Tesla FSD 12.4.1 Lacks Confidence
HTML-код
- Опубликовано: 22 июл 2024
- Sponsored:
Use my special link www.privateinternetaccess.com... for an 83% discount, 4 months free and 30-day money-back guarantee on the best VPN - Private Internet Access
Consider supporting this channel on Patreon: / aidrivr
00:00 Intro
00:50 PIA
02:00 The Drive
23:33 Summary Наука
Really wish my experience with 12.4.1 was as good as others I see 😕
Sponsored:
Use my special link www.privateinternetaccess.com/AIDRIVR for an 83% discount, 4 months free and 30-day money-back guarantee on the best VPN - Private Internet Access
It's only going to get better be patient
As far as know VPNs nowadays are useless when it comes to watching locally restricted contend 😅
@@mrmawson2438 Weird thing to say when it literally just got worse ;p
You are honest about FSD, and your viewers really appreciate that!
That Mars guy is all praise, all the time.
You definitely have a different experience than the other videos I've seen. The parking lot trap was a serious situation if there was no driver
"Main character syndrome" - love it! 🤣
Best description of Cali drivers
Don't be regionist as such people like this exist everywhere
@@strategistaow3520 someones offended
That is a real thing.
You write it like it was just invented.
You are by far my favorite FSD reviewer - I like the fact that you use voiceovers and focus on interactions that demonstrate the abilities and limitations of the software. I also enjoy that you must live near me - at one point in this video you drive right by my son’s apartment 😂
Yes, and too many *limitations* even with latest 12.4.1
Level 5 is years away.
I enjoy this because he doesn't sugarcoat anything. completely fair to the fsd
@@usernamenotrequired142 And to be fair to the fsd, it is *nothing* like a human driver. heck, any high school kid is a much better driver than the simpleton beta FSD software.
The stop sign at about 13 minutes there was a blob on the screen. It’s like the car was visualizing something that wasn’t there.
I agree. I think the shadow on the ground was tricking it.
ive noticed 12.3.6 hallucinate things like this. Ive seen it visualize a gate as a stop sign, and a drive through menu as a stop sign. i think it just means the car knows its appropriate to stop, which isnt really a bad thing.
Dont even own a tesla, nor do I live in the states, but your channel is so relaxing to just sit back, relax and learn all these developments in self driving technologies. keep it up!
1 video from you is more valuable to shareholders than the 100th „no intervention“ drive from other channels.
The system can do a lot. But we have to focus on what’s still missing.
As a shareholder I disagree. The other channels don’t have the same problems which means that the problems might be his car (hardware) specific. AI driver says so himself at the end of the video. I have no doubt some things are 12.4 related as well, that is to be expected.
Having said that, I don’t think it matters much. Because I don’t think FSD will be unsupervised any time soon. “There is less and less interventions so we cannot improve the neural network” yet there are millions and millions of different once-in-a-lifetime edge cases that are extremely important for self driving to master (safety!) that require human knowledge way bigger than just a driving neural net will ever have. Knowledge about humans, about infrastructure, weather, nature, etc. Good luck training on land slides, wrong way drivers, lost ladders in the road, drunk aggressive pedestrians attacking you, stuck red lights, presidential convoy, accident in the middle of a crossing, getting away from hurricanes, escaping an attempted carjacking, etc etc etc. Hell, I even wonder how they are going to handle reading the thousands of different unique, text-only signs in the usa, let alone all the different traffic rules and traffic signs and infrastructures all over the world… It will take a LONG time.
@@qqleq thanks for your thoughts. I disagree on the once-in-a-lifetime events. Yeah, system wasn’t trained on every thing that might happen, but that’s true for us as well. We can still drive a car. We can react to new situations without specific training. If FSD gets similar capabilities it even doesn’t need to be perfect.
The logic would be:
Unclear situation > avoid collisions > come to a stop > request tele operator to take over control.
So there’s no need to handle every situation but just to identify situations it can’t handle safely.
@@qqleq 1) Blaming the car has always been a common excuse. That is not the problem. FSD has all sorts of problems in my area and you can find many other FSD channels having problems too. It just does not perform the same in different places, nor does it improve as fast. Haven't seen any significant improvement in FSD where I live in years, despite the behavior and problems changing.
