Louis: 0:00

Hey there. Welcome to another Tesla Motors Club podcast. My name is Lewis.

Doug: 0:03

I'm Doug,

Mike: 0:04

And I'm like,

Louis: 0:05

In today's episode, we'll cover a number of topics such as the cyber taxi reveal was delayed. Elon gives us some insight into XAI versus Tesla AI. And FSD 12. 5 starts going wide to normal people, but not me, unfortunately. All that and more in episode 66 starts now. Gentlemen, we're back.

Doug: 0:27

we're back.

Louis: 0:28

while. Oh man. Normal people are getting it. No longer just the specials. I am so unspecial, I don't have it yet, even though there are other non specials that do. So, you know.

Doug: 0:40

Yeah. So it's been a little while.

Louis: 0:41

Been a

Doug: 0:42

I've been, outta town

Mike: 0:43

Globetrotter.

Louis: 0:44

Traveling the globe.

Doug: 0:45

Yeah. I had a, friend's wedding to go to in Sweden, , and I went there for about a week and then. Denmark for about a week, Then I went to the Netherlands and visited, @dpeilow , he used to be a, one of our, moderators on TMC. He's, , from the UK. And he works for the, European Space Agency. interesting guy. So I was in the Netherlands for about a week. And then, stopped in Iceland for three days. And, now I'm back. But I'm so discombobulated. It wasn't even just The different time zones is also, , up north and it's summertime close to the solstice. The days are so long, the sun doesn't set till like 11 PM and then it's still light out pretty much all night. And so it's a very, Discombobulating.

Louis: 1:23

time I was that north, I was in Denmark in winter. And so it was just dark and bleak everywhere. It's just like, there was no light. man, it was so cold. So at Least you went at a better time of year where it wasn't miserably cold.

Doug: 1:35

Oh yeah, it occasionally rained, but the weather was actually quite nice., nowadays you hear about people doing , cool cations or whatever. Basically, going to a place like Scandinavia during the summertime when, in Maryland, it was like 100 something degrees. And of course, in California, it was really

Mike: 1:51

it's been hot lately. Yes.

Doug: 1:52

Yeah. But it was nice, 60s, 70s, pretty nice weather. I was there. In Iceland, it was actually . wet and cold, like 50 degrees felt pretty cold when you're getting rained on,

Mike: 2:02

would happily take that right now.

Doug: 2:06

the neat thing I noticed just, from traveling is a lot of, Tesla's, a lot of EVs, the VW EVs and some other Chinese things I didn't quite recognize, but, lot of Teslas and, , a lot of Teslas as working vehicles,, Tesla taxis , , And even police cars. so that was neat to see.

Mike: 2:20

Before the show, you brought up street charging

Doug: 2:22

yeah, yeah.

Mike: 2:23

that's pretty interesting because from my HOA. I get nasty grams. If I'm charging in the street, they take real offense at it. So,

Doug: 2:29

in the Netherlands, I stayed for the most part in a smaller town, like maybe a half hour train ride to Amsterdam. sort of , classic, medieval town. Where you have these, buildings that are 400 years old. And, that country, and I noticed this in most of the Scandinavia, but, , particularly in the Netherlands, it's bikes are the main thing. Bikes are how people get around.

Louis: 2:47

love bikes.

, Doug: 2:48

sometimes the bike lanes are as big or bigger than the car lanes. Bikes get more preference to even, I think pedestrians,

Louis: 2:54

They do?

Doug: 2:55

you have to yield to them., but yeah, residentially you have, , curbside charging and, , as we talked before that the model in Europe for the type 2 connector is that, , you provide your own cable., so there are just these pedestals and you plug in and, I think right now where I was, they were free. They might cost more later, but this is in front of your house, like on your street, there'll be 1 or 2 of those.

Mike: 3:16

I'd like that.

Doug: 3:17

There can be the issue though, my friend was complaining, like, Oh, this guy who has a plug in hybrid, he plugs his car in and then just leaves it there, you know, that's like the worst thing, , but there are some more around the block., , part of the reason they, did that model is because. They had a mixture of competing, , plug standards, , cause they had the cars that were coming from Japan and the U S had what we call type one connector. It's, the Yasaki connector. That's company that was that form factor. Here, we call it the J 1772 plug, even though that's just the name of the IEEE standard.

Mike: 3:52

Good old J plug.

Doug: 3:53

right. So , the main thing about that is that's a single phase plug. but in Europe, , they're using three phase and three phase makes a lot more sense for their charging because they can charge, 11 kilowatts at, 16 amps, but it's three phase. 240 volts, three phase. and there was also a type three connector for a while that was, promoted by the French, they always, try to be different. that didn't add any functionality, except that it had shutters, , which is something that the safety conscious people, get up in arms about, like, oh, you shouldn't be able to touch the pins, , so there'll be like the mechanical shutters, what they call them getting in the way, but there's no point in that because there's already this pilot signal thing. There are already, , safety features built in to the J 1772 standard., so anyway, the pedestal would just And then if people have their own cable, then on one side is type two and on the other side, it could be type one or two or three, so that was part of the reason that people carry their own cable., and then, the other thing is that it prevents, vandalism.

Louis: 4:50

It's a good idea.

Mike: 4:51

Yeah. It's a really good idea. Actually, I'm thinking that'd be a great idea around some of the urban environments here.

Doug: 4:57

Yeah. So dude, this place I stayed at the Netherlands, , it was this town called Leiden, which has a lot of, physics history. Einstein spent a lot of time there. I saw the house that Lorentz lived in

Louis: 5:08

That's cool.

Doug: 5:09

and there's a, bar across the street from there. And he and Einstein would hang out there and discuss physics or whatever. Bohr and, , Einstein that would hang out there. So anyways, it was a cool place and it felt like kind of a paradise. I mean, , canals everywhere and people just hanging out laying by the canals bikes and EVs and green grass and and the temperature wasn't too hot, I had a nice time there. And then, I went to Iceland before I left one of the neat things about Iceland, it's right on the split between two tectonic plates, right? It's , the North American plate and the Eurasian plate that are separating there., there are other places like in California where the plates are, going into and underneath each other. Here it's, separating. So you get, magma coming out and, the whole country is increasing at the rate of about an inch a year, , but they have all this geothermal energy. So EVs are popular there because electricity dirt cheap , and they have the nice hot springs what was interesting to me, I stayed at this Airbnb there just on my own. And while it was nice, washing my hands. I'm like, what's that smell? and it smelled like, kind of egg smell. Yeah. And sometimes you get that, , if you have an electric, , water heater, there's a magnesium, , sacrificial anode in there. And when that starts to go bad, , your water can start to smell like eggs, so I was wondering if that was the issue, but then I realized, especially after. Doing a nice big tour around, some geysers and stuff, you would just smell that same smell like, oh, yeah, of course, this is the hot springs , and I found out, part of , their power system, at least in Reykjavik, they distribute hot water as a utility, you think about you get your water, but then you heat it and then you use it. No, they're just giving you hot water year round. You have two services. So you have the cold water, which is amazingly great. Glacier melt, so it's very clean. No chlorine in it. Just tastes great. You get it straight from the tap and it's super cold from the tap, and then you have this hot hot water. That, smells a bit like eggs, has a sulfur smell and it's really soft too. you're trying to rinse with that hot water. It just feels like your hands are Still soapy

Mike: 7:08

something there, yeah.

