-
Want to remove ads? Register an account and login to see fewer ads, and become a Supporting Member to remove almost all ads.
Tesla autopilot HW3
- Thread starter verygreen
- Start date
It has been. I see someone already linked GM's early promo materials, and they're just one company that was working on it.
History didn't start in 2009. You can literally use Google to Google "self driving car history"
I wish I believed in gambling. Because I'd love a good retirement plan. True fall asleep and wake at your destination, level 5 autonomy is without question decades away. I don't doubt it's a century away if it's even possible to begin with. Many of the leading researchers working on this very problem aren't even sure it's a tractible problem.
Not even a hint of irony in your post. That's what I love.
That’s the thing.
Tesla never needed to promise all this. They could have shipped AP2, AP2.5, AP3 all that as they did without ever making such lofty promises and laying out broken timelines. They could have just kept developing the software, releasing features as they matured (maybe charged for such features as they arrived through My Tesla)...
But no. Enhanced Autopilot with Navigate on Autopilot and Enhanced Summon in 2016. Coast to coast self-driving in 2017. Level 5 no geofence feature complete in 2019.
It almost feels as if these promises were meant to drive demand and sales of the car, but in the end I don’t believe that was necessary either. I think hubris got the best of them and now they are finding it hard to let go and instead are doubling down on it.
Or maybe they figured out Level 5 no geofence. I mean they said so and are not retracting it, so it must be true...
I do think Elon knows what he's doing - having him in the public eye saying "we'll do this by then, it's going to be amazing can be a useful fundraising strategy.
However, "Full Self Driving"? I'm afraid I don't think we'll see this.
HW2 vs HW3 = faster image recognition. It's not really progress in a driving sense. Doing something useful with the output of bounding boxes and approximate distance measurements is a whole other ballgame; and an area in which we've seen very little real advancement
I do not believe the current approach to tackling autonomous driving will work for the general public. FSD is not just a software problem. It's not just a technology problem. At some point, I think we'll have to go outside and make real physical changes to the world around us.
I'm more and more convinced that autonomy for cars/trucks will only really be solved by specialised infrastructure. Perhaps this starts with dedicated AV lanes on highways, and newer, smarter roads that are AV only. Then, a slow, expensive and laborious upgrade/transition of the existing road network to enable AV travel.
However, "Full Self Driving"? I'm afraid I don't think we'll see this.
HW2 vs HW3 = faster image recognition. It's not really progress in a driving sense. Doing something useful with the output of bounding boxes and approximate distance measurements is a whole other ballgame; and an area in which we've seen very little real advancement
I do not believe the current approach to tackling autonomous driving will work for the general public. FSD is not just a software problem. It's not just a technology problem. At some point, I think we'll have to go outside and make real physical changes to the world around us.
I'm more and more convinced that autonomy for cars/trucks will only really be solved by specialised infrastructure. Perhaps this starts with dedicated AV lanes on highways, and newer, smarter roads that are AV only. Then, a slow, expensive and laborious upgrade/transition of the existing road network to enable AV travel.
Yes, deception tends to work. That's pretty well established throughout human history.
To bring this thread a little back to the topic of hardware... I’ve only had my HW3 car a few days but so far it sure likes to upload a LOT more than my AP2 car did. I used to maybe push 200MB up every few days, but this new car pushes up GB+ without batting an eye. Either that or I have a bunch of fault logs lol. AP seems to be working fine though.
…and yes, on the subject of hardware: I bet that other than a few token people like social media slores and the random individual here and there to make it look like they're doing something, nobody gets upgraded to HW3 until the second half of 2020, at least in any sort of significant numbers… if even then.
I've been pretty vocal about believing that this is a very stupid approach if they actually give a rat's ass about their customer satisfaction. If they upgrade our hardware now, we are ready when the software goes live that requires it (if such software goes live. I have my doubts but that's for a different thread).
If they don't do the hardware upgrades until there's software that won't run on 2.5 (or 2.0), then their awful service network is the bottleneck for tens, if not hundreds, of thousands of customers who already paid for those features.
So why not do the hardware upgrades now… unless they don't actually think the hardware is going to do what they need, and/or they're just going to save the money for the inevitable lawsuit settlements.
I've been pretty vocal about believing that this is a very stupid approach if they actually give a rat's ass about their customer satisfaction. If they upgrade our hardware now, we are ready when the software goes live that requires it (if such software goes live. I have my doubts but that's for a different thread).
If they don't do the hardware upgrades until there's software that won't run on 2.5 (or 2.0), then their awful service network is the bottleneck for tens, if not hundreds, of thousands of customers who already paid for those features.
