Why Tesla FSD could FAIL in 2025
HTML-код
- Опубликовано: 1 июл 2024
- Welcome back, in this video, we dive deep into the ongoing saga of Tesla's Full Self-Driving (FSD) technology. With promises of revolutionizing the way we travel, what delays is Tesla FSD still facing? And how does it compare to Waymo, a pioneer in autonomous driving?
Connect with me:
𝕏 - / jobhakdi
🌎 Website- www.jobhakdi.com/
📸 Instagram- / jobhakdi
#tesla #teslastock #fsd 
Наука
The beauty of these AI systems is that they keep improving. FSD technology is already amazing and it will only get amazinger.
Thanks for you analysis! My read on the revenue catalyst for car business in chronological order is 1) FSD subscription 2) CyberTruck mass production 3) Semi mass production 4) Car Repair business 5) Self driving Cabs 6) Robot. I expect all of this would ramp up incrementally and mature in next 3-6 years. This would be enough to push the market value to many trillions.
Thank you! My comment: 1-3 are short term drivers but don't lead to 400+ stock price. 6) could become the central catalyst for 2025 (watch my next video!). 5) will take potentially until 2026 or 2027 to start producing scalable revenue.
Wait for 8/8/2024. Elon has already spoken about an intervention team.
The way I see it: Tesla has three major key levers to have the stock explode: it's core business, FSD and Optimus. 08/08 could provide some hard catalysts, but it doesn't matter - there are so many additional angles that make the probability high this explodes
FSD will work by the end of 2024, beginning of 2025 it comes to EU and reaches 2.000.000 cars in the subscription plan $99/month by Q1_2025, by Q2_2025 robotaxi app/service starts and all existing evaluation models cannot predict the outcome 🚀
It might be in very good technical shape but won't deploy commercially as a robotaxi service at this scale in this timeframe, I am pretty sure about that
@@jobhakdi FSD 12.4.2 is out of this world already, must try it out !!
Thank you very much for actually presenting an analytical, logical perspective of FSD.
Many YT Tesla drivers/an analysts are just wishful thinking stockholders who refuse to rain on Tesla's parade.
I've been using FSD for a month. When it works it's amazing, almost miraculous. Yet, when it doesn't work it's borderline dangerous. I would say I need to intervene 1-3x on a simple 15 minute drive. (12.3.6). Elon keeps proclaiming that the next version will be much better but they never even make it passed the testers. It's confounding that FSD can drive thru a complex round-about but screw up a simple lane change. Thanks
Exactly. Elon has the right vision and first-principle wise he is right with 10x improvements (more like 2x each iteration); BUT in order to actually launch, it needs to be embedded in some operational remote control layer and geofenced. No way around it.
And I have a question. You seem to think FSD/cyber cab is quite a ways away. Forgive me if I misunderstood. Why would Elon be announcing the speeding up of cyber cab production if they were not close, 2025 as he has indicated? It would be a waste of Tesla's time & money.
Elon has hx of over promising and hyperbole
I think Elon gets carried away sometimes, and cybercab production is not going to launch before 2027. I also predict Elon will discover a new shiny toy (or already has) which is named Optimus, and is going to realize that Optimus is 10x bigger and 10x sooner happening than FSD
@@jobhakdiOptimus can’t happen without FSD. I hope you are wrong.
@@jobhakdi why can’t a huge well capitalized company Tesla work on two projects at the same time? Also, let’s see what happens on 8/8.
@@JamesBond-qv2ht They are. It is foolish to see otherwise. All you have to do is look at the financials and listen to the Elon to see that the investment is being made to the tune of billions to make it happen.
Optimus was placed at the top of the Tesla eco system during the last earnings call.
You are missing the most important aspect here. Tesla's energy sector is going to the moon.This quarter's information is right in front of us. Some of you may remember Elon Musk told us directly that energy would be larger than auto because the total addressable market is so massive.
I think that the intervention team ramp up is a trivial part of the problem, assuming that the issue is detected early enough. the car should just send video to the team and let them control the car. You can ramp it up more quickly if you just take very sharp intervention drivers
It's trivial in a technical sense - but it's not trivial operationally and managerially. It's the kind of work Elon hates and that is going to drain him, so let's see if they make it a top priority. It's solvable this year, but only with major management attention
Medium, short term catalyst: I think the non nagging update is a big deal… it will make driving much, much better and will be a good test case for how often there is a crash with people in the car but fully distracted. I think they could allow for FSD for routes that that car has already navigated successfully many times. If you want to go to an uncharted location in the countryside, you would have to be in nag mode again when you reach the place where the car has not been to before.
They will need to go the Waymo route and do tight geofencing
$TSLA: 10-15x in 5 + 5x in 10 = 50-75x in 15.
I am a bit more bullish, >20x in 5
Their is no way FSD can be a hundred percent perfect. It just has to better than the best human to get adopted. Most of the heavy lifting had been done by wemo in terms of legislation.
Remember when people said it will be near impossible to enter china. It will take years to get approval?. Fast-forward today.