Doug: 7:09

Yeah, but that's what you shower in and it's pretty nice actually because basically you shower in this nice spa water, but yeah, when you're washing your hands, if you want to feel like they're not soapy end up having to rinse with the cold water and you're freezing your hands off, but otherwise it's pretty, it's pretty nice., so anyway,

Louis: 7:25

Interesting.

Doug: 7:26

some years ago, I think they try to do some hydrogen economy type stuff because they have surplus, , electricity and guess what? It didn't work. And if it didn't work there, why do we think it's going to work anywhere? You know, where, we have to , spend all this money on, um,

Louis: 7:41

or something.

Doug: 7:42

Yeah, it's not like solar and, wind., not like making those things is free, you know, it costs money to make that stuff., and if hydrogen, it's a third efficient,, roughly as just using batteries directly. Then, when you have the space for three times as many, solar panels or wind turbines, but now here you have an environment where there is surplus energy, , and still their hydrogen economy didn't work there. So why are you still pushing it? Anyway, that's my conspiracy theory about the hydrogen stuff.

Louis: 8:09

Yeah, For sure. And we've covered that in previous episodes so we should not dig into that again.

Doug: 8:13

No.

Louis: 8:14

Okay. Well, welcome back., sounds exciting. I definitely miss , my time there. I'm glad you made it through all the airport drama and everything else with all the stuff being shut down

Doug: 8:22

Yeah, I missed that

Louis: 8:24

incompetence.

Doug: 8:24

airport drama. So

Louis: 8:26

Excellent. So , the main thing we want to cover is recently they had the, Tesla earnings call

Doug: 8:31

for Q2

Louis: 8:32

for the investors and the big shots get to ask questions and get some feedback on

Doug: 8:38

Actually, at least the way Tesla does it. I mean they do have the big shots, but They do have a means for retail investors to ask questions., I forget the name of the service, but, people can basically put a question there and it gets voted up or down. And so Tesla is better about it being a bit more democratic, though. You could argue that, it's mostly Tesla fans that do that., and so Elon is more interested in their questions, , but often their questions are much more educated than, the analyst questions, though, feel like the analysts have gotten better. Certainly at the beginning when, people weren't even

Louis: 9:10

They didn't understand EVs at all or the industry. Yeah,

Doug: 9:13

Yeah, I didn't even understand any of these people

Louis: 9:15

right.

Doug: 9:15

questions, but these days I think they're pretty good.

Mike: 9:18

Now the analysts are driving Teslas.

Doug: 9:19

Quite possibly.

. Louis: 9:21

So I heard a few, good questions on the call and definitely felt like gain some insights out of it. I guess one of the biggest ones was, what do we call it now? The cyber taxi, the robo taxi, the cyber cab, the I've heard so many names.

Doug: 9:35

during the call. I think I heard Elon say Cyber taxi, which is the first time I heard him say that so does this thing actually have a name earlier? He said cyber cab out of nowhere

Mike: 9:46

CyberCab, and he also called it RoboTaxi.

Doug: 9:49

Robotaxi we've heard for years.

Louis: 9:51

That was the original name.

Doug: 9:52

Of course, the Robotaxi was supposed to be my model three, right?

Mike: 9:56

that was the original intent. Why you can't buy back your lease.

Doug: 10:00

whatever you call it, fact that the name changes is not really concerning, but it's two minds of it. It's like, okay, they don't really have a name for the thing yet.

Mike: 10:08

they don't have a thing yet. Mm hmm.

Louis: 10:13

Well, , like you said, he said cyber taxi , recently. And that was kind of our first time hearing that term. they also mentioned they pushed back the reveal. Two months, right? So it was from August to October., that almost makes it feel more real to me. In the fact that , one, they've changed the name for some reason. So they've come up with, here's a name for something. And two, it's not like, Oh yeah, by the end of this year, sometime we'll have it. It was like deliberately we're moving it by this much and it's a reasonably short time period. So I almost feel like maybe they're getting somewhere with this.

Doug: 10:48

the sense is that they actually have something to show. It's like, instead of a guy dancing around in a suit, which, you could do tomorrow, assuming you've already ordered the suit. when he said August 8th it sounded out of nowhere and it was probably the first time whatever engineers that are going to be responsible for getting this together heard it too,, August 8th is a, good date in Asia. It's, eight is a very auspicious number.

Louis: 11:10

Very lucky.

, Doug: 11:11

yeah, lucky extra lucky.

Mike: 11:12

not for the engineers that just heard it that day.

Doug: 11:14

maybe not, but then, yeah, pushing it back two months, okay. Two months isn't six months. Two months is like, we have something we need just a little bit more time.

Mike: 11:24

this is Tesla time though. So,

Louis: 11:26

it's a reveal. It's not a release, right? It's not.

Doug: 11:29

Right. But tends to be some kind of demo., so hopefully we see something reasonable, but,, Mike: Oh yeah. I mean, , obviously they're not ready., it's not going to be a real self driving vehicle, , unless it's heavily geo fenced or something.

Louis: 11:41

to be fair, Elon said at, some point during this call, he's been overly optimistic in the past, but he does feel like Full self driving will actually be working not necessarily by the end of this year, but next year, it is a realistic thing. he will be shocked if they don't have it next year.

Doug: 11:59

I'm sure he would have been shocked four years ago too, that it didn't have

Louis: 12:02

again, take it for what you will just, , he made a joke about him always being optimistic and being wrong in the past, but , he keeps talking about 12. 5 being this big, revolutionary improvement and, they feel like they're getting there.

Mike: 12:17

Let me ask you a question, Lewis. Do you think with the rebo taxi, cyber taxi, whatever they want to call it, they'll roll out hardware 5?

Louis: 12:24

So that's a good question., , it depends what it means, but yes, I do believe that hardware five will be in whatever vehicle they produce that is

Mike: 12:33

I could see that be the jump

Louis: 12:35

rare. Right. I, whatever vehicle they produce that is supposed to be self driving that isn't for you to buy. I do believe that'll be a hardware five vehicle. Yes.. Which, , hardware five is expected to be next year sometime, right?

Mike: 12:47

which gives FSD a fighting chance.

Doug: 12:50

We already had this discussion a bit in the previous episode, right? Because of the coincidental timing of. What he thought hardware five would come out and, , when we might get a cyber cab or whatever you're going to call it. I think that was our conclusion already, right?

Mike: 13:03

We'll see.