2) FSD does not need to be trained on every situation, nor does it even need to be able to handle every situation - that is what remote operators are for. The strength of the E2E Neural Net approach is that it learns to generalize and problem-solve situations it hasn't encountered, like humans do. The goal is to get that good enough that a relatively small number of remote interventions will be necessary.
@@catbert7Neural nets as they are now do not generalize the way humans do. Otherwise, they wouldn't need millions of examples to learn something that humans can learn almost immediately.
@@TheSpartan3669 This! Guy pretty much said FSD will turn in to AGI with enough driving related training data..sure thing.
Best FSD RUclipsr as always
We try haha 😆(other RUclipsrs lol) doesn’t help we don’t get the latest updates
12:45 it’s probably thinking there is something in the road as it’s visualizing a blob in the intersection
I'm surprised he didn't even mention it!
It's waiting for the ghost
Amazing channel, I like the attention to detail like the sound effects, just makes the video so much more entertaining!
Appreciate it!
I love the way you put the visualizations on the screen. Love your videos too, very informative and entertaining, even for someone who doesn't have FSD
Lots of good FSD reviewers out here but yours are the ones I look forward to the most. Great work!
I have a theory for what happened at 8:35 where the path planner wanted to go, but the car stayed still.
The new city streets model is supposedly trained with video+controls and outputs control directly, but we also know that they're still running the traditional object detection models (traffic lights) and some of the logic for determining which lane you're in (for the auto-canceling turn signals and to determine which light is yours).
It's possible that they're still using the results of the older systems to add an extra layer of safety on top of the new "end to end" model.
If that's the case, then the tentacle we're seeing is what the city streets model wants to do, and the behavior of it not moving seems like a hard-coded rule, a rule that's got higher priority than the city streets model does.
If anyone remembers when Elon live streamed a v12 drive, one of the things v12 did was run a red light for no reason. I don't know about anyone else, but if I were an autopilot engineer, and I had a bunch of proven software (v11) that could help me ship v12 guaranteeing it wouldn't run a red light, I'd just gatekeep v12's ability to drive unless both v11 and v12 agreed about whether the light was green.
Makes sense!
came to the comments to share pretty much this exact theory (even the part about running the red light on the live stream) but you explained it perfectly already. i think this is by far the most plausible explanation
Yes, this has been my assumption ever since v12 rolled out. Because if you stop and think about it, a pure "video in, vehicle controls out" neural network ... would not create visualizations on its own. It would be a black screen. The car would still do all of the same things you're seeing - but us human passengers would all freak out.
So I assume this is just a "split brain" symptom of having two AI engines running off of the same video feed. One that makes pretty pictures on the UI screen based on all the image training data Tesla has done for years - but that's all it does, just render the world around it. And the other (v12) makes control output signals and that's all.
@@dougtoombs9195 V12 sure does create a "world model" in real time which can be visualized. You can actually see this visualisation when the HiFi Parkassistant view is active.
"video in, vehicle controls out" is a simplification and leads to misconceptions. The controll NN (neural network) is not trained on raw video, but rather on the "world model" which is created beforehand. So a more accurate simplification would be: "Video in -> perception (aka "world model") -> control out".
You are doing a great job on testing bro keep it up
12:50 shadows seem to affect 12.3.6. That seemed similar and even showed something on the screen as being in front of you.
Thank you so much for your honesty. It'd the only honest review I've seen of fsd. The rest I've seen have been tesla fan boys on perfect routs, claiming that fsd is virtually perfect. They have tunnel vision but you are able to keep a level head. Well done and thanks so much.
I had a similar issue where a car drove out into the right-turn only lane from a gas station, blocking my way. FSD waited a few seconds and then instead of turning right at the light, up ahead, it turned into the gas station to navigate around the car... just as I would have done!! However, it ended up going the wrong away around the station and instead ended up in front of the car wash! I laughed and took over. 😛
Top right map is AWESOME!
Welcome to Walnut Creek. This drive was facinating to watch! I was thinking for that unseen left turn light at Broadway/Ygnacio was the use of a Programmed visibility (PV) head on the mast arm. It did see the PV light on the thru movement though. The upright on the left had regular LED bulbs but it didnt want to recognize those either. So maybe the issue was it not knowing you were in a dedicated left turn lane.
