Elon: FSD Beta tweets

[Daniel in SD]

Elon’s achieved the impossible enough times that he’s earned benefit of the doubt. Starting a successful electric car company and Space X landing rockets are two of the most improbable achievements anyone has ever accomplished.

Electric cars and vertically landing rockets already existed though. Solving a large part of real world AI is way more improbable.

 

[impastu]

Its not just "one year away". And Elon KNOWS that. But if he were to tell the truth about actual timelines that amount if revenue pulled in from suckers who believe him, would be a fraction of what it is when he makes false promises. (ergo, Robotaxi service by 2020). He KNEW there would be no robotaxi service and not "nyc to LA with no human intervention". But when you say that, it gets people excited to whip out their wallets.

I don’t think they would have structured M3 leases with no buyout option if they weren’t bullish on robo taxis.

 

[nativewolf]

I don’t think they would have structured M3 leases with no buyout option if they weren’t bullish on robo taxis.

Lots of reasons to do so and that has paid off handsomely as they resold used M3s for a considerable amount. Sometimes it is just the most obvious explanation. Robotaxis from used cars is never going to be the most obvious.

 

[Knightshade]

I am really curious what all this whining is about on FSD beta? I am totally loving it. I say "Take me to the dog park" and except for parking it does all the work. This is what I dreamed of when I paid for FSD and I am living the dream TODAY.

Lane changing into oncoming traffic isn't QUITE the dream I had in mind though, and I've had it happen more than once, and as recently as 10.8.

FWIW I don't think Elon's a dirty liar or anything- I've defending him more often than not on this- but the idea we're anywhere near what was originally promised just ain't so.

We might be a year or less away from what was promised to the post 3/19 folks though (wide release of L2 city driving)

I don’t think they would have structured M3 leases with no buyout option if they weren’t bullish on robo taxis.

Or they simply realized how high used Teslas would sell for, and how quickly they could sell em.

 

[impastu]

Lots of reasons to do so and that has paid off handsomely as they resold used M3s for a considerable amount. Sometimes it is just the most obvious explanation. Robotaxis from used cars is never going to be the most obvious.

Your opinion. My opinion, if profit from resale of lease returns was the motivation they would have structured MY leases the same.

 

[Sigma4Life]

Electric cars and vertically landing rockets already existed though. Solving a large part of real world AI is way more improbable.

I was being brief in my previous message to avoid writing a dissertation. I’m referring to reusable, vertically landing rocket boosters. Those did not exist before Space X and we’re considered impossible by the ‘experts’.

I agree FSD might be even harder since electric cars existed before. The thing that makes Tesla’s success nearly impossible is there hadn’t been a successful American car company in 100 years. To achieve that AND be done with EVs which the ‘experts’ said couldn’t be done at a profit makes it a remarkable achievement.

Lets not even get started on PayPal. Elon has a long list of remarkable achievements, and what’s even more impressive is they are all in different spaces. From banking to cars to energy to space and now to AI.

 

[Daniel in SD]

I was being brief in my previous message to avoid writing a dissertation. I’m referring to reusable, vertically landing rocket boosters. Those did not exist before Space X and we’re considered impossible by the ‘experts’.

Is there any source for this claim? Because reusable vertically landing rockets were definitely a thing before SpaceX.

The brilliance of Tesla was to realize that there was a market for luxury EVs and to leverage battery technology from consumer electronics. I suspect if you asked "experts" at the time if it would be possible to build a $100k two seat electric sports car they would say that of course it's possible but you'd never sell any.

 

[Sigma4Life]

Is there any source for this claim? Because reusable vertically landing rockets were definitely a thing before SpaceX.

The brilliance of Tesla was to realize that there was a market for luxury EVs and to leverage battery technology from consumer electronics. I suspect if you asked "experts" at the time if it would be possible to build a $100k two seat electric sports car they would say that of course it's possible but you'd never sell any.

Thanks for the link. I still think we are talking about two different things, though the video is impressive given the time frame. Space X was able to build small rockets (hoppers) that could launch, hover and land even in their company’s infancy. What Space X did that was supposedly impossible was build orbital class rockets that can send an object into space, land AND be reused.

You can’t go to another planet and return if you don’t have a rocket powerful enough to get you into orbit and beyond, then land on another planet and be reused to get you back home.

 

[Daniel in SD]

Thanks for the link. I still think we are talking about two different things, though the video is impressive given the time frame. Space X was able to build small rockets (hoppers) that could launch, hover and land even in their company’s infancy. What Space X did that was supposedly impossible was build orbital class rockets that can send an object into space, land AND be reused.

You can’t go to another planet and return if you don’t have a rocket powerful enough to get you into orbit and beyond, then land on another planet and be reused to get you back home.

Why would anyone make a reusable vertical landing rocket if they thought it would be impossible to make one that could go to orbit and be reused? I'd really like to see what these "experts" said because it doesn't make any sense.

BTW the lunar lander was a vertically landing rocket too!

 

[Knightshade]

BTW the lunar lander was a vertically landing rocket too!

How many times did they reuse it? Is it zero?

FWIW, I don't think anybody ever said it was impossible to make a reusable orbital rocket.

They said it was impossible to do that profitably

Market doesn’t justify reusable launchers, expendable rocket makers argue

So what if SpaceX and Blue Origin are launching and landing. Can they profit?

arstechnica.com

 

[jebinc]

Why would anyone make a reusable vertical landing rocket if they thought it would be impossible to make one that could go to orbit and be reused?

Ask Jeff Bezos!

 

[momo3605]

Your opinion. My opinion, if profit from resale of lease returns was the motivation they would have structured MY leases the same.

MY lease structure is the exact same as M3.