Louis: 13:03

yeah, I'm still standing by , the level, , five is, six years out at least or something like that. we won't actually have fully self driving anywhere kind of vehicles Till the end of the decade.

Doug: 13:14

the other thing that he reiterated. in this call, , basically staking the whole company on this. If, you don't believe Tesla's gonna achieve autonomy, you shouldn't be an investor.

Louis: 13:24

He did say that.

Doug: 13:25

The reality of that argument is that yes, if you really do have it, then, potentially the whole business model changes. And you're a super Uber, an Uber, Uber, if you will. and you're selling a service instead of a car, basically , what we said before. But the thing is right now you're a car company. Okay. And, , I would like them to concentrate a little more on being a car company. I mean, you can do both., you have enough cash to do both. But, , maybe keep the eye on the ball a little bit more. Okay., and consider customers, , , what customers actually want, obviously want lower prices., the Cybertruck is interesting as a Halo vehicle, but that's not, , what the majority of people want.

Mike: 14:02

It's not your bread and butter.

Doug: 14:03

Right. And even in the existing lineup, which is a great lineup, , having, , 3YSX, those are, good core, , Type vehicles hitting different segments of the market, but they need to be, regularly refreshed., and they do get, , minor changes and updates, , over time, but, there should be some kind of refresh every couple of years,

Louis: 14:23

Right.

Doug: 14:23

it shouldn't just be getting rid of things that people want, like there really should be.

Louis: 14:29

Stalks.

Doug: 14:30

Yeah, people want their stalks. And, um, know, I should mention, , there is a thread on TMC, by Richard C., And he says, customer choice is key to financial success., maybe, , but I, think there's something to what he's saying., and again, from our, point of view. And I'll say our, as a sort of collective thing in terms of the mission statement of TMC is really for , the customers and, , supporting existing customers. And, , customers want, turn signals,, what their stalks and , some might not want them. Okay, fine, but make it , an option. And if it costs money, charge money for it, we'll pay for it. People want it., and if you go through his, thread, he's Uh, some people want those things too,, I think probably a Tesla the first car I've ever owned that had ultrasonic sensors. I just used to, , typically drove very old, fun cars that didn't have those kind of safety features. I was lucky to have, airbags that, you know, staged airbags or whatever. But certainly user interface things like the turn signal, which is, right in the right spot, doesn't make sense to get rid of it. And I don't think you're saving anything by getting rid of it. And you are, that he points out, potentially losing a good number of sales by not having it. I mean, Mike, you had some comment on that, like your wife won't drive the Highland.

Mike: 15:38

Yeah, she absolutely refused to even get in the Highland. Won't have anything to do with it. And one of the reasons we bought the Y right now is this the last gen that we think will have stalks, the new, Juniper refresh supposedly won't have them.

Doug: 15:54

Yeah., Mike: She's just not into it and nor am I tend to think I could get used to it. But,

Louis: 16:03

I'm used to it. Honestly, I'm used to it, but the interface is not as convenient, like with turn signals again, especially in roundabouts and stuff. But it is what it is.

Doug: 16:12

And so your solution is just not the signal or what, in the roundabouts.

Louis: 16:17

honestly, my solution is. I pay a lot of attention when I'm in a roundabout to my steering wheel more so than I should.

Mike: 16:25

that's not a

Louis: 16:26

Realistically,, I am like hyper vigilant keep both hands on it, where normally I drive with one hand on the yoke, right? It's not even a wheel. but in a roundabout, I do two hands, and I am vigilant on it, and to make sure, I put my thumb where it needs to be before I even turn the wheel, so that I'm ready to go, because otherwise, it's like, while it's mid turn, when I'm like, you're not finding that. thing in the turn,

Mike: 16:50

Elon's counterpoint to that, Lewis, is you should be using FSD anyway, so why are you even driving?

Doug: 16:55

okay, is, FSD gonna safely get out of a roundabout

Louis: 16:59

I'll say is, where I live, roundabouts are usually in parking lots, Which is often where I'm going through them to which I cannot use FSD. I Only live about 20 minutes from the Tesla factory and headquarters in Texas, of course. So hopefully that's the thing that they address at some point

Doug: 17:15

don't get it. why can't you use FSD in the parking lot?

Louis: 17:17

because it's not mapped properly, right? Like self driving in a non road doesn't really work. Great.

Doug: 17:24

it works for me in parking lots, but I think they are mapped somehow. it knows where the lanes are. So.

Louis: 17:30

Gotcha.

Mike: 17:31

Mine won't work inn the Vons parking lot. It completely freaks out when it pulls into the driveway. Too many carts, cars, and people.

Doug: 17:37

Yeah, I guess it depends whether or not it's mapped. That used to be an issue with , summon,, when we first got, , summon, not just the moving forward and back, but I guess smart summon, , , I feel like it's been nerfed, but, , when we got it early days, like it was much more aggressive and it would do stuff., I think there was some hack where people were like, oh, it's not working well in my parking lot. And then people would go to Google maps and like actually add lanes and stuff, user submitted map. And then , somehow the Tesla would know the map. So it's using Google maps and it would actually use the map that people were putting on parking lots. I feel like a lot of parking lots got, mapped because of people using summon, it used to be a lot, more aggressive, but it also did stupid things to like, , I tried it and , Ran over a tree branch and was dragging a tree branch by the time I could see the car. I was like, oh,

Louis: 18:23

Sure,

Mike: 18:24

Oops,

Doug: 18:25

well,

Louis: 18:26

to be fair though, the other argument is I don't use FSD much if my wife's in the car because she doesn't feel comfortable with it. And so usually when I'm going to those places, she's with me. So I mostly use FSD when I'm by myself in the car or if I'm with friends, with my wife, I usually just drive because. as she hears the noise, she gets anxious.

Mike: 18:46

you say friends and the first thing that pops in my head is supervised driving. I'm going, I don't see those 2 being in the same sentence.

Louis: 18:54

Yeah. To be fair, when I'm doing that with others in the car, it's usually like a, Hey, let's look at this cool feature and we're focused on the self driving.

Mike: 19:02

little show and tell.

Louis: 19:04

Yeah, exactly.

Doug: 19:06

There was video that one of you guys shared, from reddit that had a car kind of freaking out, in the parking lot. Well, it's taken by someone who doesn't know, right? By some bystander and it's some car in a parking lot and it's moving on its own, but it's just kind of sitting there. It's, it looks like it wants to back up at one point, but it's just kind of stuck.

Mike: 19:26

That was Seb that donated the video..

Doug: 19:28

Yeah, okay. Well, thanks, Seb. basically it's just a video of a car. It's not even worth seeing really, but people are just like, What is it doing? There's nobody in there. But it's just, sitting there. I'm pretty sure that's somebody using summon the parking lot. But also the car isn't working very well. I use something when it first came out and I thought it was great. remember, , our previous hosts just talk about picking up his dates with it

Mike: 19:49

Yeah, yeah, it's a good party trick.

