Elon Musk working on Master Plan Part 3

[Crissa]

Automation will help smaller systems because they'll need fewer drivers to keep up frequency. Once automation exists, it'll be the hottest product.

All systems have busses like that; and even five people on a bus takes less pavement than five cars.

-Crissa

 

[JBee]

With robotaxi we need to rethink why there's a person in the vehicle at all. The car can run errands by itself inbetween human delivery jobs. ;-) (No its not a maternity ward robotaxi, although there might be call for a robo ambulance for less serious cases with telehealth, which is big in Aussie rural areas)

The biggest issue with public transport is the last mile and the speed, convenience and safety of door to door. The question in my mind is what type of vehicle is best suited for this, a van with multiple ride sharing customers or a tandem two seater trike without ridesharing?

The other thing I was thinking is vehicle hygiene. Do we want kids and animals messing up our robotaxi interiors, because they might be travelling unsupervised in robotaxi?

All this leads me to think that there is a big opportunity for smaller vehicles to facilitate robodeliveries and single/dual passenger rides, because efficiency wise its just not really an effective way to let your M3, let alone a CT pickup your pizza.

In comparison to public transport I believe that a smaller, more cost effective and less energy consuming light tandem 3 wheeler would trump any bus or small rail setup, and even a M3 in many situations.

The reasons are:

1. direct point to point routing with no hubs or change overs, therefore shorter effective distance travelled,
2. lighter vehicle operating at capacity, rather than empty and still running,
3. lighter vehicle means less road wear as does direct routing, as well as more distributed road wear because routes alternate,
4. this also leads to less congestion as some freight errands are not time sensitive and can routed around high traffic areas,
5. smaller vehicles make less noise, so deliveries could occur at night effectively doubling or tripling existing road capacity
6. Smaller vehicles use less energy, have smaller batteries and charge using less power, reducing peak loading of grid, especially if there are lots of them alternating between uses this will stabilise the grid
7. In rural areas they could even distribute power instead of powerlines
8. A smaller vehicle like this could be single peice cast, single hub motor, single peice replaceable interior, single door etc

This is the beauty of cellular distribution. We need to stop centralising stuff!

 

[TyPope]

I have zero interest in loaning out my car to be abused by some stranger. Sure, you can monitor it but it'd be sort of like letting a stranger sleep in your hotel room while you are at the conference. I mean, why leave that room empty all day?

However, being able to have my car drop me off at the concert and go back home to wait to pick me up would be awesome. I paid $45 to park at an NFL game once.

Even workplaces... No parking? Send the car home to wait. Of course, if everyone did that, it would double the amount of traffic on the roads...

 

[Ogre]

I have zero interest in loaning out my car to be abused by some stranger. Sure, you can monitor it but it'd be sort of like letting a stranger sleep in your hotel room while you are at the conference. I mean, why leave that room empty all day?

Some people AirBNB their homes or their vacation homes. You don’t need every home in the system for it to make a tremendous impact.

Tesla wouldn’t need you to RoboTaxi your car, it’s likely they need a very small percentage of owners to participate. Uber makes a pretty decent dent and it requires you drive your car. Tesla won’t have trouble getting enough volunteers to sign up for their Robotaxi program.

However, being able to have my car drop me off at the concert and go back home to wait to pick me up would be awesome. I paid $45 to park at an NFL game once.

Then when the concert is over, your car would be in a line of 45,000 other individual cars people had sent home. It would make a lot more sense to have vans pick people up who live near each other 8-10 at a time and drop them off in logical order.

 

[Tinker71]

I have zero interest in loaning out my car to be abused by some stranger. Sure, you can monitor it but it'd be sort of like letting a stranger sleep in your hotel room while you are at the conference. I mean, why leave that room empty all day?

However, being able to have my car drop me off at the concert and go back home to wait to pick me up would be awesome. I paid $45 to park at an NFL game once.

Even workplaces... No parking? Send the car home to wait. Of course, if everyone did that, it would double the amount of traffic on the roads...

I am with you. I suspect the Tesla Transportation system app will rate passengers and robotaxis. High ratings will be coveted. Leave a vehicle a mess and rides there after will be more expensive or not offered at all.

 

[Tinker71]

With robotaxi we need to rethink why there's a person in the vehicle at all. The car can run errands by itself inbetween human delivery jobs. ;-) (No its not a maternity ward robotaxi, although there might be call for a robo ambulance for less serious cases with telehealth, which is big in Aussie rural areas)

The biggest issue with public transport is the last mile and the speed, convenience and safety of door to door. The question in my mind is what type of vehicle is best suited for this, a van with multiple ride sharing customers or a tandem two seater trike without ridesharing?

The other thing I was thinking is vehicle hygiene. Do we want kids and animals messing up our robotaxi interiors, because they might be travelling unsupervised in robotaxi?

All this leads me to think that there is a big opportunity for smaller vehicles to facilitate robodeliveries and single/dual passenger rides, because efficiency wise its just not really an effective way to let your M3, let alone a CT pickup your pizza.

In comparison to public transport I believe that a smaller, more cost effective and less energy consuming light tandem 3 wheeler would trump any bus or small rail setup, and even a M3 in many situations.

