Elon: FSD Beta tweets

[AKAlex]

Except that all those robotaxis supposedly flooding the cities will need to park someplace to save costs while waiting for the next fare. And there still needs to be someplace for the delivery trucks and service vehicles to park so that the street cafe has food to serve and toilets that flush.

I would be curious to know what the average utilization rate is of a car in the city... I'd reckon that it's less than 10%. If utilization could be increased to 20% while keeping demand constant (A probably unrealistic assumption!), it would follow that a city would need only half as many cars.

Because of the lower quantity of vehicles, the space these vehicles were taking can be repurposed. Parking can also be further from high value areas allowing for better utilization of space in these areas.

 

[ThomasD]

According to UBER there are 50,000 drivers in and around San Francisco. 67% of people in San Francisco own cars. 3 million people live in San Francisco. Over 2 million people own cars in the San Francisco area. Even if 50% of people gave up their cars for AV transportation you would never have enough AVs to transport people where they want to go when they want to be there with a reasonable pickup time.

 

[planetary]

According to UBER there are 50,000 drivers in and around San Francisco. 67% of people in San Francisco own cars. 3 million people live in San Francisco. Over 2 million people own cars in the San Francisco area. Even if 50% of people gave up their cars for AV transportation you would never have enough AVs to transport people where they want to go when they want to be there with a reasonable pickup time.

SF population is only 875k.
SF metro area population is 7.8M.

Not sure where the 3M population figure comes from, but I'm sure you can draw a few maps to get there.

 

[Daniel in SD]

According to UBER there are 50,000 drivers in and around San Francisco. 67% of people in San Francisco own cars. 3 million people live in San Francisco. Over 2 million people own cars in the San Francisco area. Even if 50% of people gave up their cars for AV transportation you would never have enough AVs to transport people where they want to go when they want to be there with a reasonable pickup time.

What limits the production of AVs?

 

[azred]

Regardless of your opinion is there anyone who doesn't think Elon is too optimistic on how quickly Robotaxis will make a significant impact to the owner model?

I don’t think he is too optimistic in his robotaxi potential estimates. I think the robotaxi projections are simply blatant bullpoop. What I haven’t figured out is his motive. I can’t imagine many people buy a Tesla with the intent of running a robotaxi business any year soon anyway. So if the projections are not useful marketing ploys, are they just ”fun” comments by a guy who loves to stir up Twitter almost as much as you know who.

But isn’t fake it until you make it just a normal business practice in Silicon Valley?

 

[Sigma4Life]

I don’t think he is too optimistic in his robotaxi potential estimates. I think the robotaxi projections are simply blatant bullpoop. What I haven’t figured out is his motive. I can’t imagine many people buy a Tesla with the intent of running a robotaxi business any year soon anyway. So if the projections are not useful marketing ploys, are they just ”fun” comments by a guy who loves to stir up Twitter almost as much as you know who.

But isn’t fake it until you make it just a normal business practice in Silicon Valley?

I think he’s expecting a significant step change with each iteration of FSD. My guess is he believes version 11 is that major improvement. However I think it’ll be more like version 15 that finally becomes better than human drivers. That’s about 2 years away and still not L5.

 

[DanCar]

I think he’s expecting a significant step change with each iteration of FSD. ...

I'm not sure we should care what he is expecting.

https://twitter.com/x/status/1294374864657162240

Do we have that zero intervention yet that was claimed two years ago?

 

[sleepydoc]

I'm not sure we should care what he is expecting.

https://twitter.com/x/status/1294374864657162240

Do we have that zero intervention yet that was claimed two years ago?

I do on some drives. A lot depends on the drive and the particular traffic patterns, and in Elon‘s defense, when FSD beta was released it certainly was a quantum leap forward from what we had before. Since then I’ve seen a lot of incremental improvements (along with a few regressions,) but that’s what I expected.

I’ve said this before, but I suspect the engineers pay more attention to Elon’s commute than they do to any of ours and any bug that shows up in his commute will almost certainly get addressed more quickly. We just need to have him move around the country to each of our neighborhoods!

 

[Daniel in SD]

Do we have that zero intervention yet that was claimed two years ago?

I think most people can do their commute with zero interventions some percentage of the time.
The funny thing is he says the goal is millions of miles between interventions before robotaxi capability so there are quite a few more quantum leaps to go!

 

[Supcom]

I would be curious to know what the average utilization rate is of a car in the city... I'd reckon that it's less than 10%. If utilization could be increased to 20% while keeping demand constant (A probably unrealistic assumption!), it would follow that a city would need only half as many cars.

Because of the lower quantity of vehicles, the space these vehicles were taking can be repurposed. Parking can also be further from high value areas allowing for better utilization of space in these areas.

So, why would current car owners in a city give up their vehicles? They already have access to public transportation, taxis and Uber/Lyft. And given the high cost of car ownership in large cities, surely, they don't own cars to save money. So, a lower fare from a robotaxi is unlikely to change their behavior.

