I called out Nav on autopilot 2 years ago and detailed exactly what its gonna consist and people laughed at me and now its been proven to be nothing more than a fancy gimmick. They don't call me bulls-eye for no reason. Besides, The avalanche of myths will flourish once .41 drops with nav on autopilot.
You keep thinking you're the only one that felt like FSD was a complete gimmick. Lots of us including myself knew that the sensors of AP2 would never allow for FSD. There also wasn't a regulatory pathway to it.
I also knew that even parity with AP1 would take a LOT longer than the 1-2 months or whatever it was that Tesla was claiming.
Back then I claimed that Tesla would take at least 6 months for parity. It actually turned out to be a lot longer, and so technically I was wrong. Do I feel bad about being wrong? No, because I was trying to be as optimistic as I could be. That claim of 6 months was an optimistic claim.
I don't know what you claimed because there isn't much point to being right about something like that anyways, and so it's not something I would have paid much attention to.
Why is being right that important to you? Doesn't that limit your ability to put your head out on the table, and your ability to read a threads energy?
This entire thread is Jimmy stating things to the best of his knowledge where he acknowledges there is a lot that he can't know. He accepts that he might be wrong, and he's entirely humble in his delivery. He takes chances, and he admits when he's wrong.
For a thread like this I think it's important to be a bit optimistic, and to think in terms of what could they do considering a best case scenario.
This is a thread for people who are hopeful. Of course it's a breeding ground for myths because that's a bit of where the energy is. If you'd paid attention to most any end-user focused neural network thread they are ALL way more optimistic than the systems can deliver.
To be honest I've never even seen a neural network work correctly even 99.99% of the time let alone more. Except maybe a hot dog, and not a hot dog one.
Despite that I'll try to be optimistic in this kind of thread. Sure I might throw some cold water on the more ridiculous claims, but for the most part I'll just take it all with a grain of salt. It's really not that big of a deal to spend much time with.
I tend to be more optimistic in this kind of thread than I am elsewhere. Like take drive-on-nav for example. I think most people are highly optimistic about that, and I'm in the "Not sure if it's going to work well" camp. I say that because I'm not sure if a neural network will get us there without rear-radar. It's a little hard for me to see a car changing lanes automatically using just a neural network without any kind of rear-radar to back it up. I'm concerned that the edge cases aren't really edge cases at all, and they're hitting limits of what the current sensor system can do.
I certainly don't want to be right, and it's not even something I'll put my vote down on. Instead I'll remain hopeful, and I'll stamp the within 3 months we'll have drive-on-nav with confirmation required and 6 months without confirmation required.
Just out of curiosity where are you putting the drive-on-nav without user confirmation in terms of time-frame? Something that's doable with AP2? Or will it require more? Can it be done with the sensors on the car in a safe manner? Where things like madmax will work?
People tend to get so focused on things a long ways away that we lose focus of the challenge of even the simpler stuff.