Just subscribed I love your humor. Keep it up!
8:57 remember it running the light on the first preview we saw of v12? My guess is that at intersections they still have the old stoplight code running and the car isn’t allowed to move unless the old stoplight code says it can and v12 neural net wants to move.
Much more likely they ve trained rights on red out of the system for this version.
SOOO many Stop signs in America. In the EU we just have priority signs or roundabouts.
American system is completely retarded
What are priority signs? Are those like US Yield signs?
@@AaronJaeger In most of Europe at intersections you have one road with Right Of Way always and another with either a Stop sign or Yield sign depending on visibility.
So no 4-way Stops, or negotiation based on who arrived first.
It's a residential street. You don't need roundabouts.
@@nebulaone908 I've seen very small roundabouts on residential streets in Florida. Better than stop signs if you ask me.
*I always look forward to your videos.
Such an excellent channel!!!*
Your videos are fascinating .
Thanks! Not much but I’ll do my best to tip each of your efforts. Please keep it up!
Truly appreciated, thank you!
ai drivr out here risking his drivers license for our entertainment
Also i live near walnut creek and its really funny to get some commentary on the drivers around here lol
That was a fun watch.
I’m from the area and drive Walnut Creek often.
Keep it up!
If every single car was on FSD, how many crashes would you estimate there to be in the world?
for current versions I think there would be more “dumb” crashes overall but far fewer fatalities
@@AIDRIVR Maybe this is why Tesla mentioned recently that they're focussing on reducing repair costs 🤣
I think that these routes are not optimized even for human vision. Logic of some intersection is just "wait". Maybe we don't even need this good FSD, only fix stupid route engineering to have better UX.
I actually hope they keep improving multiple different models. It would be nice to have the option to have a ‘chill’ mode that doesn’t try to cut around slow traffic and cross double-yellow lines along with an ‘aggressive’ mode that does. During the free trial I flipped from chill to aggressive and didn’t see a single difference.
Good and fair job.🎉
Nice work. Thank you for your service
FSD will work great in northern Europe from day one. No weird left turns, no weird “no left on red”, no “California roll” at stop signs, not too many weird drivers, and a lot less traffic.
And very cohesive, optimized, easy to distinguish road signs.
lol
Umm...we don't turn left on red. We turn right on red after making a stop.
What about in Germany where you have to slow and allow drivers to enter the road where in US you assume right of way on side streets?
Your editing is at FSD12 levels. Just insane man.
I love your videos, there are the best. And there are the only ones I trust.
It’s funny how you describe how humans drives like you came from another planet and you try to explain to your friends how strange humans are
Best FSD youtuber
I think you nailed the summary of how version 12.4.1 behaves. I hope to see how 12.5 or later updates pan out.
Well done. Informative
Thank you!
You're doing Tesla FSD Team a huge service picking apart these issues
AIDRIVER, Wow, your FSD evaluation videos are the best. By far. Thanks!
Thanks!
Well, this is interesting. I'm a bit disappointing the progress the past 2-3 months. It doesn't seem like they're able to quickly make improvements to V12, at least not without introducing other regressions.
Fantastic video as always, sad to see the random lane change regression, but overall looks impressive, especially the u-turns and cul de sac handling.
about 10:50, the car moved over because it's actually the law to move over for maintenance vehicles
I have a 2018 Model X and paid for FSD when it was promised to be revolutionary- I’d say that this version is probably the point at which Tesla has delivered on the promises software back in 2018.
If you’re waiting for a robo-taxi you’re effectively waiting for 1) government approval of FSD (together with legislation changes to traffic rules and at fault liability claims with fully autonomous self driving vehicles) and 2) further software improvements / AI learning for several billion more kms.
I’d also like to see the cars get FSD horns and a loud speaker like KIT
8:45 I also suspect that there is some second system which regulates continuing at traffic lights. An intersection near me is no-turn-on-red, but from the line you can't see the signs which say so (they are too far back.) And as soon as traffic is clear, the path planner appears as though the car wants to make a turn on red, and the wheel turns like it is about to go. But it waits until the light turns green before accelerating.
Possibly some system has memory of the rules of that intersection (or its just based on map data), but the path planner generated from the end-to-end network can't see the no turn signs and so it tries to generate a path to continue and gets overridden.
