Elon Musk touts data that he says shows Tesla is on the way to fully unsupervised Self-Driving, when in reality, it shows it's years away and misrepresents it.
It's hard to take Musk seriously when it comes to driving times because he's been wrong for years.
Some argue that you can't hold that against him, even though he uses his claims to sell cars and sell “Full Self-Driving” packages for up to $15,000, because it's a difficult and important thing to achieve.
Even if you agree with this argument, there are clear problems with Musk's claims about Tesla's progress and timelines toward unsupervised self-driving.
The biggest is data.
Tesla has always refused to share any data about its self-driving progress. This is despite the fact that they recently started using “miles between required releases”, sometimes called “miles between critical releases”, as a metric for tracking progress and claiming x multiples in miles between critical releases in recent updates without any real data to back it up.
A recent example was Musk hyping Tesla's Full Self-Driving (FSD) software updates 12.4 and 12.5 that they would be able to drive “5 to 10x more miles per intervention”.
Again, Tesla hasn't released any data to back this up, but we have some crowdsourced data that points to the FSD 12.5 reaching 183 miles (all versions combined except testers with less than 50 miles) between critical misfires. Musk never specified what the “5 to 10x” improvement was compared to which version, but if we compare it to the last update, FSD 12.3, the miles between critical separations dropped from 228 miles.
None of the previous versions of the Tesla FSD in the last 3 years could add up to a 3x improvement in mileage between critical intervals. Let's not forget about “5 to 10x.”
Now, Tesla has upgraded to FSD v13 and Musk also said it will “blow people's minds.”
The automaker says the v13 will deliver “5 to 6x improved mileage between required intervals” compared to the v12.5.
That means Tesla expected the FSD v13 to reach between 915 and 1,098 miles between critical intervals based on crowdsourced data.
Currently, after more than 8,000 miles of crowdsourced data, FSD v13 is 493 miles between critical discontinuities:
That's a 2.7x improvement. It's remarkable, but it's also a significant miss compared to what Tesla predicted.
Now, Elon and Tesla fans like to say that this crowdsourced data is wrong and that the Tesla FSD actually performs better.
The data is really limited, but it's the best we have since Tesla refuses to share its data. I tend to fight this accusation both because it is arguably the best data available and because Elon Musk himself has referred to this particular crowdsourced data in the past.
Now, he's done it again and again to claim a “visible improvement” in Tesla's FSD performance, but it's very misleading:
This data only refers to highway miles and Tesla has been using the same highway stack for years. The city driving software stack is unique and based on an “end-to-end neural network”. The automaker has been promising to update it, but never did — leading to the stagnation you see in this chart.
Tesla has been working on this update for years, but it wasn't actually released in v13. It's up to v12.5.6.1. Taking all v12 updates after this one, the highway average was already 393 miles:
This is not an indication of “descriptor development”. It's just that Tesla finally released a long overdue update to the highway software stack after working on the city software stack for the past 2 years.
In addition, if we take this acknowledgment from Musk that this data represents the performance of the Tesla FSD, which should be the case otherwise it would be very misleading for him to share it, it shows that Tesla is many years away from achieving self-supervision. driving despite Musk saying it will happen in “q2 2025”, which is months away.
Ashok Elluswamy, head of FSD at Tesla, has previously said that in order for Tesla to be able to drive itself without supervision, Tesla needs to achieve an average mileage with significant interventions “equivalent to human miles between collisions,” which stands at 670,000 kilometers, according to NHTSA.
So, based on this data shared by Musk, Tesla needs to go from 493 miles between discharges to 670,000 miles between discharges within the next 5 months.
Electrek's Take
I'm not a hater. I am a real person. Without patting myself too much on the back, you have to give me credit for predicting this with v13. After Tesla's AI team released the planned improvements that came with the v13, I reported seeing it get closer to “690 to 828 miles between critical misalignments.”
But then I reported that v13 would lead to improvements but not reach that goal after v13 was delayed for a few months and released with some dumb version.
Now, it ends at 493 kilometers between disjoints. It makes sense. It's an impressive development, but it's also far short of what Tesla has said will happen and still hundreds of thousands of miles away from what Tesla itself says it needs to be to achieve unsupervised self-driving.
Not only that, but Elon is now misrepresenting the data to claim that Tesla has achieved massive growth without proof.
You purposely only look at highway data, which is misleading because the stack hasn't been updated in years.
I think it's obvious that Elon is either lying about driving himself or he doesn't know what he's talking about, which somehow doesn't stop him from making confident statements that happen to help Tesla sell cars. It's not suspicious at all.
Also, I would like to point out that I believe that if Tesla were to develop FSD in a vacuum without Elon Musk making claims about Tesla achieving random self-driving in the x timeline, making “Tesla's cars value assets”, then he would use this to sell cars. and $15,000 self-driving packages, I think Tesla's FSd upgrade will be celebrated.
Instead, it seems more like a scam to many people. That's Elon Musk's fault.
FTC: We use auto affiliate links to earn income. More.