@@Xanthopteryx "that show how bad it still is" is literally a wrong statement since v12 is out, so showing off v11 is like complaining that an iphone 10 is slow and small today
@@AIAddict Look forward to it. I love seeing these comparisons. V12 is not perfect, but it is learning fast. Di this test again with V12. I am sure you will see a marked improvement.
@@axe863 The sensors are not what is important. It is the Ai. More sensors only add cost and complexity. Waymo recently crashed into a large pole, despite all its sensors,
In a future video when you get Tesla FSD Version 12.3 Beta in your car. I want you to do this exact same route for the next video. Looking forward to Tesla FSD Version 12.3 Beta vs. Waymo.
@@AIAddict - Not true! The v12.3 is being rolled out to everyone now - we got ours a few days ago ... this test is not fair since you're using a version of FSD that has been dead-in-the-water for months now! Tesla stopped trying to fix issues with v11.4.9 once they determined that v12.x was already exceeding the abilities of v11.4.9
@@polarlight1369 ”rolled out to everyone” that’s just not the case, the majority of the FSD fleet doesn’t have it yet. I was lucky to get it on Monday and it was a massive improvement and would certainly do a much better job on this route
im curious, how much weight does the waymo system add? are there any plans to actually fit the system into the car? its got to cut range by 10-20%. also i was surprised to learn that the waymo system, without the car, costs well over 100,000. is that ever going to be addressed? people think FSD is overpriced at 12,000 what do the think of waymo at about 120,000?
Bro your life matter the most, nothing else...they why car have airbag.......waymo car take full responsibility ......So they will ensure there system work perfectly.....cost doesn't matter if system work..... people buy iPhone ....even phone are available within $100
@@Tashupost Waymo takes responsibility with huge limitations. Tesla are trying to solve a much bigger problem, and are well on their way to solving it.
no, it does matter if only 5% of the population can actually afford waymo. when the choice comes down to having a place to live or having waymo in your car the price will matter.@@Tashupost
V12 from experience is so much smoother so I think that you should redo this test with v12. Although this was a great test until v12 goes wide! Great Job!
@@Jhaakriwhat’s suspect? he started by showing honest reviews of each major FSD version, taking it on some standard, and tricky, routes to show the weaknesses (most people edit these out) and improvements. He would sometimes film in the middle of the night when he would get the new version. Testing new versions was all that he did. But, he’s also one of the few reviews that won’t edit out weaknesses, or pick super easy routes. One of his “tricky” routes is what got him fired from Tesla, where Tesla, and also stockholders in the comments section, didn’t like how his unedited footage showed that the car would drive on tracks and smash you into poles, with some of the old versions. People think he’s “anti Tesla” because he doesn’t edit his footage to make them look good, and he doesn’t only say nice things, which is good, because the system isn’t perfect, and that’s ok.
I have many issues/problems with this comparison but will try to keep it short and sweet. First, every Tesla on the road (around 6M today) with the exception of very early models is equipped with all the hardware needed for eventual FSD, while only a few hundred (maybe a thousand?) Waymos are out there gathering data etc. FSD costs in the neighborhood of $1000 for their hardware suite while Waymo has been estimated at being around $100k per vehicle. That means the general public can never own the Waymo system, only use it in those sandboxes allowed by their Google maps. Second, this is V11 in the Tesla and they just released V12 which is all about data in and control out, meaning no more written codes, no rules, the system just learns on its own from billions of miles of AI training from driver data. Now that Tesla did pretty well doing what Waymo can do, let’s try it the other way around. Have Tesla do a trip to LA from SF and see how that Waymo does. I guarantee you that Waymo won’t leave its first sandbox while Tesla will tear up hundreds of highway miles without needing interventions. There might be a future for Waymo but I can’t see it. $10+ for a one mile trip? Tesla will start off at least ½ that cost, maybe even lower and then keep dropping from there as demand means more vehicles etc. Tesla’s V12 has garnered massive positive responses from testers, most saying how comfortable and human like the driving is. This is the year that Tesla’s FSD finally arrives, next stop….flip the switch and say hello to Tesla’s Robotaxis!