Elon is underated in many aspect because of his optimism, but that is his greatest asset. I believe betting against Elon is not a good idea.
agree - never bet against Elon. I wouldn't bet on FSD becoming commercial in the next 12 months. But I for sure would also not bet against it, always long Tesla!
Yeah great video. I'm a bit bearish short term on FSD because I use it every single day and while I do see improvements with every new release, it does seem quite far from having the ability to drive itself. It does seem obvious that you will need major operational systems to account for interventions
Well said!
Great content. Thanks!
Glad you liked it!
Optimus sure feels like an easier problem to solve, as it can be ‘geofenced’ to specific tasks initially and it can mess up without catastrophic results.
It's easier and much bigger !
Do you think that it’ll be easier for them in China in the shorter term? There are 500k auto fatalities a year I believe in China, so even an imperfect system could make inroads, and continue to push for perfection.
China is easier because it's in much worse shape (fatality - wise). It's a great testing ground
How do you know they are not already working on an intervention team or the software to handle one?
Thanks for your insights. My take on FSD so far, is that version 12 was a stunning leap forward, especially when compared with version 11. My concerns now are similar to my concerns with version 11 though. The upgrades appear to generate inconsistency within the new release (just as occurred with v.11). This problem has to be solved so that with every upgrade in function, there isn't a slide or problem that had previously been overcome. Secondly, the external cameras need some type of clearance/cleaning mechanism so that in poor weather or with flying debris they can be cleared automatically. My 2019 Model S does not have that capability. And I agree that Tesla needs some type of "on-call" at the ready, intervention squad to handle potential FSD issues.
Waymo is ahead of Tesla the same way a amateur driver that has driven the same track 1000 times is ahead of a formula 1 driver the first lap. You may win the first couple of races but eventually you are going to endup behind the pro. Now take these two and put them in a brand new track that neither of them know. It's a no brainer Tesla is the formula one driver because the technology will adapt and continue to get better, I wouldn't be surprised if the new version is already comparable safety wise to Waymo. In addition to that Tesla already has a massive fleet and could potentially crank out 100,000 cars a month by just flipping a switch.
That is correct - but the first few rounds may be equal to the first two years, and it makes a difference if we have to wait for two years (or not) for kick starting the FSD value being priced in
@@jobhakdi I say 6 months to some type of ride share going small and geo-fenced. 12-18 months to actual robotaxi in multiple areas still very contained. 2 years full blown robotaxi fleet.
Keep in mind that they already have all the software and hardware needed, I'm sure they have the legal stuff already lined up at least for Texas and maybe Florida since neither state has any laws that would prohibit FSD. Ark has them up and running with robotaxi 2-3 years.
I agree with most of what you say. Probably soft catalyst coming. I believe geofencing is useless without lidar.
I agree that FSD will take a while to launch. I think the next hard catalyst will be the launch of the $25k car in 2025.
Optimus is coming ….
Great videos
I agree completely, the incremental pathway for the ai powered humanoid bot is way easier than the car. The video only, non geo fenced FSD concept is a typical Elon play, a bit like catching his monster rocket with chopsticks, but I would never bet against Elon, as you say it’s all about the timing, what we need short term is revenue growth.
I wouldn't bet they figure it out in the next 12 months. But I for sure wouldn't bet against it, either. If Elon wanted to and makes it a priority, these problems can be solved by end of year.
Looking forward to that Bot video-I agree that the Bot will be “easier” to deploy as they will be much easier to manage the danger versus a car
The bot video is in the works, it's going to be huge!
Appreciate your balanced, realistic analysis. I'm curious; what nationality is your accent? Sharing more of your background, personality, humor, etc. might make you more of a "personality" with a specific approach, which can help your channel grow.
German! Well I don't want to bore people too much and stay focused, but maybe I sprinkle some more personality things in there - thank you for your feedback!
You might have waited until 8/9/24 to release this. I don’t disagree but still things we do not know.
That’s too much of a wait for all of us, I will for sure have a video out right after
06:14 How do you know what Tesla has in capabilities ?
Do they have to share their secrets with the world?
Elon has accomplished astounding feats:
Neurolink, launching rockets into space, creating a start-up electric auto company and making it a $600,000,000,000 world-wide power house. If it was anyone else producing FSD, it would be questionable. However, Elon is not just anyone. He is THE INNOVATOR....If FSD is ever truly successful, it will be at the hands of Elon Musk and Tesla.
Geofencing is for systems that do not have the capacity to learn. That's why Waymo has to geofence an area and map every sign, curb, pothole, and fire hydrant before they can safely let a car loose in their fenced & programmed domain. Then they need someone sitting in an office who has the capacity to drive the car with a joystick when it gets into a predicament that was not programmed in for the system to understand. Tesla's FSD is built on a whole different model. No more programing. It now learns on its own and so geofencing is worthless & needlessly restrictive. A Tesla with a proven out autonomous FSD will drive just the same inside a geofence as it would outside a geofence. When FSD is at the point of 10xs safer than a human, you can let a Tesla car loose with FSD and travel from downtown Los Angeles to downtown New York City and it will get you there 10xs safer than a human can and with no need for a human at the end of a joystick. If this is not true, then Elon has been lying to us.