Doug: 19:54

when it first came out and it'd be raining and it would come pick me up. Actually for a while, it didn't work at all. Tesla said it was some known firmware issue. And I'm like, well, yeah, I'm not paying for this feature and, , , they didn't do anything to fix it for me. But after some software updates, eventually it started working again. But I only say working in quotes because It's just not great anymore. Like I'll find it, we'll just quit or is too slow to get started. And then, I'll be walking towards it and summoning it. And by the time I get there, it's pretty much still in the spot and it's just pulled out a little bit. And I was like, okay, what was the point? But it used to be really useful, especially if it was raining or you go to someplace like home Depot, for some reason, the entrance is always. like 100 yards from the exit, so you park near the entrance and then you come out and you got your stuff.

Mike: 20:38

Yeah, walk all the way back.

Doug: 20:39

Yeah. and it might be raining. So it'd be nice. Oh, summon the car. And I did that back in 20 21 with no problem. And it would just come to where I was, but now it's just way too hesitant., and it just ends up quitting but now you're blocking traffic in the parking lot.

Mike: 20:56

I gave up on smart summon when I couldn't get it to work in our work parking lot. the concrete that the structure is made out of apparently blocks cellular signal

Louis: 21:04

Oh,

Mike: 21:05

the car would either not respond at all or it'd roll like a foot and then say, I'm done. I can't go any further.

Doug: 21:12

When I first started using it was really impressive., I'd have it parked in some other part of the parking lot. It would come up. I'll be sitting there talking to somebody and just,, keep talking to him. And then the car pulls up.

Mike: 21:22

coasting up.

Doug: 21:23

cause it used to just come to you now you have to put a spot on the map for it to go to. Okay. Okay. It's nice to have that option too, but the come to me was, faster, , and, , yeah, it's just, nowadays I just don't find it usable. So I don't know what happened., maybe, too many accidents or rolling over curbs or something , and they just nerfed it to make it safer, but it's basically not usable now. So we're still waiting for that promise, right? The summon also what they called reverse summon, which is basically you getting out of the car and it go parking itself somewhere.

Louis: 21:52

that would be nice.

Mike: 21:52

would be nice. Yes. I agree.

Doug: 21:54

yeah. So when, when is that happening? Cause I paid for it.

Louis: 21:59

Yeah, well, , good luck., I guess we move on to some of the other insights and things that , we got out of the earnings call., there's some questions around, , XAI. And if you remember, there was some drama where it leaked out that Nvidia had a large shipment of, GPUs ready to go to Tesla, and they diverted them to XAI instead,

Doug: 22:18

Who's they? Elon

Louis: 22:19

basically Tesla chose to divert the cards to, , XAI instead of Tesla themselves using them., and so, an analyst asked about, the priorities here, you know, what are you doing?, and basically there's a few interesting things that came out of it. One, , specifically on the GPUs,, Elon said, well, look at the time. The data center was full. We literally had nowhere at Tesla to put in or use those cards so we could have received them and they would have sat there for, , weeks, months, whatever, not being used., so , it wasn't him prioritizing XAI over Tesla. It was simply Tesla can't use them now. So XAI can use them now. So XAI should get them.

Doug: 23:00

and he gave us that explanation before.

Louis: 23:02

Mm

Doug: 23:02

but as you mentioned it, , I'm wondering, well, who paid for them though?

Louis: 23:06

Exactly, right? and then the other argument is, okay, so when Tesla is ready for them, are they now then going to be waiting for cards? Because, He also spoke about how competitive and how difficult it is to get the high end, NVIDIA GPUs. Everybody wants them. Everybody wants them. They can't make them fast enough., they can't ship them out fast enough, which is why when Tesla had some, they gave them the XAI because XAI needed them so bad and Tesla couldn't use them yet., so he did mention that they're building out the data center and , they'll be turning on more things in the Tesla data center or something like that. So, okay, cool. Do you have video cards ready to go? He didn't get into that necessarily, but, it was some interesting insights into it.

Doug: 23:49

Let me just ask though, just thinking about it you're talking about it, so say Tesla already bought those cards and they can't use them., his argument was that it didn't make sense for Tesla to just stick them in a warehouse somewhere., I'm not sure that's exactly true, right? I mean, they'd have them now., I don't know how fast the next version or the better version comes out, but, , that's still an asset. Tesla, as an independent company, could have sold them to somebody else.

Louis: 24:16

have

Mike: 24:16

Probably made money at it.

Doug: 24:17

no, definitely. Right. if they're that high demand.

Louis: 24:21

we don't know. Maybe they did. Maybe XAI bought them from Tesla. We don't know that.

Doug: 24:26

Okay. that's probably not a competitive bid,

Louis: 24:28

They probably didn't pay as much as they could have

Doug: 24:31

Yeah. So just thinking about Tesla as an independent company and a board and a CEO that have a fiduciary responsibility to Tesla, not to super brand Elon, but to Tesla., does that make sense , that XAI would be able to just scoop them up?

Louis: 24:45

Probably not.

Doug: 24:46

it's, just feeling like, in some ways Tesla and I mean, I guess that's what we're getting to, , as we continue to talk about this topic, it's, it's competing a bit with XAI.

Louis: 24:55

hmm.

, Doug: 24:55

I think , you should be acting within Tesla's best interest,

Louis: 24:58

right.

Doug: 24:59

how that's Tesla's best interest.

Louis: 25:01

it's definitely, hard to know, right? We don't know the details. We just hear what gets leaked out to the media and then what Elon, how he responded to it. The interesting thing is, Elon did start talking about XAI researchers and, , the folks working there and Tesla and basically made a fairly compelling case of, look, Tesla is my priority. He tried to hire these folks for Tesla. He tried to get those , , machine learning researchers. He tried to hire these, very,, prestigious, well known, very skilled AI researchers for Tesla. he tried to do that and he could not get them to work there., in fact, , the reason XAI exists per Elon is because the only way he could get this caliber of. Individuals, or at least this group of individuals, not to say there aren't better individuals that he could have gotten separately., but he couldn't get them to work for Tesla. They wouldn't work for Tesla. They would only do it working for a startup and primarily focused on AGI., and so again, how Elon talked about it I gleaned from it is, , , two primary factors: one, Tesla is a big company,? And their valuation is one of the largest in the world, . , not to say that Tesla can't suddenly be worth a lot more, right? If Optimus comes out and they get self driving, right? Yeah, they could be way bigger. However, if you have a company that's worth, , 100, it's a lot easier to become worth a thousand times more than that. Then it is if you're already worth many billions of dollars or hundreds of billions of dollars to be worth a thousand times more than that. So as an engineer, right? I work in industry, not in auto, but in tech., when I work for startups. A big part of that potential pay package is I can make millions of dollars if I work at a smaller company that then does well and grows versus working at a big company that the rate of growth is limited., the flip side is big companies generally can pay more. It just depends.? , so anyway, so he tried to hire these individuals and one of the reasons they wouldn't do it is they want to work in a startup. They want to work in an environment where they have a potential, much bigger upside, right? Where they can turn around and get, , tens of millions of dollars or more out of it if it does well. The other thing, the way that he said it is they really want to work on AGI, right? So that's AGI is what we would traditionally think of as AI, like just a self aware, actual generalized AI, , concept. So like a human and a computer, if you want to think of it that way, to some degree., and so they want to work on that problem, and that is not the same problem as focusing on full self driving and autonomous driving and what Tesla is focused on. Now they're related, there's overlaps, they need the same GPUs, as we've made clear., and they need a lot of the same types of researchers and brilliant minds and people that work on those problems, but the problem itself, , the search space, or , the area of like, where you need to focus on for AGI is much bigger than the problem for FSD and self driving, not to say that's not a huge problem. It obviously is. there's overlap, but there are differences. And so to get those types of researchers, they don't want to work on self driving. So anyway, it's a very interesting insight at least from my perspective it sounds like Elon wanted to get these types of individuals, either these specific individuals or, , researchers of their caliber or , skillset. And he could not get them to go to Tesla and likely probably could not get them to go to X., but he could get them if he started a new company. Was focused on AGI and I feel like that's maybe a thing that he did specifically just to lock up this talent because it is a problem that Elon cares about AGI in general, right? It is a thing that he wants to influence. So anyway, that was just a very interesting insight that I got out of listening to his answers to these questions, , on the call.

