Search results

Tesla pushed an update to fix the camera, dunno maybe sometime last year? Don't remember exactly. Painful many years where it was unusable and Tesla Service denying the issue. I'm going to guess they haven't sorted out sending the same update to an MCU2 retrofit in an older car...

They don't need the short-and-distort.

Random question since my searches are failing - is there a calendar or standard date the furthest out LEAP's become available for trading? i.e. expecting June 2026 next.

You're making some unfounded assumption if HW3 had 100x as much compute as it does today that the car would suddenly drive itself. Extremely flawed view. BTW almost all the failures I experience are in static and low speed environments, not that this will change your view.

I'm talking about training because that's been the core problem. Nothing so far has shown HW3 inference does not have enough power, plausible but a different topic. Energy to train a model of a certain size is the right metric. THAT has been scaling very very well. But you were probably...

Statements like these make me root for Tesla to finish quickly even though I'm spending this thread calling out their obvious flaws. It's a talent limitation not a hardware limitation.

This is absolutely incorrect 3 times over. Comparison to general purpose compute cores is completely unfounded. You're just not paying attention. Custom ASIC's used for AI training is because it's doing an entirely different task. It's got nothing to do with VC's. It's all about how many...

That's all the vision/perception part and is essentially solved now, mostly thanks to an explosion in compute power and a tiny bit to improvement in algorithms. But computing proper paths and driving behaviors once the environment is known is just a normal hard engineering problem. Although I...

Oh and by the way, I think the code they used to get it to run a path using the size of the ego car without hitting curbs is also what makes the car randomly turn the wrong way and freak everyone out when it's trying to make a turn where it's not going to clip a curb, i.e. left turn on to a 2...

Vision stack has clearly been refined. But that has resulted in almost no perceptible benefit to an outsider given that path planning was garbage before and still garbage now. Clearly path planning has changed over that time, the biggest change I can recall was when they actually taught the...

do you understand that cars drive on roads?

Sounds like whataboutism. Which part of letting people vote for these locations solves the above? Oh yeah, none of it. Planning the approximate area and size of new stations is perfectly suited to an algorithm.

Same, 11.3.x misses offramps often. Lane changes towards exits are sometimes improved, sometimes not. I'm beginning to suspect no one on the Autopilot team is a native driver (i.e. grew up in the US, learned to drive on things like go-carts even before license age, etc). FSD's driving probably...

People who own cars drive where there are roads. This is not hard. 1. Make all roads reachable. 2. Scale depending on demand. They didn't do 1 yet.

Demonstrating the dangers of direct democracy in real time. Tesla could easily figure out optimal placement by analyzing reachability from map data. They also know already where existing Tesla owners live. The voting is an exercise in futility.

Do what I do now, don't pay attention to the road, look down at the display instead so you can not miss the next nag warning. PS, don't sue me, sue Tesla.

This is not even an FSD thing. Strikes during autopilot are being used against FSD. Although on a road trip today I was remarking at the number of tesla's we saw on the highway that were clearly NOT using autopilot, as they didn't center in lane at all. Funny the autopilot strike happened at the...

Finally got hit with this ~38 days after getting 10.69.3.1+. Apparently it didn't like me keeping me hand on the wheel like I always do, immediate detection and strike. The message was ~"take over immediately", but I see defeat device detected in notifications. Seems like my car without interior...

That said I think something is wrong over there where everything seems to be single threaded through Elon. Old problems just fester until it's Elon's problem. If you tweet or youtube you can make it Elon's problem, if you're lucky. Took something on the order of 4 years to get the backup camera...

I doubt there has been any NoA updates in a few years. I would not expect any until single stack FSD is released.

Same story as it's been since the beginning. One of the routes around here is tentatively getting completed in Q4 after 9 years of waiting, making the trip plausible, but not comfortable (if the power happens to be out as it does during the winter - SOL).

Under-windshield wiper heaters work. Can't disagree with the rest.

Likely your 12V just needs to be replaced, the car refuses to sleep because it may not power on again on a bad 12V.

It is totally if driver monitoring is disabled. And the lawyers (and Elon) will keep that system on, again, for the 4th time in the last page.

NoA is an L3+ system on-ramp to off-ramp. You can argue whether or not it's good at that, and even what the definition of good is, but Tesla is not taking any liability that they don't have to. Why would a company that that got their beta testers pay THEM $10k on top of ~$90k for the hardware to...

