Articles/megaposts by DaveT

[Zhelko Dimic]

I was just thinking that if Musk prevails against SEC easily, it will massacre SP.

Investors like me will be scared to stay fully invested with emboldened Musk, riding high on the wings of victory. We will be scared of him screwing up something badly.

SP would benefit from Musk getting slapped, not getting a pat on the back. Actually, draw would be the best outcome for SP, Musk and SEC both warned to behave.

 

[JRP3]

On the other hand investors like me will be glad that Musk will have been proven right once again, that logic won out, and that another meaningless impediment has been eliminated. I'd think the stock price would take off.

 

[azaz]

Have something that's been bothering me lately and wasn't sure where to post without just being labeled FUD.

So I'm very skeptical that Tesla will achieve full FSD in the next ~3 years, or frankly Waymo or anyone else, simply because there are just way too many corner cases and the current sensor hardware can easily get obstructed in bad weather. Now Tesla has taken a lot of deposits for FSD, and are taking even more every day. Frankly I'm long on Tesla and I ascribe basically $0 value to FSD but I wonder if they are at any legal risk if they can't deliver? I worry about the risk that Elon and team may just be wrong about cameras+NN being sufficient and if that would open the doors to having to refund customers an amount that may be in the billions.

 

[hershey101]

They absolutely cannot do FSD within the next 3 years (or even 5). But no, this won't cause any expenses due to refunds (the contracts are pretty air-tight). It'll probably piss enough people off that they'll stop buying the cars, or at least the upgrades. But that's fine.

 

[DaveT]

Have something that's been bothering me lately and wasn't sure where to post without just being labeled FUD.

So I'm very skeptical that Tesla will achieve full FSD in the next ~3 years, or frankly Waymo or anyone else, simply because there are just way too many corner cases and the current sensor hardware can easily get obstructed in bad weather. Now Tesla has taken a lot of deposits for FSD, and are taking even more every day. Frankly I'm long on Tesla and I ascribe basically $0 value to FSD but I wonder if they are at any legal risk if they can't deliver? I worry about the risk that Elon and team may just be wrong about cameras+NN being sufficient and if that would open the doors to having to refund customers an amount that may be in the billions.

I was just thinking about this. When will I be able to sleep in my car and just wake up at my destination? Doesn't look like it's going to be happening any time soon. Probably at least 5 years out, if not more.

Elon in 2014 said "Maybe five or six years from now I think we'll be able to achieve true autonomous driving where you could literally get in the car, go to sleep and wake up at your destination." (ELON MUSK: In 5 Years You’ll Be Able To Get In Your Car, Go To Sleep, And Wake Up At Your Destination)

Talk about bad predictions. Geez.

 

[JRP3]

6 years from then would be next year, and the technology may exist for that to happen, but of course it won't be legal yet. Of course we know Elon usually has overly optimistic timelines.

 

[wdolson]

Have something that's been bothering me lately and wasn't sure where to post without just being labeled FUD.

So I'm very skeptical that Tesla will achieve full FSD in the next ~3 years, or frankly Waymo or anyone else, simply because there are just way too many corner cases and the current sensor hardware can easily get obstructed in bad weather. Now Tesla has taken a lot of deposits for FSD, and are taking even more every day. Frankly I'm long on Tesla and I ascribe basically $0 value to FSD but I wonder if they are at any legal risk if they can't deliver? I worry about the risk that Elon and team may just be wrong about cameras+NN being sufficient and if that would open the doors to having to refund customers an amount that may be in the billions.

I'm skeptical too. I know a fair bit about how embedded systems work, I've developed a number of them. Though I've never worked on any system quite as complex as a self driving system, I can imagine the challenges. We have had near level 4 self flying planes for many years, but regulators still require two pilots and as we know from the 737 MAX problems, a minor bug that slips through can cause catastrophic problems.

The variables you need to deal with in a ground transport system are many more and each variable makes the problem geometrically bigger.