Doug: 28:55

my sense was. XAI was really just a response to open AI, ? Open AI was getting a lot of traction in the public , , and Elon wanted to be there too., it's interesting though, , if we believe, okay, my assumption, and then if we believe what Elon was saying, that. He basically wanted to do that within Tesla.

Louis: 29:15

Mm hmm.

Doug: 29:15

I mean, it's interesting to think about like, if Tesla was doing a chat bot?, would that make sense if Tesla was working on a chat bot?, I suppose if you said, Oh, well, this is. Yeah, I was gonna say it could be part of what we're using for Optimus, or maybe, Tesla would have its own Siri type thing, where you could talk to Tesla , and, drive me to this place, whatever I would like, navigation to be a little bit smarter. In terms of when I'm trying to find something, because often , I'll put in, go to Safeway and then the Safeway gives me is 20 miles away instead of prioritizing the one that's right next to me, the other thing about it, though, and, talking about the NVIDIA GPU is also, I, we know that everybody is in this space, right?, all major companies, Apple, , Microsoft, of course, Facebook meta, , and they're all having to buy NVIDIA's, GPUs, ? And so obviously they're all gonna wanna make their own, ? Because they don't want to be a slave to Nvidia., Nvidia obviously is doing gangbusters right now,?, and of course Tesla is working And, , as we heard in the. 4th quarter earnings call, for the previous year for 2023, , the basic sense we got is that dojo wasn't working as well as hoped and that, actually, we do need to get those NVIDIA cards. and I wonder then, gee, , how far ahead is NVIDIA, ? Because they got all the money right now because they're doing so well. And so you would think, this is their bread and butter now, making these GPUs for this application. So you know that they're got to be putting tons of resources and making the better ones and continuing to be better and better and better, Now, Tesla's designed their own chips before. But they're not NVIDIA, right? They're not as focused in this , as NVIDIA is. So they're trying to make their own thing, Dojo, and that makes sense. And, The Tesla bulls are like, okay, well,, then Tesla could sell Dojo to, whatever Meta or they wouldn't, but, you know, unless they got the cage match out of the way or whatever. But, um, , the idea that, Oh, then Tesla can be a major competitor to NVIDIA, but I don't

Louis: 31:09

Sure.

Doug: 31:10

NVIDIA is probably so far ahead. can you give me some insight there?

Louis: 31:14

So to be fair, I agree with you, I thought it was very interesting. I learned a lot about, where they are feeling they are with dojo right now. But I'll preface and say, Tesla is not alone in trying to do this. So obviously Nvidia is the market leader and they've been the market leader for GPUs for many years, despite my AMD , family. I used to work at, global foundries. Used to make AMD stuff. AMD bought ATI, you know, the big two companies were ATI

Doug: 31:39

Right.

Louis: 31:40

and NVIDIA. NVIDIA has been the leader for GPUs For, a long time.

Doug: 31:44

But obviously Microsoft has got to be trying to make their own at this point.. Louis: But Google. has their own, right. It's called TPUs. tensor,

Louis: 31:51

Apple has their own, , built off of ARM. So they basically have, Apple Silicon. It's not exactly GPU, but there is a lot of GPU stuff in there. Intel has been working on it too, ? They're coming in from the CPU side, but they're essentially building GPUs in their chips., so it's, not like, Oh, , it's just Tesla trying to reinvent or come up with this new thing. And there's a lot of other companies too, just, I'm just naming a few of the big ones.

Doug: 32:15

right.

Louis: 32:16

what Tesla's trying to do isn't inherently wrong, but the challenge is, like you said, NVIDIA has the money and the expertise at a scale that nobody else does., so It's really hard to compete with them. Now , what I feel like a lot of the companies like, for example, what Google originally did with their TPUs and what Tesla was trying to do with dojo, , is that they're looking at it from, well, yes, NVIDIA is doing, you know, general purpose GPUs. And if I can specialize and focus on the thing that I really need, then I have a huge advantage. The problem is What's happened in machine learning and with AI, , development is that because of, , essentially tensors, , are tensors mathematically. And because of, how neural nets work and the modern approaches, how LLM, the architectures and GPTs and all that kind of work., with transformers, essentially that general purpose thing kind of works for everything anyway. And NVIDIA has also done a good job. I should specify in making specialized units on their GPUs. So , there are general purpose, but they also do have specialized cores that are really good at certain types of operations. So the bottom line is, , yes, in theory it's good. And I do think it's good that there are companies competing and putting money into it., but it's really freaking hard to beat NVIDIA. it's going to be hard for a while. Now, I don't think Nvidia will stay at the top forever. Every big company, when they get that current lead and they get to a certain point, they get lazy, they get complacent, they're kings of the world. And then another, somebody will come out of nowhere

Doug: 33:48

I would, say that the main thing that happens is , that innovators dilemma again, right? Some kind of change is going to happen. but they're focused in on the thing they're doing and that's still making the money. And then it's a risk for them to change because , there might be a new model, instead of transformers, it might be something else.

Louis: 34:04

Right,

Doug: 34:04

but for them to start doing the other thing, maybe that starts to invalidate what they're currently doing. And so, They want to protect what they have.

Louis: 34:10

They don't want to cannibalize themselves.

Doug: 34:12

Yeah, so my point is that, , everybody's trying to make their own thing. And somehow Nvidia is just,

Louis: 34:18

they're kings of the

Doug: 34:19

yeah, they're the king of the world., and Tesla, Tesla. They're not Intel, but they're

Louis: 34:23

right.

Doug: 34:24

of things.