They could, the system is effectively has this capability now. But Tesla does what is good for Tesla. No reason for them to transfer liability from the driver to themselves. So keep tugging on that wheel.

I keep hearing about this phantom braking thing, and I'm puzzled. AP1 behavior was great up until later releases where they turned up the paranoia on the radar and it started phantom braking all the time, and I sold that car before it ever stopped. Although I hear there haven't been any AP1...

Procedural just means you using a grey matter neural net to tune the algorithm. In either case, a difference between actual steering angle and commanded steering angle could be used to identify when the path planner is doing really badly. Right now all they ahve is "user disengaged".

I suggested this ~5 years ago on the first release of AP v1, and people looked at me like I was crazy. It makes a lot more sense to allow this for training FSD, where you can make the appropriate arc through a turn even though FSD is engaged. Oh, and it needs the help. I'm surprised how bad the...

disabling TACC on the highway when taking control of steering is clearly the wrong choice and would be very annoying. Tesla should just make it more obvious when the car is trying to control any axis.

Going to take a wild guess and say there is no log data because it didn't happen. Report seems like it is written by someone who has never even used regular autopilot before. Not to mention after weeks of hard work to fight to get the chance to test FSD, the first thought is to file a NHTSA...

Well, Tesla is bringing the charade back to 50 states, eventually. But I guess we have until then to take a break about whining about a dumb algorithm finding tyrannical correlations (not causation) to insurance claims. I would have been out of the game last week but I spent 2 days "driving as...

Anyone notice how BPTRX counts it's allocation of SpaceX? I expected this to go up from 4.2% this week, but perhaps this is not real time? BPTRX - Baron Partners Mutual Fund - All-Cap Growth

red hands/take over does not count against you. The first one could have been FCW after you disengaged autopilot, but I doubt the second. The main problem was trying to use autopilot on those roads.

Well, with the FSD vision stack, highway driving should go L3/hands free long before city streets. If that can't be managed, not a good sign for FSD.

80mph autopilot restriction is interesting. Based on behavior, I believe my Model S has been running on vision for quite some time. Notably, very large following distances and lack of false positive braking from bridges/etc. I'm also able to autopilot over 80. What does this mean? The speed...

Uhh, because we're all gaming the system. Normally I don't take 100+ mile drives at 1am.

You can if you calculate accelerations based on delta velocity and steering angle.

Searching through this thread, I couldn't come to a clear answer: Does the car filter out gravity from it's accelerometer reading for braking and turning? Because it sure feels like it doesn't. Turning slowly around a high grade corner seems to ding aggressive turning.

Loose cells on the ground. Hard to believe that happened without an impact. Fire crew was trained on fighting Tesla fires, so they wouldn't have opened the pack.

Waiting for confirmation that it actually fixes the problem before trying the schedule my car. Who's got a bunch of time to waste?

This is still ambiguous. It could be a simple mechanical variable gear pitch rack or computer controlled variable ratio rack. i.e. a computer decides what the ratio is at any given time, meaning you can't actually know the front wheel angle you are actually applying. The latter is disgusting as...

Now there is a TSB. Seems unclear if this is a permanent fix. Can someone compare the part #'s perhaps?

Legal and regulatory avenues are available.

Don't do any of this. Plug in a USB and take a sentry mode/dashcam capture. Playback the video files on a computer, observe clearly it's recording garbage from the rear camera only, side/front and just fine.

No. The camera/software are experiencing what's known as "black crush". This means video levels below a certain level go to full black, i.e. if it was working right 10% video input would show 10% video level on the screen, but just to make up numbers let's say any signal lower than 20% is shown...

Has anyone filed an NHSTA complaint? My car does this too, and it's not an upgrade. Service already replaced the rear camera 2 months after I picked it up, didn't make a bit of different.

The deja vu.... Is so strong.... If heat is enough, why do you bother showering? Just use a hair dryer to clean yourself. It's heated. After 15 minutes of driving down the highway the day after a snowstorm. Salty liquid covering the highway.

Second thoughts on this. This behavior is probably on purpose. Just like removing the "low" regen setting, I believe this behavior allows Tesla to increase the claimed/tested range of the car. Therefore the behavior will likely remain.

I'd settle for vanilla cruise control working in all weather conditions. Spoiler: it doesn't. Salty water disables the radar, makes the sonar false positive, and blocks all the camera's.