Tesla FSD definitely drives lightyears better than my mother and most human drivers.
Does you mom have accidents every 10 miles😂🤦♂️😂😱
bro went personal with the mom there, and I couldn't agree more, my mom definitely has "main character syndrome", at least compared to Tesla FSD
No it doesn’t
In the clip where it stops at the intersection and refuses to go through, check the visualization - it seems to be hallucinating an object in the road from either the shadows or pavement or some combination. So an issue with Tesla relying on cameras rather than the FSD decision making?
Good video...good criticism.
Makes me wonder what would happen if you had multiple Teslas all with the same version of FSD and dispatched them one at a time right after each other. Would they all behave the same? Would the disengagments be the same? Would it make a difference with no traffic vs. mild or heavy traffic?
I'm sure not a new question but have you performed a camera recalibration recently? There are some recent anecdotal posts (Tesla Motor Club Forum) that problems FSD owners have had with v12 were resolved with a calibration. No guarantees of course but since your experience doesn't seem to reflect other v12.4.1 experiences seems like a prudent thing to do. If nothing more just to tell people like me that you've done it. Watch all your videos and they are excellent by the way.
21:27 Based on my free month trial, I had tons of interventions and disengagements. Give me a longer free trial and I'll absolutely rake them in in Minnesota where there's of course much, much less data than California.
Always enjoy your videos and commentary.
As a MS 2020 owner with paid-since-new FSD in Australia I do wonder whether we will ever get what we bought?
This feels more like my experience.
I find myself having to intervene more than other videos I’ve seen.
I also probably have less patience than those creating the videos.
Seems like a rousing success! Definitely ready for Robotaxi😂
Cheers mate
🍻
do you feel it is getting better and better?
so no 5-10x improvement...shocking!😮
hey, I might be dumb but where is the snapshot button for my future use?
lol those pist humans honking 😂😂😂
I am wondering: does the performance of FSD change significantly in more challenging weather conditions like fog, heavy rain or snow?
As far as I remember it’s pretty much always sunny in your videos which I’m jealous about for once but I also am wondering if the camera system would struggle in situations with less visibility 👀
Those random lane changes happen to mine as well, its almost like it cant decide which lane to get in, ive had it just straddle the lane at some point
Nice
Could you test if the FSD software recognizes warning lights on cars? It seems like it identified an emergency blinking light on a towed car and gave it extra space. I'd love to know if it consistently detects and responds to such signals.
Seems like they're making it extra cautious for robotaxi?
I know that spot by the Bentley, good for moving overall. Id like a tippy toe preference for that dip out of a drive ❤
FSD hard braking for yellow and flashing yellow is a deal breaker for me, for now. Going to have to cancel until 12.5 maybe
It's the .1 release. 12.3 didn't go wider release until .3 or .4. A lot of low hanging fruit will likely be fixed before you can touch it.
This version was only released to like less than a couple hundred people.
@@tHebUm18 bro I’m on 12.3.6 😭
Seeing your videos, I wonder how Elon's vision of robotaxis will become a reality
Why my tesla didn’t get this update?
My FSD still on Vers 12.3.6
On what happened at the intersection: it seems that multiple subnets of the system have to agree on a situation (lights are green) and in this case a part of the system did interpret the lights as red.
Probably the visualization is not representing the entire system state, but only of a part of the system.
Great content, sad I was really looking forward to fewer disengagements from 12.3.6.
Perhaps an improvement here is that 12.3.6 will sometimes be way too confident to the point of forcing critical safety disengagements. (Happens a lot where I live)
Overall I love FSD and still super bullish, but really disappointed in the apparent regression of 12.4.x
As a product manager in big tech, I think Elon's claim that there's 'not enough disengagements' needs a qualifier to note this means "validated disengagements" -- it's the problem of identifying signal from noise at scale. I guess the voice feedback feature isnt helping with this problem as much as I assumed it would (I use it every time).
We probably need to move past RLHF and go to alpha go style training (AI training AI, reinforcing the model) but this is a hard nut to crack still. I doubt it's a hard limit but general progress in the foundational models has been slowing down and regressing even a lot more than I expected.