Well, I do think FSD is getting better all the time and could become 10x better than humans soon. But I also think Waymo solved orthogonal problems that Tesla also has to solve (car getting stuck; etc) and that requires a whole new / additional approach. Elon is only focused on the ultimate outcome, which is fine, but to actually launch, TSLA WILL need to add a tactical layer (which Elon tends to hate). That's why I think it will take longer, and geofencing is the first step.
@@jobhakdi Like I said, geofencing is worthless if you solve the autonomy problem. Why would you fence in an autonomous car that drives in the real world anywhere and everywhere 10xs better than a human? Makes no sense at all.
Geofencing will be necessary for the pilot trials which Tesla will have to do. Need to prove it is safe without supervision. Will obviously need intervention team. Robotaxi is 2 years away at least - may need AI5 but can probably start trials sooner with HW4 vehicles to prove it is safe.
@@facetious1999 some analysts are talking human drivers to prove it out but geofencing makes absolutely no sense when FSD will drive exactly the same on either side of a geofence. A geofence purpose is so that programmers can map out every stationary sign and obstacle to eliminate some risk. FSD learns itself and already knows what an obstacle is even when it has never seen a specific obstacle before because it can compare it with the billions of obstacles like it that it has seen. So like I said, a geofence is worthless accept for maybe a regulator’s piece of mind.
@@RealRadNekit is just practical. Trials will be run within cities and municipalities, each will have its own local government regime. Obviously want to incorporate as large of an area as possible but likely not possible to get a whole state to agree on FSD Robotaxi trial. That will take a long time. To prove safety in trials, just need smaller areas first before can make case for larger trials.
I agree a 100%, the march of 9s is going to be the problem, unless FSD is perfect OR TESLA has found a solution to every possible scenario then I hold my judgement on the hard catalyst of FSD. Optimus could be the real one except that the market for humanoid robots have many many competitors.
I am skeptical on FSD mid term, and I am Uber-bullish on Optimus. I do not think there is real competition, as I will outline in the next video ;)
@@jobhakdiYou could be right about some AI start-ups not being competition but that would take a huge leap in AI, something like solving the AGI holy grail of artificial intelligence.
You're right, there's competition but they're all start-ups losing a shit ton of money every quarter and having to raise more capital.
Tesla already has like $30B cash, lots of workers, lots of factories, expertise in building efficient factories & scaling their business and also expertise in manufacturing.
I don't think the competition will be any real competition.
On top of this Tesla will bring in probably additional +$10B cash into the company this year and grow that cash stack to +$40B.
They can quickly build new factories if they need to.
Good points, but I still have a strong conviction that Tesla already has figured out a plan on how to deal with possible interventions once robotaxis start operation. Maybe we will hear also this on 8/8?
I didn't get the feeling Elon had a plan at the shareholder meeting. He is probably just getting started , but that means it could go fast now. I wouldn't exclude the possibility he makes it a high priority and they have figured it out by year end. I wouldn't bet on it - but I sure would even less bet against it ;)
Tesla is building data center with new NVIDIA chips that will significantly improve FSD. Let’s see what’s the timeline. We are living in AI generation. Technology could improve exponentially.
I agree. But FSD requires too much to be in time. AI translates into much more value faster through OPTIMUS (video coming in the next few days )
It's just realistic. Betting on things happening quickly with Tesla has been a losing game. Betting on things happening eventually has been good.
The question is when "eventually" is
So how do we get 2k by 2030😂
We are going to get 5k by then. Mark my words. The trick is called Optimus
@@jobhakdi I hope you right
Tesla Energy
It will definitely be geo-fenced at first because of the legal aspect where they are not going to get permission to operate everywhere all at once. In those cases I can see them deploying robotaxi with drivers temporarily until they have enough data to show how safe it is. They can probably even make money if not break even with drivers and in the mean time the stock price will start to jump with the projections. $$$
Agree ! But not sure Elon likes these compromises . Would be smoother for the stock
@@jobhakdi absolutely agree I already heard someone say that the remote real-time support is not a priority for him. As per the drivers he may have no option. He's a purist until he has no option and then he becomes pragmatic.
@@didier_777 that’s exactly what I think. Elon lives in dreamland (which is a great thing , because side that’s where the future is invented). He forgets about the pragmatic implementation , focuses on strategic wins. Then, when there is no other choice, he confronts the annoying intricacies of implementation - and gets it done. I think he was a bit behind schedule on this , but now probably embraces it and if that’s the case , it can move fast
FUD
Just pointing out important considerations. Code biz and Optimus make up for it - my Optimus video is super bullish (coming soon )
I don't think Tesla/FSD will fail at any time . . well it will with the woke msm reviews.
People like this are not very bright. Sad