of the planet.
I realize that the end of this OP is quite a long distance from car autopilots. But, anyone who still has critical thinking skills should understand where all this automation is leading. It is not going to help anyone but the professional class (who create the automation) and above. As a former member of that class, I just want to register my dissent from a society ruled by biased, black-box algorithms, a society owned by the likes of sociopaths like Jeff Bezos, amoral businessmen like the current management of Google, and Aspberger-y CIA pawns like Mark Zuckerberg.
I, for one, do not welcome our new AI overlords.