100k is the total cost of the vehicle with the hardware built in. So the equipment is in the $30k range and keeps dropping as they scale it up. V12 is still at level 2 so we can't say Tesla has a cheaper system at this point. MobilEye expects $5-7k for their L4 HW with 1 lidar and 6 radars.
Also don't mix up "hundreds of hw miles with no intervention" with a level 4 system. Level 4 doesn't need a driver. Waymo in 2015 had 1500mi/disengagement. Almost 10 years ago.
First of all: No. Tesla does NOT have all the hardware needed for self driving. This has been proven wrong several times, and it continues to be false. Second of all: It does not matter how many stock..i mean how much you have invest...i mean what you believe or thing: Tesla is NOT self driving. Waymo IS self driving. Tesla will NOT be self driving this year, nor next, nor year after that. I would say that maybe, a huge maybe, Tesla could reach a L3 state within a few years IF they really really go all in for it. But they don't so... They could also reach a very Limited L4 system (like certain motorways during certain times of day) IF they go all in for it, and then it might be here within five years. But they do not go all into that either. Level 5? Forget it. Will not happen with any Tesla on the roads today. Impossible. Level 5 from another company? Well, maybe? But that is like at least 15-20 years in the future. Mostly for research and testing, not for mainstream public. Could be that it will never happen too.
You guys have been saying FSD is "finally here" since 2016. Having lots of cars that can collect data means nothing, what matters is how well you use that data else it is meaningless. Garbage in, garbage out
I think a letter grade A to F would be better. In a robotaxi test, any human driver intervention is an automatic F IMO. You intervened twice and probably should have when it ran the stop sign. Tesla Fsd gets an F for now.
If you have Android or iOS simply download the Waymo app and yield a ride similar to an Uber! It’s open to the public. It’s available in Silicon Valley, Los Angeles, Phoenix, Scottsdale, Austin, Tempe, Chandler and Mesa!
During the Tesla route, I do say in the video that I hope V12 can help resolve these blunders indicating software level. Sorry if you missed that part.
@AIAddict I caught it which is why I said "not being upfront" and "borderline". Also, the first few comments posted were asking what version FSD was, so it was clear others hadn't caught it. I'm excited to see the comparisons, but given v12 is widely distributed when this was released and it's profoundly different than v11 builds... this video is sort of either obsolete or misleading at time of release and the information wasn't shared well enough to allow people to discern for themselves.
@@AIAddict Thank you for adding the info to the video description. Unfortunately, I dont read the description before choosing whether to watch so it needs to be in the title or stated at the beginning of the video, but having the FSD version in the description is better than nothing.
We all know Waymo is really good. It has over ten million robotaxi miles in real cities with no bad crashes. What we need for Tesla V12 is a city mileage test, as in driving around in SF for an hour at a time over many months, totalling maybe a thousand miles. Real AVs are evaluated over tens of millions of miles, because driving is safety critical; one bad crash in 100,000 miles makes you a terrible driver; and every driving record is just a list of crashes, not a list of all the good days where you didn't crash. For a big AV fleet, one bad crash in ten million miles is potentially catastrophic. If Waymo were to plow into a crowd of pedestrians, it could sink the business, even after ten million really safe robotaxi miles. Testing FSD over a few miles is silly, unless it messes up that fast, indicating that it is so bad that it's a joke. So let's see V12 in a series of one-hour drives in SF during business hours, or maybe half-hour if you don't have the time. What is the FSD failure rate, or average time between failures? We won't get this data from Elon or any of the fanboys. They don't want to show the bad stuff, or real driving record. I think you John are perhaps the only RU-vid FSD tester who wants to just show the raw truth about FSD, whether good or bad.
Bro your life matter the most, nothing else...they why car have airbag.......waymo car take full responsibility ......So they will ensure there system work perfectly.....cost doesn't matter if system work..... people buy iPhone ....even phone are available within $100
v12 will actually stop when it's close to the pin and ignore the rest of the route. it's an interesting override. It can even do a short cut near the end. It's curious why it tried to take a left at the start, if it's conditioned to fear that winding road. Maybe the drivers have chosen to avoid it when nav went there and it mimics that. You can test it on the light rail in san jose that older versions were confused by.
