We're running out of low hanging fruit for advances in speed (and going to hit hard physics limits soon after that), and things like quantum computers are only good for a small subset of tasks. If it does eventually happen, it's going to take some pretty massive computing clusters to do it. And then who's going to pay billions of dollars in invetments and millions of dollars in yearly maintenance for a single human-equivilant mind?

In reality (and assuming all of this breathless hyperbole about what AI actually is doesn't set it back another 20 years), what I'd expect to see is machine learning become more common everyday consumer areas. Things such as a system where you put a microphone under a cars hood and feed it to a computing cluster, then 10 minutes later it spits out a list of things likely wrong with the engine for the techs to look into.

I've never run out, but I've come quite close a few times.

One particular time I was driving through the desert, passed a town on a low (was reading about a quarter) tank. I realized about 30 miles later that the next place I could stop for gas was still 60 miles ahead. I put 13.8 gallons into a manufacturer claimed 12.6 gallon tank at the next stop, so I'm sure I was close to empty.

I had another one where I left the house on a quarter tank with ~100 mile drive up to a friends house planned. That should have been enough to easily make it to the convenient station just down the street from him, but I hadn't realized how much more mileage making a detour to pick someone up on the way added. Noticed the needle buried in the red about 3/4ths of the way there, took the next exit, filled up, and was back on the freeway again within 5 minutes.

Then there are the countless other times I've driven to work (yay 60 mile commute), then stopped for gas before leaving for home since I knew I'd be risking it to try to make it back on that low of a tank.

And the heater is a problem when idling in traffic in the snow. If you run the battery down maintaining a comfortable temperature, it's not like AAA can bring you a little can of electricity to make it to the next charging station where you can fill up and be on your way in a few minutes.

Not only are gas stations more prevalent, but there are also a huge number of other options available if you run a gas powered car out of gas ranging from a passerby letting you siphon some out of their tank to someone driving a can over to you to worst case of having to hike a few miles to the nearest gas station. Run a battery powered car out of battery and your only option is to be towed.

And the gasoline equivalent would be to walk up and ask for a little can of gas. Which is easily carried by a single person and can get you another 50 miles no problem.

Yes but it's one more layer to defeat. It might also keep the casual maliciousness out (say the neighbors kid just read about this cool thing you could do the hack a car). Either way, it's a really simple step with no downsides.

It would be an extra device, just like a hardware firewall. My $30 desktop switch has enough brains to let me configure it to block some basic stuff (like MAC flooding) plus act as a switch. I'm sure it cost a 10th of that before all the retail markups. The cost argument is why I'm fine with the lawsuit. That will give the people who want to design with an eye for security a cost savings argument ($300 worth of chips is potentially cheaper than a massive lawsuit).

Flammables are easy to make. Detonators that are consistent and reliable are hard. I'd be surprised if you could come up with a design that's less work than ordering a $20 device (one that no one would even think twice about you buying, unlike say, thinly shaved magnesium strips) and downloading a program.

Again, because of the controllability and the way it lends itself to casual use. Also, this is exactly the kind of argument people made against securing computers in the 90s, and we're just now getting away from those consequences today. Why not spend a tiny bit of extra effort designing this stuff with a least rudimentary security in mind?

That's the point when you take the source code (that's the whole point of open source right?) and make a new project not managed by an idiot.

That's news to me then. My impression from watching a video a while back of how these worked was that they were simply using the OBDII port to send false signals and/or flooding the bus with so much traffic that the signals couldn't get through. I could have sworn they specifically said that the dash was only apart because they'd been monitoring signals while developing the hacks and couldn't be bothered to put it back together again.

Throw an interface in between it and the rest of the car that will do the following:
- Disallow any communication that wouldn't be expected to come from that port. e.g. I would not expect the ABS sensors to talk to the brakes via the OBDII port
- Limit the number of messages sent into the port if it exceeds some threshold (assuming that the attacks relied on spamming messages. But even if not, is there some reason you'd have to flood the bus with messages?)
- Possibly put it somewhere where casual physical access isn't as easy, e.g. inside the glovebox that locks when the doors are also locked.
- Add an indicator to the dash lights to say when something is plugged in and/or diagnostics are being run.

Note that this doesn't do any of the following:
- Hinder your ability to use diagnostic tools to read the values reported by the ECU
- Hinder your ability to access the port
- Cause additional processing on the part of systems other than the firewall between the port and the rest of the network

Because as I said before, if you use a hacksaw you have no control over when or where the accident happens. It becomes a higher threat electronically because of controllability, variety, and ease of use.

