Altimeter Capital ($20 Billion Venture Capital company) thinks Tesla FSD 13 has caught up to Waymo, one critical disengagement every 15000 miles and will reach one critical disengagement every 100,000 miles in Q1 2025.
Altimeter Capital has sold all of its Uber holdings to buy Tesla. Brad Gerstner explained the sale of Uber by the belief that Tesla Robotaxi could disrupt Uber.
Tesla reach reach one critical disengagement every 500,000 miles in the middle of 2025. This is the average human level driver for critical disengagements.
Brad Gerstner believes Tesla will be able to remove the safety driving mode for robotaxi in the middle of 2025 by placing the car in robotaxi mode. This is a mode of defensive driving where the car is more careful about speed and how cautious it is. The human will be a passenger who will not disengage but will accept the driving. Just as a human accepts the driving choices of a bus driver.
Brian Wang is a Futurist Thought Leader and a popular Science blogger with 1 million readers per month. His blog Nextbigfuture.com is ranked #1 Science News Blog. It covers many disruptive technology and trends including Space, Robotics, Artificial Intelligence, Medicine, Anti-aging Biotechnology, and Nanotechnology.
Known for identifying cutting edge technologies, he is currently a Co-Founder of a startup and fundraiser for high potential early-stage companies. He is the Head of Research for Allocations for deep technology investments and an Angel Investor at Space Angels.
A frequent speaker at corporations, he has been a TEDx speaker, a Singularity University speaker and guest at numerous interviews for radio and podcasts. He is open to public speaking and advising engagements.
We are going to have to live in Elon world where when you are run over by his disfunctional self driving – it is always your fault. I have a Tesla – 2024 model Y. If you think it takes 10,000 miles for disengagements you must be kidding. But for musk boys, they could care less about the truth – they would rather die for self driving for their idol.
Please seek professional therapy.
Agree.
to me the big question is which city or neighborhood will be the first to outright ban human drivers. when you think about self driving the hardest part is actually dealing with other humans. remove that unknown and its orders of magnitude safer. Will we have a futrue where insurance companies even insure human drivers in populated areas?
Yeah ban human drivers because your self driving can’t drive. Ridiculous ideas. Freedom fighters are just trying to constantly impose something on others – even when their ideas are not ready for prime time
None.
However, the first banning will be in today’s HOV lanes. And if done right, speeds will be increased to 100-120 MPH, with variable based on weather, accidents, etc.
What would it take to crash a Tesla at night? A few flashlights?
Maybe.
2 nights ago, I was using FSD latest on a 2 lane rural type of road with no street lights and is very active. One car out of the stream of oncoming traffic had high beams on and FSD alerted and suddenly braked. It didn’t disengage and thankfully didn’t run us off the road (elevated roadway with no guard rails) in avoidance, it just slammed on the brakes and then resumed. I don’t know the reason, perhaps it lost visibility due to the oncoming high beams.
As an AI person I understand what the limitation is but as an end user this is still far from mass adoption.
This is why I say that FSD is not ready until:
1) cameras have a way to deal with precipitation, mud, etc. and
2) a way to deal with varying light esp with sunlight, bright lights, fog, snow, etc. they need back with hi-def radar/uss.
This: https://www.google.com/search?client=firefox-b-1-e&q=roadrunner+fake+tunnel#fpstate=ive&vld=cid:3d901fa6,vid:4iWvedIhWjM,st:0
Om http://www.teslafsdtracker.com, we see zero increase of miles between safety critical interventions for the whole 12.x series of releases.
And nobody is asking Elon about the difference between the Tesla estimation oand the crowd sourced data during the quarterly reports.
Why??
Depending on what area we’re talking about. The numbers are pretty correct for CA.
Because Teslafsdtracker measures how often a normal human feels that a intervention is necessary and at a certain point its not a useful metric
Imagine you were a goodish chess player and you were testing a new chess computer (before they got really good) when would you choose to override the computers move choice? .. well that depends if you are good enough yourself to understand the computers move choice doesn’t it, there would come a spot where the average human could no longer understand thus would keep over riding even as the chess computer sored beyond them .. same here if Tesla is using very good players (drivers) it would have a much better insight of WHEN the computer was actually critically wrong
That does not mean they have succeeded I am just pointing out that the average persons disengagement becomes a non useful data point
You may be right.
One thing that speaks against it are the videos. There are plenty of videos where the errors are actually objectively dangerous.
Take for example where the FSD makes a U-turn on a 55 mph road, and slow walks the acceleration after having turned 180 degrees. That is objectively dangerous; the cars are approaching with a delta v of 40-50 mph relative to the FSD car.
Or when the FSD sw decides to drive through a left turn lane when the traffic light has turned red.
Note that we don’t have watch 1500 miles of videos to see one these types of examples….
But even if that were not the case, it does not explain why retail investors do not ask about the glaring difference during an earnings call….
Yes but
1. how recent are the videos, what version etc
2. The claim is it will be ready by mid 2025 not now and 6 months in A.I is a long time who knows
3. I am a trained advanced response driver, not great but fairly good, I have watched many videos with disengagements where its the driver that’s just wrong, they clearly don`t understand why FSD is doing what FSD is doing
time will tell we will see what happens
Impossible. When my car had version 12.3 it would try to try to crash into stuff multiple times a week. With V11 it was multiple times of day.
Since I got 12.5.4, it has only made a few mistakes – once every other week or so.
It blindingly obviously has at least a 10 times lower disengagement rate.
My car is hardware 3.
So a company that owns 589,700 shares in Tesla, and has a history of pump & dump, is talking up FSD…well, gee, I better go get a second mortgage & buy more shares. 🥴
Maybe not just yet, since it’s already had a 100+ point move in the last month or so since Musk became best bud with Trump, who just got re-elected. I’m sure that had nothing to do with Altimeter’s timing though…