Louis: 34:25

So, the interesting thing out of the call in his comments was that, , Elon is very optimistic very., pro, , Tesla and what they're capable of doing, obviously he's their CEO, he wants them to do well. So he's generally going to spin them in the most positive light you can, whenever a situation comes up. So it's very shocking to me for him to basically come out and say, Yeah, NVIDIA is really impressive. Their high end chips are really good. And pointed out that, Tesla's chips are not competing with NVIDIA's chips., not even close by the way in which he framed it and that they almost aren't even sure that they'll be able to, but that they still value Dojo and they still are looking at putting money into it and doubling down on it purely out of the fact that you can't get enough Nvidia chips because they're so competitive and everybody wants them. So it's almost like, well, this is these other chips that we can get more easily and use them because we can't get the better ones, that everybody wants.

Doug: 35:25

Huh.

Louis: 35:26

an interesting take, at least, , that's how I interpreted , his

Doug: 35:29

does that balance out cost wise, this is a, I guess a common issue cost of developing something in house versus buying it from somebody else. Buy it from somebody else. You get their markup and, whatever overheadd and their profit. But, doing it yourself, there's certainly some risk. Especially if you can't do it as well, and obviously they're not manufacturing themselves. It tended to TSMC in Taiwan.

Louis: 35:51

So that's actually the thing that helps in this space is the fact that none of them are making their chips are actually physically making them. So, including NVIDIA, none of them are actually making the chips TSMC predominantly is the company that manufactures them and how that generally works. Is companies like TSMC are improving their technology nodes. So all the work they do for Tesla and all the work they do for NVIDIA essentially helps everybody because they improve their process and how they're able to do things. And then what they do is they generally sell their technology nodes to these customers and go, look, I have this much capacity at two nanometer or that I can do this type of gate or whatever. Anyway, the improvement, the money that goes into NVIDIA and Tesla and everybody else is helping everyone., you're right though. They're still overhead. They have their own chip designers. They have their own, , hardware folks that are, , coming up with concepts and working on these problems and then also like building the data center and all the other, like, it's not just a chip, right? It's all the other things involved and all the software that goes into it. So there's definitely a cost for Tesla. I think it's good that they're willing to put research dollars into it., and they're developing things and it's good that they're improving and they're doubling down, but it's feels like it's a bet that didn't really pay off in the way in which they were hoping just because of how far ahead NVIDIA is for everyone. but , I don't think it's a bad thing for Tesla that they have dojo. And I think that, I'm happy to see that they're going to continue investing into it.

Mike: 37:18

The, ethos of Tesla has been vertical integration. We own everything top to bottom. So doing your own chip ties into that rather well. even if they don't do it as well right now, that doesn't say they can't do it well in the future, but they can certainly move a whole lot faster. If they have a change.

Louis: 37:38

It's definitely a potentially a better, longer play like a perfect parallel to this whole conversation is Apple. So, for example, Apple for years and years, they used power PC and they used IBM to make their chips and they had their own Apple chips and then they switched to Intel. And myself included, everyone was like, thank God. Thank you for moving to Intel because Intel was so far ahead and they were just so much better and suddenly Macs were great and things were working well and all this other kind of stuff. And then a few years ago, for those in the know, Apple switched to what's called Apple Silicon again, which is basically ARM chips.

Mike: 38:14

series,

Louis: 38:15

at first I was just like, no, why you're so stupid, Apple. Why would you do this? Except what we've learned after a few years is it was absolutely the right choice. Apple Silicon. I have, I literally have a Mac book next to me that is an Intel and I have a Mac book that is Apple Silicon. I have the right next, you know,

Mike: 38:37

is awesome.

Louis: 38:38

and Apple Silicon is killing it right now. Because the things that ARM allowed them to do and the things that they focused on in the design is all about power and efficiency and heat and all these other things. Which are really useful. So it, could pay off, , five years from now, 10 years from now, where suddenly it's like, Tesla is a genius. And it's so good that they had dojo and they were doing all this because it suddenly makes things way easier and cheaper to do what they're doing., as it did in my opinion, with Apple Silicon, we don't know, honestly, one of those things we can't know until it

Doug: 39:10

in the apple case, , Intel had stagnated, right?

Louis: 39:13

Exactly.

Doug: 39:14

Intel one of the few companies that actually does their own manufacturing and not getting the benefit of everybody else's, like TSMC does. So their. Shorter nanometer, , processes , were just stuck, And not getting anywhere., and then at the same time, apple had been already developing their own chips for the iPhone where, these are portable devices., , they have to be energy efficient and maybe not so high power in terms of compute, even though that was continuing to increase, and then along with the iPad too, uh, So, they basically plenty of practice and I was like, okay, let's scale this up into a proper processor where we can, have some thermals where, you can make it hotter and, , have a heat sink and whatnot, then you can start getting performance as well as efficiency. So, general sort of, you know, and I had this feeling when I went to this, uh, , AI conference, I guess a couple of years ago now, Intel, man, they, had a good

Louis: 40:06

They dropped the ball.

Doug: 40:07

they had a good run

Louis: 40:08

the ball. Yep.

Doug: 40:09

trying to go ARM now, Microsoft is on their way as well.

Louis: 40:14

Windows 11 already compiles for ARM.

Doug: 40:17

Linux has done it too,

Louis: 40:18

Linux is on ARM. Yep. Mac is on ARM.

Doug: 40:21

once your normal guy that buys his Windows machine, once that's on ARM, what's Intel doing? maybe Intel needs to do their own ARM and start paying those royalties.

. Louis: 40:31

is an interesting thing., I could go on and on, but I won't. and having, , worked in semiconductor, I do love the industry. So, it is possible, but as Doug did point out, there is a difference here where basically Apple moved off of somebody that was stagnating. And one of the advantages is that TSMC as the world leader foundry with bleeding edge, they could make things that, the previous vendor couldn't,, while TSMC is making Tesla's chips, they're also making NVIDIA's chips, so there isn't another pivot there that you're not getting another advantage by having your own chip from that perspective necessarily, but,

Doug: 41:06

not to get too off topic, but let me ask about AMD. Are they, still using GlobalFoundries or do they use TSMC as well?

Louis: 41:13

They use

Doug: 41:14

They use both, because the AMD model, another way that they were doing better than Intel is that they have that chiplet model, right? Which maybe can add some latency because you have like little components, but man, can that increase your yield?? If you make your chip smaller, And then, you just package them together, your little pieces together. Uh, just seems like pretty smart way to do things.