It's possible that the technological challenges of FSD may be solved in 3-5 years, but it will take regulators significantly longer to approve any system. There is a possibility that the problem is not solvable to a level regulators can approve FSD under all conditions. It's likely it will be approved in limited access highways like interstate highways and in relatively closed loop systems like limited bus routes (such as around airports) it may be approved. There may even be a fair weather certification, but require human control in bad weather.

 

[azaz]

yeah I think getting in the car and login to sleep is at least 10 years out. I think Elon was hilariously wrong about that and its frankly something that he's never addressed. What were the unexpected challenges?

 

[DaveT]

yeah I think getting in the car and login to sleep is at least 10 years out. I think Elon was hilariously wrong about that and its frankly something that he's never addressed. What were the unexpected challenges?

Another thing I was thinking about today was I wonder if the SEC feels like they need to reign in Elon’s over-promising and while they can’t prosecute him for his claims of FSD and such, perhaps all of Elon’s over-promising is the background for why the SEC is so aggressive toward him.

 

[azaz]

Another thing I was thinking about today was I wonder if the SEC feels like they need to reign in Elon’s over-promising and while they can’t prosecute him for his claims of FSD and such, perhaps all of Elon’s over-promising is the background for why the SEC is so aggressive toward him.

I think this is the crux of my worry. how aggressive will they be in 5 years if Tesla can't deliver FSD? Is it something they'd seek relief for customers through a DOJ referral? If say 500,000 people order a 5k FSD option that is never actually full self driving that's a potential 2.5 billion dollar reimbursement.

 

[mongo]

Another thing I was thinking about today was I wonder if the SEC feels like they need to reign in Elon’s over-promising and while they can’t prosecute him for his claims of FSD and such, perhaps all of Elon’s over-promising is the background for why the SEC is so aggressive toward him.

I believe that would be more of an FTC violation, not SEC...

 

[heltok]

I'm skeptical too. I know a fair bit about how embedded systems work, I've developed a number of them. Though I've never worked on any system quite as complex as a self driving system, I can imagine the challenges. We have had near level 4 self flying planes for many years, but regulators still require two pilots and as we know from the 737 MAX problems, a minor bug that slips through can cause catastrophic problems.

I worked with developing SDC. The problem is hard, but progress takes big leaps at times. One day Go is 10years in the future, next day it is ~solved. First part is solving it, that can take anywhere from 0 to 10 years. Second part is getting it legalized. I think Teslas approach is very interesting, getting to scale and getting the data to statistically prove that it is 10x safer than the average driver. Once that is done I think some states will compete in approving it fastest. Some places will ofc lag behind, but I imagine many many places will be pretty quick.

 

[replicant]

I worked with developing SDC. The problem is hard, but progress takes big leaps at times. One day Go is 10years in the future, next day it is ~solved. First part is solving it, that can take anywhere from 0 to 10 years. Second part is getting it legalized. I think Teslas approach is very interesting, getting to scale and getting the data to statistically prove that it is 10x safer than the average driver. Once that is done I think some states will compete in approving it fastest. Some places will ofc lag behind, but I imagine many many places will be pretty quick.

Why focus so much on corner cases? Tesla can achieve self-driving on the most travavelled routes/highway. That's geofenced FSD but still FSD, and a big benefit to the car passengers. I guess we went from horses to cars gradually, too.

 

[wdolson]

Why focus so much on corner cases? Tesla can achieve self-driving on the most travavelled routes/highway. That's geofenced FSD but still FSD, and a big benefit to the car passengers. I guess we went from horses to cars gradually, too.

The holy grail for everyone in the FSD world is the ability to drive anywhere with no driver in the car. That would be a massive money maker for any business who employs people to drive now, but would lead to massive unemployment for anyone with a job driving. That's why all the ride sharing companies are exploring self driving, they want to ditch all the drivers. Tesla has promised a Tesla ride sharing network where people can rent out their cars and have it make money for them when they are doing something else.