 

[Daniel in SD]

This NY Times article from 1995 has no mention of the experts who thought it was impossible.

Companies Hope to Build Rocket That Would Take Off and Land Vertically (Published 1995)

www.nytimes.com

Dr. Charles A. Lurio, a spokesman for the Space Frontier Foundation, a citizens' space advocacy group that has lobbied for a vertical launch-and-landing rocket, said, "Studies show that vertical takeoff and landing is much superior to using horizontal-landing winged rockets for commercial customers, and the best way to making access to space affordable."

Proponents say that vertical-landing rockets are more flexible in their choice of operating sites and would need less preparation time between flights.

SpaceX beating the the ossified legacy aerospace companies is an impressive achievement but definitely not something that was thought to be impossible from an engineering standpoint.

 

[Sigma4Life]

Why would anyone make a reusable vertical landing rocket if they thought it would be impossible to make one that could go to orbit and be reused? I'd really like to see what these "experts" said because it doesn't make any sense.

BTW the lunar lander was a vertically landing rocket too!

The technology used to reach orbit is not the same as what’s required to just launch a small rocket a few hundred meters into the air. The sheer size, amount of fuel required and heat generated by an orbital rocket made landing and reuse seemingly impossible until Elon came along with Space X.

The lunar lander is so different I almost don’t even know where to start. It’s not even a rocket, it’s a space craft designed solely for traveling in space with little to no gravity. The device used to leave the moon was basically the equivalent of a single use bottle rocket, they lit the fuse when it was time to take off. They couldn’t even fully test the system because it was basically a single use explosive.

 

[dramsey]

Since you've worked on software, I'm certain you've had a problem you were working on and were certain you were an hour from being done. 1 hour later you're still convinced you're 1 more hour from being done. Then 5 hours later you're still working on it because it's just not working the way you expected. Now if that problem were as complex as FSD and 1 hour was a year, it's easy to see how year after year Tesla could feel like they are 1 year away.

Yep, that's certainly happened-- many times-- in my career. The difference is that I never reported a problem as "solved" before it could actually be demonstrated as solved. Now, maybe someone in the FSD division got really excited at a meeting and reported that Level 5 autonomy had been "solved" and it was just a matter of implementation. But I kinda doubt it. (The rumors are that the FSD engineers are continually blindsided by Elon's progress tweets...)

And sure, I can understand how FSD is exponentially more complex than anything I ever worked on. How many times has Tesla dumped their entire approach, fired the head of the FSD division, and started over? Three I think?

But you'd think 5+ years of blown schedules would _at least_ imbue Elon with a little bit of humility and the possible better wording of when we could expect it. But...

• He still promised, unambiguously, that everyone who'd pre-paid for FSD would be able to get the beta. Yeah, yeah, I know: NHTSA looking into Autopilot accidents, lawyers doubtless telling him he needed to CYA with the safety requirement. But still.

• It could be worse. I bought an Acura NSX in 2017, right after Tesla announced Roadster 2.0. I was widely derided on the now (thankfully) dead Tesla forums for this, with people who'd happily plopped down $50K or $250K for their Roadster reservation, about how stupid I'd feel in 18 months or so when they were whizzing around in their Roadsters. Yeah...

 

[dramsey]

For those who believed him literally that FSD is just one year away, shame on you. That may have been reasonable to do before the first beta release, but is that super-optimistic rosy outlook after years of testing and constant improvement? By now, we should've all realized how difficult this problem is to "solve". Literally short of creating a true AI-like algorithm that's capable of learning (perhaps like alpha-go) it may never be solved, but that certainly won't discourage me from helping FSD get more data points. If you made the $10K investment in buying FSD, then you have already made your contribution in moving the world forward.

Heck, I'd be happy if Autowipers™ were out of beta. It's been almost 4 years.

Fun fact: No Tesla software introduced as "Beta" has ever come out of Beta. Yes, I just checked, and the Autowiper™ functionality my car gained in late 2017 (which, BTW, doesn't work as well as the sensor-based wipers used in my 2013 Model S) is still shown as "Beta".

 

[mtndrew1]

I’m too lazy to go back and see when the last FSD tweet was referenced in this thread but it’s been a while. Good grief.

 

[dramsey]

But, Elon is the Boss.

That is a good point, and probably the only one that really matters.

I would like to know the inner stories of the FSD software group over the last few years. It seems that they've hired a new head for this group several times, each with great fanfare. One (very charitable) hypothesis is that the current head of FSD tells Elon they're just about there, Elon goes out and says FSD is imminent, it turns out not to be true, he fires the FSD head and hires someone else...

 

[qdeathstar]

It seems that they've hired a new head for this group several times, each with great fanfare. One (very charitable) hypothesis is that the current head of FSD tells Elon they're just about there, Elon goes out and says FSD is imminent, it turns out not to be true, he fires the FSD head and hires someone else...

Well, I don’t think Elon is that stupid. I don’t think the churn in the FSD program is higher than anywhere else…

 

[Daniel in SD]

The technology used to reach orbit is not the same as what’s required to just launch a small rocket a few hundred meters into the air. The sheer size, amount of fuel required and heat generated by an orbital rocket made landing and reuse seemingly impossible until Elon came along with Space X.

The lunar lander is so different I almost don’t even know where to start. It’s not even a rocket, it’s a space craft designed solely for traveling in space with little to no gravity. The device used to leave the moon was basically the equivalent of a single use bottle rocket, they lit the fuse when it was time to take off. They couldn’t even fully test the system because it was basically a single use explosive.

You have provided zero evidence to backup your claim. I have provided evidence that experts thought it was possible and that a sub-orbital prototype was built, flown and reused multiple times.
If Tesla achieves FSD with the current hardware I will be the first to say that they did what every expert in the field thought was impossible.