Reading these comments there are a lot of butthurt Tesla fan boys here complaining he did not use V12. The vast majority of the Tesla community does not have that software yet. I hate to be the bearer of bad news but current Tesla’s will never reach level 5 and I think Tesla knows that it’s been offering “FSD” for over 8 years and is still level 2 Tesla is also deceiving people with selling “full self driving” for 12k when clearly it is not capable of doing it how they aren’t hit with a false advertisement suit needs to be studied.
I love how Tesla fans are so toxic just because waymo was better their feelings got hurt lmao. to all the people who say this isn't fair because Tesla doesn't have version 12 y'all did the same to Waymo years ago. When waymo was just released and it didn't want to make left turns or it took a whole detour because the car wasn't confident on busy streets Tesla fans in that time were clowning on waymo. now that waymo has upgraded their driving skills and it's clearly shown Waymo doesn't drive like how it used to 2 to 3 years ago, it's a problem now that Tesla doesn't have version 12 ok lol. This test was completely fair let's see how waymo does when you do it again with volume 12. If it doesn't beat waymo I don't wanna hear anything about wait till volume 13 guys 💀.
I have been following self driving for years, and both systems have their plusses and minuses. I don't think anyone can argue Waymo is the safer option. But with huge limitations. Tesla are tackling a much bigger problem, and if they can solve it, will be in a dominant position in this space.
@@thewatcher5822 I agree but I just don't like how Tesla fans can't give waymo credit it's literally the first driverless car that's already training on the highway. the only reason why waymo can't expand fast is because they have rules to follow and many people also protest against them. Also 1 Crash can damage their whole company which is why they only expand when it's safe. Tesla literally ignored a stop sign, almost crashed no one said anything everyone is like where's version 12, but if waymo did that it would be on the news talking about driverless cars are dangerous etc.
@@pennytrui1149 There is no doubt Waymo taking on all responsibility is a huge weight to carry. Which is one of the huge advantages Tesla has. Keeping it's system at level 2 and putting responsibility on the driver is a very clever strategy by Tesla. Like I said both systems have their plusses and minuses. From what I am seeing, Tesla are best placed to dominate in this space. Time will tell.
You engaged FSD on the fly less than two seconds from a stop sign going too fast and the Tesla was to blame? Would you prefer it slammed the brakes hard? Also when testing FSD, need to know the version you are using. Jerkey "I don't feel safe" was noted for V11.x . Early 12.x had some recidivism. Current V12 has 1/3 the interventions of V11.x "They have been at this since 2016, don't see it improving much". Version 12 is whole new architecture (about a year or so old) and has been progressing much quicker than 11.x and previous due to AI learning and is more responsive due to much shorter cod stack. The "brute force approach" uses 9 cameras and maybe $2k worth of electronics and wiring. The Waymo approach uses 26 cameras, multiple lidars and ultrasonic sensors totalling 51 sensors, maybe $50k worth of electronics, they have to map changes in each city they support and at Least 50 lbs of additional weight. I would argue that Waymo is using the "brute force" approach. Also worth noting is that Tesla can be driven anywhere, while Waymo is geofenced to 4 cities. Tesla 12.x is moving along rapidly now (given actual machine learning) vs the slow progress of previous versions. So I imagine there's a very good chance that the gap will close over the next 12 months. This is speculative, but AI learning appears to be on that trajectory and that started with V12... or we will hit another ceiling and have to wait again... Time will tell (probably about a year). Can Waymo map the top 50 cities on each continent rapidly once they get a robust product that they like? I haven't heard anything but time will tell.
6:05 you should've brought a traffic cone .D 8:29 this what makes this comparison a bit skewed (I'm assuming you're driving your own Tesla). In your Tesla you're responsible for any damage by sitting in the driver's seat (or at least feel responsible, as IDK what's in the EULA, but I assume nothing good for the "EU"). Still the critical blunders (see below) in the test should struck-out FSD from public roads 10:07 sorry, FSD did not. 10:26 exactly what I've meant (stop sign, getting lost (less safety-critical, but still functionality-critical), passing a vehicle on a blind hill).