Anything less than that is a convenience for drivers, but doesn't make the huge piles of money people are hoping to make off FSD. But the hurdle to cross is to solve all the edge cases.

Semi-FSD (driver still required) is a feature car makers can sell with their cars, full-FSD is a license to print money and a complete revolution in transportation.

 

[Esme Es Mejor]

The holy grail for everyone in the FSD world is the ability to drive anywhere with no driver in the car.

But the “anywhere” part is the least valuable part of the equation. The ability to drive the most heavily traveled routes with no driver in the car will be incredibly valuable. Just being able to autonomously drive major commute routes + connect people with airports & mass transit should be enough critical mass to establish a financially viable network for riders. More routes could then be added each time a corner case is overcome.

The same applies to business fleets. The first application would just be in convoys, with the lead vehicle driven by a human & the rest of the vehicles following autonomously. Then, no human needed on the easiest, most common routes (schools, ports, rail depots, etc). Then, add more difficult routes as they become manageable.

 

[tlo]

I think a lot of people are missing the fact that if it takes many more years to achieve FSD (for any company), then Tesla stands to benefit because it is the only company making money from semi-FSD (in the form of cars sold and autopilot/FSD software).

 

[replicant]

The holy grail for everyone in the FSD world is the ability to drive anywhere with no driver in the car. That would be a massive money maker for any business who employs people to drive now, but would lead to massive unemployment for anyone with a job driving. That's why all the ride sharing companies are exploring self driving, they want to ditch all the drivers. Tesla has promised a Tesla ride sharing network where people can rent out their cars and have it make money for them when they are doing something else.

Anything less than that is a convenience for drivers, but doesn't make the huge piles of money people are hoping to make off FSD. But the hurdle to cross is to solve all the edge cases.

Semi-FSD (driver still required) is a feature car makers can sell with their cars, full-FSD is a license to print money and a complete revolution in transportation.

You're proving my point: "The term holy grail is often used to denote an elusive object or goal that is sought after for its great significance."

I understand the significance of pervasive FSD, and how it might impact Tesla's revenue (and SP). What I don't understand is why people care more about the corner cases of FSD vs the enormous market of geofenced FSD that is right around the corner. Trains don't go everywhere (contrary to horses) but that didn't stop the rail revolution.

I want Tesla to start making money with FSD somewhere ASAP, while they keep working on making the tech work everywhere. That should be today's main target IMO.

 

[colettimj]

You're proving my point: "The term holy grail is often used to denote an elusive object or goal that is sought after for its great significance."

I understand the significance of pervasive FSD, and how it might impact Tesla's revenue (and SP). What I don't understand is why people care more about the corner cases of FSD vs the enormous market of geofenced FSD that is right around the corner. Trains don't go everywhere (contrary to horses) but that didn't stop the rail revolution.

I want Tesla to start making money with FSD somewhere ASAP, while they keep working on making the tech work everywhere. That should be today's main target IMO.

This. I just completed a road trip from PA to FL and back. It's hard to explain the value of EAP during freeway driving to someone who hasn't experienced it. I didn't purchase FSD, but I can easily see the value even if it only works when on limited-access roads.

 

[farzyness]

This. I just completed a road trip from PA to FL and back. It's hard to explain the value of EAP during freeway driving to someone who hasn't experienced it. I didn't purchase FSD, but I can easily see the value even if it only works when on limited-access roads.

Exactly this.

Even if true FSD isn't out until 10 years from now, Tesla will be able to roll out Level 4 autonomy to their cars at a seemingly accelerated pace. In the last couple months we got NoA, merging recognition, traffic light recognition, and enhanced summon. Next will be stop light recognition, round about recognition, and pedestrian crossing recognition.

I think folks are truly underestimating the progress path that Tesla is on right now, ever since they hit their 5k a week milestone with Model 3. This complete excludes the massive cash generation machine that will be Model Y, and the potential of the Pick Up and Semi vehicles.

Everything else is truly just noise.

 

[DaveT]

A risk factor for TSLA is if the economy heads into a recession within the next year or two.