What questions would you ask Elon Musk if you had dinner with him?

[jerry33]

Ibut the vast distances between the stars means that interstellar colonization will be much more challenging.

You'd need a P85 for sure.

 

[Norbert]

I'd ask about what he things are good candidates in the solar system (in addition to Mars) for human colonization in the next 100 years. I'd also ask if he has any thoughts on how humans might colonize other star systems. Getting to other planets and asteroids within a reasonable time is doable with today's propulsion technology, but the vast distances between the stars means that interstellar colonization will be much more challenging.

Maybe someday we can custom-design planets and assemble them from asteroids. Then maybe send some of them on a course to other galaxies.

 

[Raffy.Roma]

I would ask "when the AWD Model S with 110 KWh battery pack, driving assistant package and ventilated/cooled seats will be ready for production?"

I know that it's a very low level question with respect to other questions quoted by other TMC Members :wink:

 

[James Anders]

Elon, what about carbon-ceramic brake rotors? How about pioneering a way to make these much cheaper the same way you're doing on batteries?

There's a huge potential gain in weight reduction and rotating mass.

 

[Jack6591]

How money would be required to build ans test a demonstration model of the "Hyperloop"?

 

[eepic]

How is your team tackling some of the unavoidable 'tough cases' required for even a freeway-only autopilot? For example, consider the case of autopilot engaged and cruising at 100 km/h (~60mph) on a single lane highway and something, perhaps a drunken person or a deer, stumbles onto the freeway 20 meters ahead.

- Let's say the adjacent shoulder or oncoming traffic lane has no cars, and it's possible to swerve there and avoid a collision completely.. will autopilot try to do this if it evaluates it as the 'safest' option? Or will it maintain its lane and just attempt emergency brakes to minimize the imminent impact?

Intuitively, I believe the answer would be it's only allowed to stay in its lane regardless of circumstance and therefore it would only brake. The ability to evaluate multiple non-ideal alternatives and potentially balance passenger safety vs. external vehicle/person safety would be a much more difficult problem for initial versions of this AI. But then it's arguable that a human could easily do better by swerving. To the public eye this would raise a lot of questions, even despite the fact that a human behind the wheel may overreact and flip their car (a much worse outcome).

 

[lorih]

Not a question really, but I would remind Elon that he promised to make the Model X easily adapted for wheel chairs (if Telsa did this is would be so much better than third party option).

I would also tell him that although autopilot makes the most sense for most people (i.e who would own an Tesla and NOT want to drive it?), there are some people that can't drive for medical reasons, and the Google self driving car is going to take care of that, but be ugly, slow, etc.

So..I'd ask him not to forget his smaller market, Tesla lovers that can't drive, and ask him to make a fully self-driving Tesla for me. A girl's gotta dream....

 

[Doug_G]

The correct response is both: brake hard to shift weight to the front, and then immediately steer to avoid the obstacle. By shifting weight to the front you dramatically increase the grip on the front wheels, helping you swerve quickly. It also has thr benefit of slowing the car if there is an impact.

Presumably the car would be able to do this better than a human.

My concerns about a self-driving car would be handling winter conditions: poor visibility, snow collecting on sensors, variable traction conditions, no lane markers visible, etc. I'd be astonished if a computer could figure out how to get a car unstuck even if only slightly encumbered.

 

[eepic]

The correct response is both: brake hard to shift weight to the front, and then immediately steer to avoid the obstacle. By shifting weight to the front you dramatically increase the grip on the front wheels, helping you swerve quickly. It also has thr benefit of slowing the car if there is an impact.

Presumably the car would be able to do this better than a human.

My concerns about a self-driving car would be handling winter conditions: poor visibility, snow collecting on sensors, variable traction conditions, no lane markers visible, etc. I'd be astonished if a computer could figure out how to get a car unstuck even if only slightly encumbered.

That's interesting, never thought about the weight-shift traction. I suppose the control electronics could also optimize that with the traction lost from braking to reach the ideal amount. But I was thinking the difficulty of allowing the car to enter an oncoming lane in certain circumstances would add a lot of difficult cases (say, this situation happens with a sharp bend coming up and the car is not 100% there is no car approaching in the oncoming lane).

I think the exponentially more difficult cases Elon was referring to (first 90% easy, next 9% 10x harder, next 0.9% even harder, etc) was referring a lot to poor weather. In other words, autopilot would initially only be usable in ideal weather and visibility.

 

[Doug_G]

Around here that probably means self-driving wouldn't work for the entire winter.

Oddly enough young people around here are debating whether it's useful to get a driver's license, because they assume self-driving cars will soon make it irrelevant. I think they're delusional. It's not going to be widespread until they can cover the 0.1% cases. Who wants to be stranded whenever the weather turns bad? People would get stranded on highways in dangerously cold conditions.

 

[eepic]

Around here that probably means self-driving wouldn't work for the entire winter.

Oddly enough young people around here are debating whether it's useful to get a driver's license, because they assume self-driving cars will soon make it irrelevant. I think they're delusional. It's not going to be widespread until they can cover the 0.1% cases. Who wants to be stranded whenever the weather turns bad? People would get stranded on highways in dangerously cold conditions.

Yea, pretty much a seasonal feature for us in the great white north. Probably kids who have no appreciation for current day engineering challenges and limitations. It'll be quite a while before we can go fully driverless outside of climates like California (and even there it will be a decade or two at the very least).