The reasons are:

1. direct point to point routing with no hubs or change overs, therefore shorter effective distance travelled,
2. lighter vehicle operating at capacity, rather than empty and still running,
3. lighter vehicle means less road wear as does direct routing, as well as more distributed road wear because routes alternate,
4. this also leads to less congestion as some freight errands are not time sensitive and can routed around high traffic areas,
5. smaller vehicles make less noise, so deliveries could occur at night effectively doubling or tripling existing road capacity
6. Smaller vehicles use less energy, have smaller batteries and charge using less power, reducing peak loading of grid, especially if there are lots of them alternating between uses this will stabilise the grid
7. In rural areas they could even distribute power instead of powerlines
8. A smaller vehicle like this could be single peice cast, single hub motor, single peice replaceable interior, single door etc

This is the beauty of cellular distribution. We need to stop centralising stuff!

You bring up many valid points and this is precisely why the AI and the app is the most important part of this system. The needs, resources, traffic, and price points will be constantly shifting. The AI will run things as efficiently as possible in the short term. It will predict and plan for stadium and conference events, maybe bringing in resources from another city.

Long term it may determine that more Tesla busses are required or as you say 2 seater buggies are needed in a particular zone.

Then think about handoffs. Would it be possible to transfer a package in motion. The AI might look for opportunities to do this.

Ultimately we want to minimize the trip miles. I have ranted before about getting charged by the ton-mile. The AI generally won't send a CT to deliver a pizza when a 2 seater would work better. However if a CT was already on the same route it might actually make sense.

 

[Tinker71]

The more I think about this, the more bullish I get on Tesla AI business. Tesla is thinking levels above everyone else. The possibilities of their AI combined with their energy network and transportation network are insane. The margins on microtransaction will be insane.

The Dojo processor for machine learning and the team they have built and are building will make for an insurmountable lead.

MP3 will be about the AI, maybe a splash of Tesla heat pumps (really part of the energy network).
A few more Tesla vehicles will come, but this will be a smaller share of Tesla in 10 years.

The AI stuff Google is working on is interesting and powerful, but it is not connected to much. Tesla will connect to Terawatts of power and millions of vehicles.

 

[JBee]

What I'm hoping for is that they come up with full car integration into the power network with V2X, plus a currency alternative, like Tesla flyby points, to trade energy and miles. Cut out the middleman.

Then I'm expecting a lot more on Teslabot and Co. and age/medi care range. A van announcement would be nice at some point, but will probably much later in the year after CT is being built.

There's a lot of potential for AI, but I think that route selection and traffic avoidance predictions are probably not that different simply because there is limited variables. But routing of robotaxi will definitely play a big roll in efficiency.

I wonder if the CT will have a auto frunk for placing deliveries in.

 

[TruckElectric]

The Dojo processor for machine learning and the team they have built and are building will make for an insurmountable lead.

The D1 chip - Each of the D1’s 354 chip nodes reportedly has one teraflops (1,024 gflops) of compute. Venkataramanan said the entire chip was capable of up to 363 teraflops of compute as well as 10tbps of on-chip bandwidth/4tbps of off-chip bandwidth.

“There is no dark silicon, no legacy support, this is a pure machine learning machine,” he said. “This was entirely designed by Tesla team internally, all the way from the architecture to the package. This chip is like GPU-level compute with a CPU level flexibility and twice the network chip-level I/O bandwidth.”

 

[TomGriff]

It kind of depends on how you are defining rural. The vast majority of the US population is living in metropolitan areas of one kind or another. My guess would be that robotaxis would work fine even in most towns down to a population size of maybe 10,000. With lower populations though, the ratio of long vs short trips skews much more towards long and the distance between ride drop offs to next pick ups gets much larger.

But it will take time for enough robotaxis to get built so that there are enough available to cover these rural situations.

Rural absolutely holds many definitions, but where I grew up, a town of 500 in NE Washington, we drove 90 miles one way to go shopping once a month, as did most people in that area. It was always a long day of driving and shopping. We are a long way from having an autonomous cars that can make that trip economically and safely, especially in the winter on 90 miles of rural snow covered, twisty highways.

 

[Crissa]

Autonomous vehicles are just fine on the remote, bad roads. It's really the people that are the problem.

Kind of ironic, honestly, we could already be using them in remote, rural places, but we don't - because there's no money out there to do it, no trust or impetus to do it.

So we keep holding it back until it works in literally the worst place - busy, fast downtowns. Rural downtowns could have been solved yesterday. Empty, snowy roads could have been solved. But we have it dealing with poorly marked, badly maintained, busy roads where there's no time to take a minute to match GPS and wait for the humans to get out of the way.

-Crissa

 

[swengl]

What would be nice would be if you could setup your own "network" of friends that you would trust not to trash your vehicle and allow the RoboTaxi service *only* to that group. The profit would have to be really high for me to want to take my primary vehicle and let complete strangers frequently ride in it. I know folks like doing that for Uber and other ride-sharing companies, but it is definitely not for everyone.

 

[Ogre]

So we keep holding it back until it works in literally the worst place - busy, fast downtowns. Rural downtowns could have been solved yesterday. Empty, snowy roads could have been solved. But we have it dealing with poorly marked, badly maintained, busy roads where there's no time to take a minute to match GPS and wait for the humans to get out of the way.

There isn’t a ton of value in partially solving autonomous driving for a small number of people.

If you “Solve” for small towns, you only help out a few people, and even then they will need to go to the city once a month and their autonomous car is suddenly no longer autonomous in some of the worst roads.

Big cities are essential to any FSD solution.

 

[TruckElectric]

What would be nice would be if you could setup your own "network" of friends that you would trust not to trash your vehicle

Have cameras installed in front and back and make riders liable for any damage.

 

[TruckElectric]