Louis: 41:36

for those that remember whatever it was 10 years ago ish, not quite when AMD suddenly stock went through the roof and things were going well, so they came out with these Zen architecture. I actually still worked in the industry at the time and was there when that happened., but yeah, so they basically came out with an innovative architecture for stuff that was super good on the server side and, low power device side. But yeah, we'll see where it goes. So dojo is not down and out yet, but it was eye opening to see that Elon acknowledged, how critical, , Nvidia,, GPUs are in, , developing, , AI. For both FSD and for AGI. And so both XAI and Tesla need those chips., and , it'll be interesting to see how that goes. I will say I was a little surprised because, , Tesla was talking about, Elon mentioned the data center and all those types of problems. I know they have deals with Oracle's cloud cause Larry Ellison was on the board of Tesla at some point and they have, , deals there. And I know Oracle has done a lot for buying GPUs. So I'm not sure which part of it is necessarily Tesla's dedicated data center versus just their cloud provider that they're using., but, we'll see , the, , development is there, the advancement is coming,, and related to that and what we 5, which has started to come out for some of the lucky, some of you out there that are luckier than me. and we'll hopefully start leveraging.

Doug: 42:57

think interesting about this 12. 5 rollout is that it's only gone to people with hardware 4.

Louis: 43:04

Yes, boo.

Doug: 43:06

I have hardware 3 in my car That makes us kind of sad because, hinted right at The annual shareholder meeting. That there's fork coming but I thought it would be a while from now And it's feeling like that fork is a little earlier since these are only rolling out on those hardware 4 cars now starts to make me feel like I'm getting left in the technology dustbin. and I saw some grumbling about it on Twitter., , Chuck, Mr. Left turn, , I saw him saying, well,, I'm buying a new Tesla and he bought a new, Model Y, because those are the first cars that are getting it. And he put it like he's specifically getting it so that he can be testing 12. 5. And what's interesting too is that, his left turn, like literally Tesla sends people out to do that left turn, I guess, because It's somehow special.

Louis: 43:54

it's an edge case.

Doug: 43:55

And so here he is buying another, Tesla specifically because 12. 5 is only available on that vehicle. And then I saw some other folks, , that were a long time Tesla fans, but that are starting to feel a bit burnt at this point, this one guy, said, well, , I've bought, several Teslas and I've bought FSD on it every time and, haven't gotten what I paid for, and I have to buy another car

Louis: 44:18

They keep offering to transfer my FSD, by the way. I've gotten another email, like not that long ago in the last few weeks. that was like transfer your FSD. If you buy another car. So

Doug: 44:28

I would hold out for it to actually work though, right?

Louis: 44:31

I'm very happy with my plaid despite the FSD where it is. but yes, I find it frustrating that it's only hardware 4 cars. I don't think they've worked it yet. I don't think we're going to be stuck we're hardware 3 is not going to get the next gen., if they did that, it would be pretty bad in my opinion of like the backlash that they would receive because Elon has repeatedly said hardware three will get it. It's just, it won't necessarily be as safe.

Doug: 44:56

what could be the reasons then for rolling it out this way?

Louis: 45:00

There's a few things. One, , it could be due to technical limitations, obviously.. That is what you're fearing., it could be the camera signals they need, the , higher res, better color space, that kind of stuff. Or the camera positioning,

Doug: 45:10

Or the slightly better processing,

Louis: 45:12

better processor. So there's more throughput, for his decision making that being said, I don't think that what they're doing so far is limited to that, you know,, personally, I don't think they're hardware limited yet. I don't think the processor limitation is really going to be the main factor there Again, I could be wrong, but from what I've read, a lot of the improvement in hardware 4 over hardware 3 is in the redundancy that they have. They kind of brought back having two different systems running in parallel. So they have redundancy of the decision making and it allows them to do more of the shadow testing. So for example, it's possible that this model is beefier and they want to do more shadow running in parallel of models to see what would happen. So that's why they want the better hardware that has the redundant systems and then once they're satisfied with how it's working and this version is what it looks like. And they don't need to do as much parallel testing on it. Then they'll it to the wider,, hardware three folks and others. That's what I think the primary reason is. Again, I have no insight into knowing that, that is accurate.. It might be just copium on my part because I have a hardware three, plaid. I'm expecting, 12. 5 to go out in the next, I want to say two months widely. Certainly this year, , but we'll see Elon seems to indicate that, he's been optimistic in the past but he said basically he will be surprised, like genuinely surprised that they don't have full self driving next year.

Doug: 46:41

Well, I'll be surprised if they do do it.

Mike: 46:43

and me

Louis: 46:43

Yeah, exactly. I'll be surprised that they do have it.

Doug: 46:46

I'll be happily surprised.

Louis: 46:47

but , he seems to indicate that 12. 5 is a revolutionary improvement over the others, which he often does., if you read Twitter Reddit TMC and all the other places where people post. It basically some people think it's the second coming. and some people think that it's, terrible and horrible and it sucks. So, as always, who knows?

Doug: 47:08

it's such a mixed bag in terms of the reactions of people that have used it., it does tend to correlate to, , how much of a fanboy people are., but, I've seen people saying, Oh, that's really great. But I've heard some people say , even 12. was a step back from 12. 3, and, , with 12. 5, the kind of quirks that you read about are like, When the car is sitting at a red light, it seems like it's impatient, like it starts inching up, when the other traffic has stopped moving, it's like it's anticipating , , the light to change, which is a very human thing to do,

Mike: 47:38

a very human thing to do.

Louis: 47:38

Right.

Doug: 47:39

but also pretty annoying if you're not the one doing it, so I'd want that, dialed out., Perhaps that's something that, showed up in the training.

Louis: 47:46

We'll see. To be fair, I have 12 4. I haven't used it enough to have enough of a difference of opinion for 12 3., nothing's jumped out at me of like, oh, this is terrible or, , oh, it's way better or way worse. but yeah, so 12. 5, that was supposed to be the big deal one. So according to Elon,

Mike: 48:04

I just put several hundred miles on 12. 3 and highway miles s. It's awesome. It's really well behaved highway. In town, it still sucks.

Louis: 48:15

well, you're going to likely be disappointed then because 12. 5, one of the big points that Elon said is that it's finally combining a lot more of the newer model to the highway and that the highway driving was much older and hasn't been updated in a long time. And so your city driving experience of the model is now going to be applied to highway.

Mike: 48:36

Awesome.

Louis: 48:37

Maybe it'll do better on the highway though. Right? It is a much simpler problem space. There are way fewer edge cases.

Mike: 48:43

the only complaint I have on the highway is it still tries to change lanes unless you manually override it.

Louis: 48:49

Hmm.

Mike: 48:50

And tell it, just be chill on the lane.

Louis: 48:52

Just cruise here. I'll let you know if I want you to change lanes,

Mike: 48:55

it still has troubles with highly reflective, tanker trucks. It doesn't like the light coming off the tankers. but those are kind of edge cases. Now the complaint my wife has when she's been in the car with me is it wants to shave like two or three minutes off trip by taking an off ramp, going down and getting right back up on the

Doug: 49:15

I see that all the time, man.

Mike: 49:18

it drives her nuts.

