Search results

I am taking it on both theoretical and demonstrated performance of heuristic vs end-to-end systems. Why won't you answer my simple question? Do you consider the performance of FSD V10 to be comparable to FSD V12?

Yes, and the engineer who did it had to pay damages to Tesla. It is still not known for sure if XPeng actually received or used the stolen data.

I have never claimed that XPeng's system does not exist. I have claimed it is not comparable to FSD V12. And unless XPeng has made some AI breakthrough that nobody here knows about, I am absolutely correct. I notice you haven't answered my question. Do you consider FSD V10 and FSD V12 to be...

Yes I have. They do not provide comparable performance because the XPeng system is using inferior technology. Again, if you have evidence to the contrary, please post it.

This article contain a quotes from XPeng about their "Infinite XNGP", which is their "go anywhere" system. It was first announced on February 29th of this year. So it's quite new. Here is the quote: This sounds like they have taken a similar approach to what Tesla did before V12. I will admit...

It sort of is theory, but it's a theory that seems to have been proven for now. It's still possible that someone could create AGI that achieves automation with a lot less data. But we haven't seen it yet.

You keep insisting that just because you call them "L2" it means they are equivalent. Do you consider FSD V10 and FSD V12 to be comparable systems?

I have pointed out the difference to you already and now again to another poster. If you have evidence that XPeng's "go anywhere" software is comparable in performance to FSD 12, please post it. Again, just because someone defines two systems as "L2", does not mean they are equivalent.

XPeng, but the "go anywhere" version is very new and apparently uses heuristics. From everything I've read, it is not comparable to FSD V12.

All L2 systems are not equivalent. Which part of that fact are you confused about?

You are right, of course. The same thing was said about seat belts and then air bags. We will get there.

As I said, "No other automaker offers anything comparable to FSD V12." I am correct as nothing offered in China compares to FSD V12. You've had this pointed out to you multiple times but you keep talking about L2 as if that makes it equivalent to FSD V12. Why?

Last time I checked, no other automaker offers anything comparable to FSD V12.

In 2017 I got my first EV and after a few weeks I started telling people I would never buy another gas-powered car. Given the millions of EVs sold since then, I guess a lot of people agreed with me and Tesla stock shot to the moon. Since I got FSD V12, I have started telling people I would...

As I recall, Tesla said they still count those crashes as a strike against autopilot.

If you are investing in $TSLA right now, you are betting that some unproven technology will work. And you are betting it will start producing big profits in, say, 5 years instead of 15 years. That's OK with me, but it's not for everybody.

That depends on who the pals are.

Your standard was "robotaxi-level FSD". You will need to define that term before I can comment further.

I agree, but unfortunately, the rate of progress will still be hard to measure. The thing I'm looking for is that 12.4 will be "noticeably better" than 12.3. And I'm looking for 12.5 to be "noticeably better" than 12.4. But when we get to 12.5, it may be harder to notice. Two reasons: 1)...

If you are talking about megapack demand, it is indeed quasi-infinite as Elon said. These things typically pay for themselves in a year or two.

This is not what I said. I said, "I thought of a possible explanation that went under the radar....Employees can buy stock. Employees can show their friends who buy stock. If the right person/people are impressed enough, they might decide to buy stock on the strength of 12.4." I offered it as...

Robotaxi level FSD this year is pretty likely. Of course, it depends on how you define "robotaxi level". But you might just turn out to be the circus clown.

Some are wealthy. And some have wealthy friends. Some may even have friends who are fund managers. Just a possibility.

So we had a big jump in $TSLA yesterday without any clear explanation. There was some positive news, but I've seen plenty of down days on news like we had. So why the jump? I thought of a possible explanation that went under the radar. Besides some new semi info and decent China numbers, the...

As I understand it, this is already happening.

If we are flattish YoY in Q2 I think the market will like that. It could be taken as a sign things are turning around.

Still flying...

That is correct.

Another FSD vs. Waymo video from Callas in Pheonix. Both performed well. Tesla won again by taking a faster route, but it was close. The only challenging part was an unprotected left at the beginning of the drive. Both handled it perfectly. I think that places like Pheonix should be easy...

No.

I've heard it said, "In the US, people are afraid of the government. In Europe, the government is afraid of the people." Movements like eXtinction Rebellion and Fridays For Future have found some success in Europe. In the US, not so much. I think Europe is leading the world in climate...

Unfortunately, I don't think that will happen. The airline business is really tough. And robust air service is seen as vital in any developed country. I'm afraid that governments will actually do the opposite and prop up those businesses if needed.

Same here. Flying is terrible for the planet. And the whole experience has just gotten worse and worse. Even first class doesn't make up for the annoyance of airports and delays. I've pretty much stopped flying altogether. Only once in the last 5 years. Hate flying. Love traveling via FSD.

I don't think he's referring to the "beta" label. It appears he's using the term colloquially to just refer to the manner of testing in a limited release.

Probably true.

When Elon said interventions are rare, I assumed he was talking about internal testing of 12.4 and/or 12.5. Given the context of his post, it seemed like that's what he was referring to.

That's true. If he's wrong we should know very soon. Either 12.4 will be noticeably better or it won't.

I don't care about that just yet. Overall improvement is what matters to me now. The reason is that Tesla is betting that its end-to-end system will continue to improve as more data and compute are added. There is reason to believe that Tesla will not hit a local maximum any time soon. But...

As I said before, I put no stock in the FSD tracker. The data collection methodology is deeply flawed. Testers come and go. Definitions of disengagement is open to interpretation. Consistency in reporting is not enforced. Consistency of routes and time of day is not enforced. The number...

Elon says 12.4 is a 5x to 10x improvement in miles per intervention over 12.3. No matter what number you start from, it's quite significant. Rate of improvement is what matters here, even without all the details we would like to have. At 5x improvement in miles per intervention, we can...

Yeah, I'm thinking the comp plan reinstatement will be voted down. With Elon and Kimbal abstaining, I don't see where enough "yes" votes will come from. The glimmer of hope is that for the comp plan vote, they only need a majority of those voting. This analysis from Harvard says that the move...

I believe this is the first time Elon has given us any clue about their internal intervention statistics. It's extremely bullish that Elon is willing to share this at all.

The problem is that Elon is one of the main attractions over on X. He has to be somewhat provocative as it's good for business on that platform.

Mobileye does not have the data they need to create an end-to-end system that comes anywhere near what Tesla has. They have neither the quantity or quality of data required. I know what NVIDIA does not have. Like all the others, they do not have a huge fleet of cars with the onboard...

No. They can't do it. They absolutely can not match Tesla's system. They don't have the training data.

It's not a matter of "wants". No other carmaker is capable of producing a system like Level 2 FSD w/ Autosteer on City Streets.

No release until fully ready? Then they will never release anything at all.

You must not question the SAE. Theirs is holy writ, the infallible word of God.

My point was that you never know how the street will react or when. But when a breakthrough technology finally looks "real" to investors, they all pile on. The fabled "Chat GPT moment" came about because of what was essentially a demo version of Chat GPT. And related stocks like MSFT and NVDA...

Tesla will initially have the same monitoring and intervention costs as Waymo. But it's not at all delusional to see that Tesla's end-to-end system is technically superior to Waymo's heuristics and hd-maps approach. This is why Tesla's system will quickly start to turn a profit while Waymo is...