Technology is going too far

I'd rather die than live in a world where everybody can read my mind. But that's just me...old school. I'm not going to criticize anybody who disagrees. 

Paul Muad’Dib Butlerian 'crusade' time! (funny how they changed the original term for the new film - but not surprising)

Improved artificial intelligence and now these things. Terminators are not too far off now. Time to break out the heavy munitions and directed energy weapons.

It reminds me of the line from the kid in one of the terminator movies when he realized technology has come too far...
"We're not going to make it are we?"
Actually this thread was inspired by car shopping. I hate all these luxuries that they call neccessities. Oh well, my wife always drives the newer car and she likes most of that stuff. I just hate it when computers tell me what to do as if they are my boss and not vise versa. 

That is more or less my position on this. I am personally in favor of transhumanism. I think integrating with our technology, our tools, will be what ultimately saves us from our short sighted and self destructive natures. We still don't know the limits of our neuroplasticity, but all signs currently point to the likelihood of ready adoption and internalization of extra-corporeal sensory devices and direct neural/machine interfaces.

I'm a fan of most of Masamune Shirow's manga work, and the various translations of it to other forms of media (e.g. anime). GITS is a particular favorite of mine, and it definitely plays into my tendencies toward dark cyberpunk and military tech fetishism. SAC (and 2nd GIG) are particular favorites of mine.

What model are you looking at? Most of they keyless/push-to-start fobs are grossly overpriced by dealers. You can often come by an aftrrmarket fob and they'd "only" charge to program it, or you might be lucky enough to have aftermarket fob and programmer modules available.

Completely random example:
www.amazon.com/Dorman-99389-Keyless-Transmitter-Select/dp/B07Y2LGGPS/ref=asc_df_B07Y2LGGPS/

I use an OBD2 reader for my old and quite used car for diagnostics. I noticed a warning light and some worrying symptoms. I did my diagnostic, then took it to a shop. They said it was "probably" a bad spark plug but they would have to replace all of them to be sure, and may replace some other things, too. I acted surprised, then said, "Huh... the computer says it is registering misfiring on cylinder six. Does your equipment not get that level of detailed information?"

Long story short, I had a different shop replace all the plugs (cheaper to do them all at once anyway, I knew that) but these guys were honest and asked an honest price, unlike the first crew. That little reader has saved me a lot of money.

I don't know what Tesla's development cycles looked like, but training AI for self-driving vehicles inside of games has been an academic challenge/personal hobby for many people for years:

alzaibkarovalia.medium.com/building-a-self-driving-vehicle-in-gta-v-using-deep-learning-and-convolutional-neural-network-696b38b4c81e

Of course, so far we are deploying the real-world things in more or less ideal conditions. Put them on the busy streets of Rome or any congested city in India and it may become so decision-locked it defaults to not moving at all.

Of course, there actually are cases of self-driving vehicles crashing in the real world. Some involving fatalities. In almost every case I read, the driver was either sleeping or otherwise not fit to provide the emergency override input when needed. Civilian available self-driving cars aren't meant to be FULLY autonomous, the human is expected to be observant and intervene as needed.

The latter raises two points (I won't call these well-formed hypotheses) about automated vehicles in my mind.

1) In theory, the more self-driving vehicles on the road the safer it will be as there will be less human unpredictability in the system. Indeed, if they all connect to a centralized computer to share their intended destinations some traffic could be rerouted so everyone arrives where they want to be sooner than if humans created a traffic jam. Likewise, they could coordinate better to ensure clear passage for Police, Fire, and EMS response, saving lives.

2) There are always edge-cases. A ball bounces into a street, something falls off the back of a truck, a sensor goes out and the AI no longer has sufficient data to navigate safely. In these cases, having an attentive human on standby is crucial. However, the closer we get to full adoption of an organized system as described in point 1, the less time people will log actually controlling a vehicle. Wouldn't we then be falling back to inexperience drivers in an emergency? They may over-react or, more likely, fail to react in time from surprise. Taking control of the wheel or hitting the breaks may feel so alien they may not be first reflex.

3) There will be non-zero-sum scenarios. You will have situations where an accident is inevitable, and the AI units must make decisions. Say you have an AI car, followed by a human on a motorcycle, followed by yourself in an AI van with your family including young kids. In opposing traffic are several AI cars and vans, on the other side is the cliff on this mountain road. Some sudden obstruction occurs in front of the car in your lane. Because of how the human has been erratically handling the motorcycle, you cannot stop in your AI van lane without hitting him with great force, with near certainty of causing his death and minor risk to your family. In the other lane there happens to be another AI van with a full family. Both AI vans could break, with a planned head-on collision that has a high probability of injuring all parties but a much lower possibility of fatalities, thusnsaving the motocyclist. The only other option is to go over the cliff's edge, with a high probability of fatalities in your family but still a possibility of injured survivors. Does the AI prioritize life over death, causing the two vans to collide and potentially causeb life-long injuries to two entire families? Does minimizing casualties take priority and the van plow under the human motorcyclist to remove the other van from the equation? Does the AI try to minimize AI caused injuries/fatalities against humans operatingnnon-AI vehicles and have your family take it's chances over the cliff?

These sorts of things are open questions in AI research. Of course, AI does not work like a human mind and cannot make moral judgements. They weight probabilities and compare them against their goals. HOW those goals are defined by the programmers can have tremendous and, often, unintended consequences.

Here's an old video about problems with (currently only theoretical) Generalized AI:

I highly recommend Isaac Asimov's Robot Series, to include most especially the Foundation series (they exist in the same universe, over tens of millennia of human history). What his version of "AI" does while interacting with humanity over that time is fascinating, if entirely out of line with how contemporary AI actually works.