@@StefanNorbergi think it would be a good idea in case of an emergency. imagine being stuck in the car that is driving down the road when you want to exit. its a pain to click on the touch screen and politely press the "buttons" to make it stop. even worst for the back passenger. in a real car you can scream at the driver to stop.
@@Francois_Dupontit has a stop button on the front and back screen for the passenger that literally take like 2 presses to make it stop . When you click it the car will immediately look for a spot to pull over and immediately go to the side of the road. Also the only reason it takes 2 presses is because people have accidentally pressed it in the past and the waymo immediately pulled over and immediately ended their ride. So people were constantly calling oh no i wasn't scared I pressed the wrong button etc. you're really not stuck when you get inside you can even open it while it's driving you might get called by their support team tho 💀
@@AIAddict "top of the hill" and "half block away" is completely different than "middle of the road". stopping in the "middle of the road" is illegal and dangerous. stopping on "top of the hill" could mean beside the curb which is safe and legal.
have you noticed something strange? your video has only 2000views~ and many people are writting the same comment complaining about the FSD version and saying they are unsubbing and stuff. it really doesnt seems organic. it woudnt surprise me if SOME company were hiring ghost account to complain.
Np, actually he doesn't. You're both dead wrong if that's what you honestly think. Waymo is very screwed, it's obvious to anyone with a brain. Waymo is not even close to impressive as fsd Beta. Only one brand will have self driving cars...
@bobhope3940 One of the only brand today that has 100% driverless car. One of the only brand that takes the full responsibility of any accidents. This is also the best self driving car. That is Waymo. Edit: Added one of the only as they are cruise and zoox too.
@@violety_indigo52 Best self driving? Maybe in tiny areas of America. But best can mean many things. Is Waymo the cheapest system? Nope, that would be Tesla,.Can Waymo be used across America? nope that would be Tesla. Can Waymo use highways? Nope that would be Tesla. Has Waymo got more real world miles of data? Nope that would be Tesla. Waymo is very good at what it does, But has huge limitations, as does FSD. But from what I am seeing Tesla are well on the way to solving Self driving.
I get that you don’t have v12 yet, but you really should make it extremely clear that you are on a dinosaur version of fsd. 12.3 is a whol different level than 11.4.9 and you basically didn’t even mention it all
A “dinosaur version” ? V11.4.9 is the LATEST software that went full release. In fact, it was a NHTSA Recall Update version. V12 has been pushed to less than 10% of users. Regardless, V12 still requires a human. Waymo does not. We plan to showcase V12 with the same test when it goes wide release.
@@AIAddict yes. A dinosaur version. The end to end nueral network of v12 is years in the making, and it’s entirely different than version 11. Also, you even admitted in the comment section that you pumped this video out quickly when you found out about v12 release. And didn’t even mention it in the video. Clearly prioritizing views over accuracy
@@AIAddict v11 might be the latest before this, but it’s an architecture that is years lagging what’s out now. Your channel is named ai addict, surely you should know this
@AIAddict you mean the non recall nuisance recall that was accomplished over air to put bigger fonts on the warning to remind everyone that this was a beta software and that the driver was still actually ultimately responsible for driving the car. Something that you may well be advised to take heed of.
This is one of a two part episode Next upload will showcase the exact same test but to see how vastly different V12 handles it. Can’t wait to show you. This test is on 11.4.9 which is the current version my car is on. This was a NhTsA recall update, should work.
The latest version can not even reverse... And the last video a saw earlier today, it stopped on a railroad track. Duuumb waaays to die... humdidum.. **singing**
Yeah! most of 11.x version builds were not reliable. would love to see v12 vs waymo. when's the part 2 dropping btw. hope it'll be sooner then later cause i could see this video being misunderstood by many as fsd's current state which would be false from what i've seen from other testers.
@@AIAddict I don't think it is an excuse. V12 has seen a huge leap in performance, so it is important for you to mention which version you are using. I didn't mind watching this video, it was obvious to me it was not V12, and it will be interesting to see how V12 handles it. and the improvements or otherwise being made Looking forward to it.