Louis: 49:19

it offers that all the time for me and I'm just like no

Doug: 49:22

Yeah. The funny thing is I think it's right though., like often I'm seeing that and I'm like, what the heck is it doing? And I'll override it. But then, , , traffic has slowed down and by getting off I'm passing all these cars that are all slow and then coming back in, it, it's a kind of a jerky thing to do because it's probably

Louis: 49:38

being a jerk

Doug: 49:39

It's probably causing more, slowing down as I'm trying to merge back in.

Mike: 49:42

would think it's more dangerous to try to merge back into traffic that second time. people being people.

Doug: 49:49

think that's a Tesla thing though. I think that's a Google thing. whatever the, traffic aware mapping, which I assume is just Google. Yeah.

Louis: 49:56

Well, no, I'll tell you because I do that with my phone a lot Google Maps does not do that Waze, I don't think does that But it's possible.

Doug: 50:03

That's a good question, though. How much of that is Tesla versus Google?

Louis: 50:07

Yeah

Doug: 50:08

well, talking about FSD. I'm still on 12. 3, which I found mostly to be an improvement. But, I've been away for almost a month, but then the month prior to that, I basically have had no FSD to use for a key reason. And I wonder how common this is, I've had an issue. With my, , cabin camera being unreliable. That's been driving me nuts., and I feel like it's somewhat software related. So while back, and I'm sure I mentioned this on a previous show that it basically was 50/ 50, every time I got in my car, whether or not the cabin camera was working and when the cabin camera doesn't work.

Mike: 50:48

No FSD.

Doug: 50:49

FSD doesn't work because it, wants to know that it can see

Mike: 50:53

We're watching

Doug: 50:53

And if you. taped over the camera, it won't let you. have FSD. Right. now The camera thing is driving me nuts because I use fsd quite regularly and when I don't have it I have to resort to autopilot and then end up missing turns because I forget that the car is not going to initiate the turn for me. So it gets pretty annoying. Then I guess some software update happened and intermittency went away and I was able to use it regularly. I was like okay great everything's working fine and then there's whatever software update I got and then it just stopped working altogether and every time I got in it wasn't working. It was so freaking annoying and I'm out of warranty now. I was like, well, I got to get this fixed because, , I want to be able to use FSD, Even though FSD isn't really FSD, I like what it is, I like using it over not using it., so I call up, the service thing on my phone and they want a hundred bucks to try to remotely diagnose it. that was somewhat disappointing. Like I was in communication with, , a service tech and they couldn't just look at the logs like, Oh, there are so many logs. Give me a time when it's not working. I'm like, okay. Well every time I get in the car, Okay, I give him a time and okay. Thanks Then they want like 800 this is starting to be real money and I'm going away. So I'm okay. Well, look, let's just postpone the service thing. They were going to send out a mobile service to deal with it. I was like, , I'm going to be out of town. Let's push it out. So I'm leaving. So packed my car up, I'm parked outside, but I put a cover on it when I'm away., and I turn off sentry mode because all the cameras are covered anyway and so then as I get ready to leave, I just want to check on cameras, just sort of see what they see. And then all of a sudden the cabin camera is working. Like what the heck. So the only difference that happened was I turned off sentry mode and then sentry mode had to be momentarily turned back on so that I could see live camera view the phone. And that's a feature that you get if you pay for, the conductivity package. Dude, all I had to do was turn off and on sentry mode.

Mike: 52:56

bucks, man

Doug: 52:57

The service guy, could have mentioned that, yeah. They wanted their 800 bucks., so anyway, I just tried it. At least I was able to confirm now that I'm back in town that it is working. So, , who knows with the next software update, hopefully it still works. But, , I just wonder how common an issue that is with the cabin camera I feel like in my era of car that the cabin camera is such a afterthought, a good part of the image is blocked by the mirror. and in the newer cars, I'm sure your car, Mike, even though it's not a juniper the cabin camera has done much better.

Mike: 53:27

it's pretty nice.

Doug: 53:29

it's nice better color space, obviously, but it also is more of a fisheye. It can actually see the steering wheel, for

Mike: 53:34

Yes, it can.

Doug: 53:35

so, , you wouldn't need these, torque measurements. You could just see if the hand is near the steering wheel. And that should be good enough.

Mike: 53:42

Well, unless Doug's driving with his one finger on the bottom.

Doug: 53:46

I suppose someone could put a fake plastic hand at the top.

Louis: 53:49

you, sound like Elon. Now, Doug, all I need is cameras and I can do everything. Get rid of all other sensors

Doug: 53:55

Well, let's just say , when autopilot first came out, there was no cabin camera.

Mike: 53:59

That's right.

Doug: 54:00

And, people like, , even Lex Friedman , would argue that there should have been some sort of driver monitoring back then. Of course, Elon thought, why do you need a monitor the driver? Cause the driver won't be driving because Elon was living in the future. Right. So at least they stuck it in there, but it , wasn't well executed. Now it's, pretty good., it could be better obviously it should be infrared and there should an IR emitter as well to illuminate you during the nighttime. I, saw on TMC, someone was complaining about FSD not working at night Because it couldn't see., there wasn't enough light in the cabin, so it assumed that the camera was occluded and FSD stopped working for him. That's kind of similar to the, errors I get because I'm driving in the middle of nowhere , and , there's no light, and so it thinks one of the outside cameras , is occluded when It just is too dark.

Louis: 54:44

To be fair, I would argue. Certainly at the current era of where the technology is. You probably shouldn't be using it when it's so dark out that the camera is not detecting anything. I'm just saying from a safety standpoint,

Doug: 54:57

Well, You have headlights I've concluded. It must be one of the side

Louis: 55:01

right.

Doug: 55:02

can't see anything and to be fair, you can have a better system that can tell the difference between something blocking the camera and nothing being there. It can have a little illuminator or something, right?

Louis: 55:13

Or you decide you don't need all cameras to do it,

Doug: 55:15

and often you don't, , you could just say, okay, don't allow it to change lanes or something. If you're that concerned,., you can see in front of you and autopilot

Mike: 55:24

And to your point, Doug, just checked it to make sure with the cabin camera on the Y, or at least my Y, you can see all the way to the bottom of the steering wheel.

Doug: 55:33

And with my error 2018, you can't

Mike: 55:36

there you

Doug: 55:36

go Anyway,

Mike: 55:37

Improvements.

Doug: 55:38

Improvements, it's just kind of sad , especially when you're out of warranty, but even while I was under warranty, anything or related, just felt like they weren't that. Eager to fix things for me. So

Louis: 55:51

At least I'm glad to hear your FSD is working. Hopefully we all get FSD 12. 5 soon. but more than likely Mike will get it before any of us and then he'll tell us that it sucks. and then we'll get it and we'll be like, Hey, this isn't that bad. Mike, what are you talking about? Cool. That's our show., thank you for hanging with us episode 66. I know we had a long break before this one. We will hopefully not have as long of a break for the next one. as always, thank you for your support and your comments and your likes and your dislikes and all the things you send our way. check out teslamotorsclub. com become a supporting member, , post things, interact with other users in the community. That's it for us. We'll see you next time.

Mike: 56:33

Peace.