 

[dramsey]

I do on some drives. A lot depends on the drive and the particular traffic patterns, and in Elon‘s defense, when FSD beta was released it certainly was a quantum leap forward from what we had before. Since then I’ve seen a lot of incremental improvements (along with a few regressions,) but that’s what I expected.

I’ve said this before, but I suspect the engineers pay more attention to Elon’s commute than they do to any of ours and any bug that shows up in his commute will almost certainly get addressed more quickly. We just need to have him move around the country to each of our neighborhoods!

Would it be pedantic of me to point out that a quantum leap is the smallest possible leap?

 

[sleepydoc]

Would it be pedantic of me to point out that a quantum leap is the smallest possible leap?

“an abrupt change, sudden increase, or dramatic advance”

…just to be pedantic!

That’s another expression that doesn’t really make sense - in quantum physics, a quantum is the smallest increment, but the term ‘quantum’ refers to a large amount.

 

[nicholb]

I think it refers more to the leap part. A quantum leap is a change of energy states without going through the states between. So in this case it would be like going from a buggy FSD to one that works without having a sorta works in between.

 

[JulienW]

I'll be Head Over Heels for the Quantum tunneling Beta update........wait I'm head over heels right now and have been all day and so is everyone reading.

 

[EVNow]

Do we have that zero intervention yet that was claimed two years ago?

I can do quite a few trips without disengagements - esp. if I don't care about etiquette and let the car take its own sweet time navigating roundabouts. Infact I don't think I have many safety related disengagements / interventions at all. This also explains why Omar can have so many zero disengagement videos.

So, my guess is - we are at a place now that Elon was claiming we'd be at in Q3 '20.

Anyway, the progress has been slow and definitely lower than the ever optimistic Elon. So, everyone should really be at "show me" mode when it comes to FSD progress, rather than go by what Elon thinks will be achieved in a year's time.

 

[FloridaJohn]

Dallas, TX had a rental bike infestation a few years ago, followed by electric scooters and electric bikes. They've virtually disappeared now. I believe that after the novelty wore off, people decided that you still had to contend with traffic and summertime heat.

Looks like they were actually outlawed in 2020.

Dallas will remove scooters from streets and rethink program

Transportation News

dallas.culturemap.com

 

[EVNow]

Looks like they were actually outlawed in 2020.

Dallas will remove scooters from streets and rethink program

Transportation News

dallas.culturemap.com

Similarly in Bellevue here (easter suburb of Seattle), there was a pilot e-bike program. The company pulled out of it in < 2 years.

Bike Share

The City of Bellevue launched a one-year bike share pilot on July 31, 2018. The electric-assisted bikes were owned and operated by Lime, a private company, under a permit issued by the city. Over the course of the pilot, more than 9,000 people took more than 40,000 trips on Lime’s e-bikes in...

bellevuewa.gov

 

[ThomasD]

The current metro area population of San Francisco in 2022 is 3,318,000, a 0.15% increase from 2021. The metro area population of San Francisco in 2021 was 3,313,000, a 0.03% decline from 2020. The metro area population of San Francisco in 2020 was 3,314,000, a 0.12% decline from 2019. If we just go with San Francisco and not the entire metro area that leaves just over 586 thousand people who own cars. How many of those people could get by without a car at all?

 

[S4WRXTTCS]

I think most people can do their commute with zero interventions some percentage of the time.
The funny thing is he says the goal is millions of miles between interventions before robotaxi capability so there are quite a few more quantum leaps to go!

They're going to need to a better implementation of speed setting to possibly achieve a truly zero intervention commute for a lot of us. There are few areas of my commute where I need to override default settings.

The next low hanging fruit would be giving owners some mechanism to use to fix the maps. I have to intervene to override the map glitch on my way to work.

Then they have to eliminate turning the turn signals on during a corner. This one causes me to turn it off on my way home as I don't want to falsely convey to other drivers that I'm getting off the road.

Lastly they need to resolve the issue with 4 way stops I'm one of those that are impacted by it being too slow to get going again (it kind of stutters it). I also think it starts to stop too late, and isn't benefiting from the regen like it should.

All those things are necessary for me to have any chance at a zero intervention commute, and my commute is only 18 miles round trip.

 

[planetary]

The current metro area population of San Francisco in 2022 is 3,318,000, a 0.15% increase from 2021. The metro area population of San Francisco in 2021 was 3,313,000, a 0.03% decline from 2020. The metro area population of San Francisco in 2020 was 3,314,000, a 0.12% decline from 2019. If we just go with San Francisco and not the entire metro area that leaves just over 586 thousand people who own cars. How many of those people could get by without a car at all?

Can you define what is meant by "SF Metro Area"? It clearly doesn't refer to the city of SF (under 900k), nor any common understanding of the SF bay area (nearly 8M). I'm puzzled.