@@thewatcher5822 Of course it isn't an excuse. There is something out there that is just blowing minds and this clown is showing stupid videos of the outdated stack as if it means something. I remember when this guy was the guy to watch. He was the first to take FSD on lombard street with V10. He would consistently test fsd in downtown SJ which is just as sh*t show to drive in. Now, he got his dumba*s fired from tesla and he is just bitter and trying to remain relevant. His strategy now is to be Dan Odowd's BIT*H
@@dsfsdfpewofca8280 Tesla has over a billion miles driven on FSD. Just because a person was supervising doesn't mean the car wasn't driving. It is one of the reasons Tesla can go with a pure neural net.
Clickbate… releasing this video after v12, which most ppl say is a huge improvement, and not mentioning that your video is in v11 in the title. Well if u did, most ppl wouldn’t watch it I guess.
We plan to showcase V12 on episode two of RoboTaxi Wars. We filmed this episode before V12 went into people’s hands. Had to scramble since V12 dropped while this episode was being edited.
@@Xanthopteryx there was a major leap in capabilities with the newly released version 12. We are talking about beta software here, where every version can and will change things massively.
Waymo does not pay influencers (as far as I’m aware). The Waymo maintaining 6mph down the curvy stretch felt as smooth and comfortable as if I was being driven by a chauffeur. The Tesla on the other hand was very jerky. Constantly using accelerator and brakes like a new driver.
@@AIAddictpeople whenever you compliment waymo must be paid lmao Tesla fans are so toxic omg. Just because you prefer the way waymo drives means you got paid by waymo what lol even you said the ride was expensive 10 bucks for just a mile. waymo will always drive better then Tesla because they are responsible for a every crash. Tesla just says oh you crashed you should have been paying attention while waymo can get sued. This is why Waymo expands slowly they also have to follow rules etc
That is literally human like it shows the car is confident. So in you opinion braking every 0.5 jerking the car around wouldake you more confident got it lol 💀😂.
@@pennytrui1149 V12 is the version that is showing true human like qualities. Which makes sense as it is learning from videos of human driving. Gone are the heuristics of earlier versions, It is purely end to end Neural Nets,
V12.3? If not then this video was dated before you even posted it, not a good comparison. Always show the latest when running comparisons so that you are showing the best there is offered by a brand.
11.4.9 is the latest wide release. Been waiting to receive V12 for 5 months and currently less than 10% of FSD owners have it downloaded. Happy to re-run the test when V12 actually goes wide
OMG the Muskrats making constant excuses for Tesla is so pathetic and obnoxious. But what can you expect from people that paid money to be beta testers? The Jag seemed far more confident and safe, the Tesla feels like you’re teaching a kid how to drive and have to double pay attention just in case it does something unexpected.
The Tesla is a level two system and requires monitoring. which is good and bad. V12 has seen a huge leap in performance, and if it continues to make the progress it is making, Tesla are set to dominate in this space.
They are both self driving software suites. I’m simply showcasing the difference between Tesla’s approach and Waymo’s approach. If you dislike that Tesla is only Level 2 despite everything Elon tweets, than I think you should take your issue up with the big man, not me. I’d choose the Waymo over the stress of monitoring FSD any day.
@@AIAddict I think it is fair to showcase the difference between the two systems. I think it is also important to show the latest versions of both systems, as the latest version of FSD is showing remarkable promise. Waymo is good at what it does, but compared to what FSD is trying to do Waymo is very limited. From what I am seeing of V12, I think Tesla are onto something.
@@bogdanivchenko3723 Oh was it this guy? Didn't realise. However, it must be said his comments in this video were fair, if not up to date. Hopefully he will recognise and give credit if he gets a chance to try V12.
Waymo is extremely limited in where it can go and what it can do. Tesla FSD is basically trying to drive anywhere in America. From what I am seeing of V12. Tesla have the solution to self driving. It is just a matter of feeding it the data to finally see Tesla dominate in thias space.
I sure hope you’re right about the feeding the data. I hope they can solve it. V12 appears to be the first indication IMO of breaking the glass ceiling if true