That said Anthropic's Sonnet 3.5 is showing initial signs of significant progress so thats bullish; notwithstanding that, its still too early to say we've hit a hard wall with our basic approach as we havent exhausted the work arounds.
12:05 there is a really good Tesla approved auto body shop there. They will fix your car better than when you took delivery. They're called Brooks motors
Totally agree. One of the first things I thought of when this update installed was that it has no confidence. Speed control and maneuver smoothness are also bad.
At 12:53, the occupancy network seems to think something is in front of the car that isn't, which is probably why it wouldn't go.
top things that bother me about FSD: stop signs, going from one lane to two: it doesn't decisively pick one. The way it slams on brakes to stop at yellows instead of smoothly going through.
Everything you pointed out it doing in this version is exactly issues I have with 12.3.6
The autopark perpendicular parking feature could be adapted to do three a point turn imo. It might be as easy as the path planner virtually creating a parking space for the feature to work.
On top of what I said what happens if there is a computer failure when driving with FSD at 60 miles PER hour and the driver is not quick enough to take over?
FSD 12.4.1 is soooo good that no one has it!
One word, SUPERVISED
I find that at a 4-way stop if you want to turn left, FSD is a bit too careful and hesitant, eben when it is your turn to turn.
Glad to see a more critical take on 12.4, I suspect the polished driving style comes with more updates and a more realistic comparison here would be 12.3.1 rather than later 12.3 updates. Regardless as you said they need interventions to feed the data collection, good review.
It would be cool to see what would happen if you put only FSDs on the road in a limited area for testing, that were on continuous taxi mode.
Also I wonder if it would be possible to in the future quick swap on the fly between a less and more assertive mode.
Hello AI DRIVR,
can you add "Lance change indecision" to your statistics?
It’s clear to me as to what the issue was. There is a red light still to the right of the green lights. The code probably say if Red or if there is even a chance at it being red, Do not go! It appears that they haven’t prepared the cameras to differentiate, no turn on red signal from the traffic signals. And actually looking back at it again, it also is making a pair of red lights because it was only the left side that was green.
@8:40 It appears to me that the Rules of being the Left Hand Lane was consistent with what the driving ai wants to do, although, what I think happened, is the lights for the Left Hand Lane didn't appear on the visual HUD and seems to have been lagging. It was picking up the lights for the right hand lane and perhaps the system is a bit behind. I think your assumptions were correct, it was being held back by that Street Lights parameters even still it fully knew what to do next.
I’d like to see your review on 12.4.3
Is 12.4.1 coming with new vehicles yet?
mythical timing pull
Best FSD reviewer...no🧢
I feel like since v12 "seeing" and "making decisions" is trained separately. I mean 1 neural net but on different layers. Maybe this is the reason of camera calibration makes effect. I can imagine that on the confused red/green light situation the car sees the light red, but based on the learned other car movements the decision making logic wanted to move because it felt that time to move.
12:45 it sees something on the road... so it doesn't want to proceed.
12:40 - I dunno if I'm helping at all, but there is in fact a human in that car, flashing brakes and everything. I don't see that reflected in the visuals, yet I wonder if its reading that and getting confused? Although you are right next to them so probably not.
Although I will say, that the car that popped up afterwards, I'm almost 100% certain the car got scared and did an emergency stop cause it didn't show on screen even before the stop sign so I think that's what that was
The behavior differences are really interesting. A lot of this is what worries me about an AI approach. Troubleshooting is going to be so hard.
21:26 I wonder if he was referencing the versions that are widely released. 🤔
Still waiting on the coast to coast 0 intervention drive including charging stops, once acheived that would be a great advertisement
was this on assertive or casual?
Is the car HW 4?
15:45 : Is this the First time that AI Driver acually prefers the previos version over the current?
@AIDRIVR : Your gettin comfy bro :)
I seriously wonder whether Tesla is randomising a bunch of parameters (which greentheonly found a while back) for each card each time you updated the software, to stop "supervisors" becoming too complacent? As FSD continues to improve, yet isn't good enough to drive unsupervised, complacency becomes an ever greater risk. Also it helps further diversify the car's actions so that they can get intervention data on a variety of different reactions to each type of situation.
At 9:50, I have that same issue all the time, instead of taking the right to turn right, it goes straight then tries to turn… has never worked in any version