@callebtc I used to operate this way, and recently decided it made sense to start buying subscriptions from multiple vendors. As a software engineer, the multiple I am getting from AI agents is unbelievable, and there is absolutely value in paying for multiple subscriptions.
Two reasons Jev is an overnight success (two years in stealth, btw).
And they’re complete opposites of each other:
1. Jev unlocks a whole new class of low latency classification and decision features.
2. Folks are doing the math. They’re realizing they’re using expensive LLMs for a huge number of decisions Jev can make faster and for a fraction of the cost. In some cases, the potential savings are an order of magnitude.
The FSD example is actually what somewhat concerns me. In order for FSD to become great, it needed like a decade of real world data to train on. What if Optimus follows a similar path - yes, Tesla needs to build 5000+ in order to build a fleet, but the purpose of that fleet might only be to train and to lead us towards a future where Optimus can do most human-level tasks.
The greatest difference between FSD and Optimus is that FSD is life-or-death so the bar of utility is astronomically high, whereas for Optimus the bar of utility is much lower but the complexity of outputs is vastly greater.
It will be very interesting to see where Tesla is at when they finally reveal.
I think a good proxy for us is to look at what others in the space are doing. If Figure was deploying thousands into factories, I'd basically expect Tesla to be in that ballpark as well in terms of capabilities, etc.
While I believe Tesla has an advantage both in terms of software and manufacturing, it doesn't change the difficulty of the problem trying to be solved.
Is it likely that Tesla has a version of Optimus ready that is 3x more capable than what Figure has been demonstrating? I'm skeptical of that. Does Telsa have the manufacturing advantage - absolutely. But for a product like Optimus, the biggest question in my mind is one of capability.
Figure's demos have been impressive, but still not at the point of being real-world useful.
I'm waiting to see what Tesla's next reveal looks like in this regard.
Yeah, I hear you. Even if the real economically viable product is 2+ years out [very plausible], it's still difficult to know how to think about investment side. Most people are sleeping on the ability for a humanoid to be able to do actually useful things. What happens when they see proof of what's possible?
@WorldlyReviewer Up here in Ontario Canada, that type of turn is illegal which is why it caught my eye - I just looked up Texas driving rules and I see that it's allowed there, so it makes sense that Cybercab is driving that way!
@HansCNelson@Gfilche@Tesla@Tesla_Optimus Right, I think what's interesting here is the thought experiment - those numbers would be instant blockers for most products, but there is still a path to economic viability for a product like Optimus. Obviously not ideal, not the end goal, but also not necessarily an impediment.
@HansCNelson@Gfilche@Tesla@Tesla_Optimus Hans, what are your thoughts on replacement cost - what if in a worst case scenario, we need a new set of hands every 3 months at $8k per set, that's $32k/year. Isn't it possible that Optimus could still be still economically viable now, while they work on improved duty cycles?
@JOBhakdi My expectations are for 10 Cybercabs to start running in the smaller geofence in Austin only (South Congress area). If they launch that way, it's still looks like we have training wheels on. I think for the stock to move, we need something significantly more than this.
@butala_aryan@teslasi Did you watch until the end of Chuck's video? It seems like Tesla has updated navigation to prevent parking near his chain, I'm not convinced that this is evidence the underlying FSD vision issue is resolved yet.
As a Canadian, I can confidently say that having a dry and occluded rear camera is quite literally a daily experience between December and March each year, mostly due to the salt on the roads which mists onto the car. I need to manually wipe my rear camera almost every time I walk past the back of the vehicle. This would be the case for pretty much all of the northern states as well.
I agree with the logic stated here for southern climates, but in the northern ones it will be much more of an inconvenience to keep that camera clean.
@WorldlyReviewer Has been doing excellent on the ground reporting of the Austin Robotaxi fleet, and he's found 12 new unsupervised in the past 2 days.
It's great news to see the fleet starting to scale up a bit, but I also wonder why this wasn't highlighted in the earnings call?
@DevinOlsenn How maddening is it to not get release notes pointing you in the direction of what changed? Do you feel like you second guess yourself on what’s different between each point release, or with enough drive time does it become obvious when a point release changes something?
This images confirms that the Cybercabs currently being produced with decals do not have steering wheels.
Look at the second and third image. To install a steering wheel, Tesla has to remove the white trim piece on the driver’s side of the dashboard because that’s where the steering column mounts. You can see that the vehicle with a steering wheel is missing that white trim section, while the decal-equipped Cybercab still has the trim intact.
$TSLA
@JOBhakdi@JOBhakdi I’m pretty sure that in previous analysis, we learned that when we see the silver reflective piece on the drivers side that meant there was no steering wheel. So while we can’t confirm 100%, I would lean strongly towards no wheel in this one.
I’m in Canada - a friend ordered a Y standard in April and delivery was August (5 months). another friend ordered a Y standard in late June, delivery date is showing in their app as “Nov 1 to Dec 15”. I don’t understand why Tesla doesn’t increase production of the Y standard to meet the demand here in Canada.