So why not do the hardware upgrades now… unless they don't actually think the hardware is going to do what they need, and/or they're just going to save the money for the inevitable lawsuit settlements.
Anyone else seeing this behavior? Pretty interesting anecdote, at the very least.
motocoder
M3 LR AWD w/FSD and premium interior
Data from the "fleet" being used to train the neural nets for a future release, maybe.
My HW2 car used to upload a lot with all the disengagements I used to get since I drive AP in adversarial conditions, but ever since HW3 came out, I found my car simply stopped uploading anything significant. I suspect they're only collecting from HW3 now?
Posted on another thread but HW3 retrofit coming. Confirmed it at my service center today. Not going wide to masses but starting with select group of people. First customer planned here is a Tesla employee.
Not sure when they will open it up to the regular consumer but it’s a good sign that they will keep their word on this one.
Not sure when they will open it up to the regular consumer but it’s a good sign that they will keep their word on this one.
mtndrew1
Active Member
Isn’t every single retrofit an FSD owner? They aren’t going to retrofit any EAP cars with HW3.
thats my understanding. I think this priority is to employees first to gauge/refine the process of the retrofit service before expanding to all the FSD purchasers. Tech today said it doesn’t require a full tear down of the dash like the screen replacement. But caveated he hasn’t done it yet. Wouldn’t get too excited that it’s coming within the next few weeks, but does make it look promising that it is en route.
There's no actual proof or even strong evidence to indicate that FSD purchasers will get early access to anything.
I'll believe it when I see it being done on my car.
Cirrus MS100D
Member
Patiently waiting to hear what they’re going to do about AP2 owners who purchased FSD at order time...
Tesla probably just figuring out final details for our spa day schedule, how we like our steaks cooked, and preferred spirits, to occupy us while they upgrade our cars.
Tesla probably just figuring out final details for our spa day schedule, how we like our steaks cooked, and preferred spirits, to occupy us while they upgrade our cars.
Ah, you must not have heard. The problem is actually Texas and the state auto dealers association lobbying the state to keep Tesla out. Texas is pro-business as long as it's the business already in Texas.
I know you aren't making this argument flippantly, but I absolutely hate this response. It's given so many times by so many people, and it's completely played out. Why would I keep my Tesla? Because the car is a good car. Tesla as an energy / vehicle company is a great idea. Tying to do that, while also trying to make research breakthroughs on autonomous driving systems is not a great idea.
What I'd love to see Tesla do is to iterate on image recognition systems, methodically, over time. Make steady progress, add safety systems nobody else is offering, and continually improve the vehicle without trying to make headlines and turn the ignorant public into test subjects. Simply detecting a driver running a red light or stop sign could have a massive positive impact on occupant safety. That's not a feature that's completely out of their grasp. But go-to-sleep-and-wake-at-your-destination autonomous driving? Simply put, there's no evidence that problem will be solved in a decade or even this century. We've been at it since the 1950s, and we've made poor progress so far.
Be careful giving out stock advice. Even if Tesla is successful, the global economy is fickle and you're asking for trouble investing in individual stocks. This is real f**kin life. Protect yo' godamn neck. Aight!
Step to the Wu. Wu-Tang Financial.
Don't think it's necessarily Texas, this has to do with the mobile service division for Texas and the way they communicate/provide customer service. That's another story.
The argument is used a lot. But it is a way for me to gauge where you are with your skepticism. Are you skeptical Tesla will deliver FSD level 5? Are you skeptical Tesla will deliver anything? Are you skeptical and hate how your car is functioning/working now? Are you skeptical of all things Tesla (ie products, stock, future growth as a company)?
But if you are a guy that has lots of problems with the vehicle, hate the customer service, don't see any future promises being kept. I just don't understand why you would keep the car. Now if you premise your argument with a statement that you really enjoy the car and that it is the best EV out yet but are skeptical about Tesla keeping its promise to provide FSD for your current hardware configuration, it's guides the discussion along a different path.
This next part isn't directed at you, you were collegial in your discussion, especially including the Wu-Tang.
Just like other forums, comment sections on youtube videos, political issues, we have this innate desire as humans to "win" the argument. I'm guilty of that desire as well. When will we realize how futile it is? Even with discussion of FSD, so many people are skeptical, others believe it will happen and so many more are along that gradient. We don't all have to believe the same thing. Especially with FSD, if someone does their research and in their personal situation (financial position/priorities) chooses to purchase, that's on them. Same with someone who chooses not to purchase, it's their decision. Time will ultimately be the judge of who made the right call, if the endpoint of FSD is even the metric.
Similar threads
- Article