Hell, you keep mentioning hacking and modifying your own car... do you drive with a bluetooth OBDII interface plugged in so you can log data on your phone? (I did for a while when trying to troubleshoot an airflow issue) Would you still do so knowing that someone could leave a transmitter somewhere near the road just waiting for someone like you to drive by so it could send your car off into a tree for the lulz?

That's basically what I meant by "vacuum seal". Something like those space bags where you remove most of the air to compress blankets or whatever but aren't necessarily creating an industrial vacuum of just a few torr.

This is what I was going to suggest. Portable DVD player with a few different DVDs of different brands (in case one uses some kind of corrosive label or something) and vacuum seal it all with some moisture absorbing packs. Take the battery out of the thing and make sure there are some instructions about what kind of input power it needs.

But what you can't do is cause the tie rod you cut to fail at exactly the point where I'm a hairpin turn along a cliffside road. Or the brakes to fail, steering to quit working, and airbags prematurely detonate as I try to come to a stop from 70mph with a semi truck in front of me. I guess you could put some remote detonated explosive or something on a brake line, airbag sensor, and steering linkages, but how long will it stay there while exposed to road and weather conditions? A disguised bluetooth adapter would have a good chance of sitting there for the life of the car.

I do too. I don't think that encryption is necessarily the way to go, but at least some basic stuff like isolating the control systems from the entertainment systems, and hardening the OBDII port against casual exploitation (limit the message rate, inspect for certain kinds of malicious packets, etc.) shouldn't be too hard to do.

Give me a Bluetooth OBDII adapter and I'll show you wireless exploitation. I bet one could be made slim enough that no one would ever notice it plugged in. 5 seconds alone with someones car when they leave the door unlocked (or hell, if an attacker has the knowledge required to send malicious commands to the ECU I bet they have the knowledge to defeat a locked car door. It's not like they need to start the car and drive away).

Benevolent Dictatorship.

Make it clear from the start to everyone on the project that while you're going to remain hands off as much as possible and let everyone do their thing, you're still the ultimate authority and you won't hesitate to step in and start cracking heads if people start causing drama and/or forget how to be adults and let their disagreements get out of hand.

You're going to have to define both depression and what you mean by resolving itself if you want me to agree on that one. Otherwise I'm going to maintain that an actual diagnosed as a serious problem by a medical professional kind of depression isn't going to resolve itself without specific actions on the part of the individual and/or said professional to resolve it. I'll grant you that actions on the part of the individual might come naturally to them in the same way that one naturally favors a sprained foot, but this stuff doesn't just magically go away (unless we're including depression caused by some specific event, which is why you'll need to define the scope for this discussion to be useful).

I'm not saying it won't solve every problem, I'm saying it won't solve the vast majority of actual problems involving real depression/bipolar/BPD/schizophrenia/etc.

But what specific actions are justifiable at what cost? I've often seen people justify things with "if it saves just one life it's worth it", but by that logic dogs and backyard swimming pools should be illegal.

Interesting idea. Though I think a lot would hinge on how well the counseling and intervention worked.

Just curious, you're also against DNRs and various end of life things for terminal patients with painful diseases right? I don't care to argue about those one way or the other, but for the sake of understanding I'm kind of curious.

This is only true for the "oh I don't want to do anything today and my life sucks, I must be depressed" kinds of depression you see people post on facebook. Actual clinical depression is serious, and for the most part won't resolve itself. It's also a symptom of other mental illnesses such as schizophrenia, bipolar disorder, borderline personality disorder, etc. Those illnesses also won't just resolve themselves over time.

At one point I used to think a lot like you in that I assumed mental illness was one of those things where someone could just suck it up and get over it. The best explanation I heard that got me on the path to understanding it was something along the lines of "telling someone with clinical depression to "just get over it" is like telling someone with their Humerus snapped clean in half to lift something with that arm. It's physically impossible for them to do". Only in this case the damage is in software, so it's harder to observe from an outside point of view. Since we don't have direct access into the human brain to reprogram it and fix the problem, we have to rely on things such as medications and therapy in an attempt to get the brain to reprogram itself. This often gives the illusion that someone just "sucked it up and got over it" when that isn't really the case. To continue with the broken arm analogy, there might be a few people who can "tough it out", set the bone properly themselves, and then lay motionless for a few weeks so it can heal without a cast. Then there are the rest of the people who go to the doctor, get it x-rayed, set, and put in a cast. Then there are the people who tried to just "tough it out and get over it" who walked around with their arm dangling by the skin and lived a painful last few weeks until infection killed them.