inevitable in retrospect
Anyway, I look forward to hearing about how Cruise was secretly 5 years ahead of everyone 10 years from now, as GM tends to do
I think an issue with robotaxis is that they'll encourage monopolies. The more taxis you have, the cheaper you can make rides because of scale. The more they're used, the more people will associate 1 company with quality and others with, well, [Wish.com](http://Wish.com) crashing sort of stuff like Cruise
Lol yup, literally the opposite. Worked for 4 different companies in that industry, 3 of them used lawn mowing to build a customer base and then axed it as soon as they could to exclusively offer more profitable services.
Diseconomy of scale is everywhere too.
This is why, for example, there are still a lot more than one car company. Big giant car companies (like, say, GM) have a lot of waste internally that smaller startups don't, and those smaller startups eat the big giant's lunch all the time.
Uh, cars? The worlds biggest manufacturer has a market share of... 10%.
Now, it's true that there's a *minimum* size at which you can't really compete with the majors (outside of niches like ultra-high end luxury/super cars). But, once you reach a couple million units/year, you can't really increase production without increasing costs approximately linearly with units sold.
It's not like cable TV or electricity where the costs are roughly fixed, and it's just a matter of how many customers you divide them by.
In which direction? Because as far as I’m aware Nike still has people lining up to buy $200 sneakers manufactured by Vietnamese middle schoolers for $5 each.
It is true that people love buying Temu slop. But the same people also have no issues dropping rent-tier money on shitty sneakers or luxury supplements.
Everything in marketing can be boiled down to the marketing mix, or the 4 Ps: product, promotion, place and price.
Monopolies tend to be formed by products that are commoditized by place and price (I.e taxies, fast food, groceries, etc) while products that emphasize promotion and product (luxury items, influencer brands, fashion, etc) tend to be more spread out over multiple companies and have much higher margins.
The very same consumer can choose dollar store items because of convenience (place oriented) but also spend hundreds of dollars on name brand fashion (product oriented); the consumer psyche has layers.
Yup. Some choose textiles, cars, shoes, jewelry whatever to splurge, then go promote monopolies for everything else.
We all watch amazon and dollar general decimate the competition. But nobody complains till they do what monopolies do and start price fixing with their leverage.
Look at what walmart did to vlassic. Walmart needed consumer support to pull that off.
Luxury items arent affected much by monopolies. Luxury shoppers are more resistant to the consequences of monopolies.
The current anti-trust enforcement action is weak.
In order to service the debt, robo taxis would have to drive faster than the current technology is safe to drive.
And when it rains and snows really hard, that's when they have the most problems.
And more wrecks happen.And it's also when taxis make the most money.
And also, the $/mile peak is in the high traffic areas. But ....the mph is so low that they don't really make THAT much money /day.
And it turns out that the tech is still to expensive. They can't drive fast enough to make sense on the highway. They can't do enough miles on the city where they are better to drive.
And enough uninsured drivers bonk into their semiconductor/Faberge egg based exoskeleton that their $/mile liability cost basis is through the roof.
Yeah.
It turns out driving risky, breaking laws, not yielding to pedestrians, and cutting people off are actually an important part of taxis being profitable, lol. More $/hr and $/mile.
Maybe, but I think the compute is going to get less expensive and there will be broader (non confidential) developments in self driving logic that will probably democratize the technology enough that there will be at least a few viable competitors.
More than likely (imo) car companies will simply partner with one of a few self driving hardware companies to make their versions of their cars compatible and self driving capable. Kinda like how Waymos are Jags or Chrysler vans.
Case in point, waymo is currently partnering with hyundai, mercedes partner is nvidia, BMW and Lyft are partnered with mobileye, we’re already seeing this
MobileEye and NVIDIA are computer vision chip companies. It‘s just one part at the very start of the in-vehicle ADAS/AD pipeline. There is a whole universe of additional complexity.
Waymo are the only credible player in this space who has a whole stack and is being public about that fact. They are keeping control of the business case themselves, though. So far there is no indication that they are going to license their system to those OEMs who manufacture the cars they modify.
SW is the difficult part. There are a handful of companies that have a SW stack and the whole training and evaluation environment required to get to 99.9999% reliability.
Which is why it will be unattainable for most OEMs for quite a while. Those who have made the effort will want to exploit their advantage, meaning that they will not share anything unless they get large license payments.
Or it will be something that people don't really care about who manufactured it. Do you know who made the subway car or bus you got on? I can't choose what plane I ride when I book a flight.
This is potentially the closest answer. I personally see more “ride sharing” rather than automated robotaxi just because people can be careless when anonymous.
Who cleans the robotaxis after every drive? Does a human just sit and get wheeled around 12 hours a day? What if it breaks down or you get in an accident? Then who’s liable? Insurance would be outrageous for any company.
> Anyway, I look forward to hearing about how Cruise was secretly 5 years ahead of everyone 10 years from now, as GM tends to do
It's gonna be a double or triple whammy when that happens since they effectively had three advanced ADAS/driverless car teams working in parallel with little resource sharing. Cruise, SuperCruise + regular ADAS, and SuperCruise's now-canned successor. Basically spending 2-3x what one consolidated team would cost.
They run out of patience. And Cruise also had lots of issues along the way. But it comes down to what is considered a core competency for GM? Certainly not software or fleet operations. They are integrator - they are god at defining specs for subsystems, having supplier build them, and integrating them. They are manufacturer - they are good at building millions of things. They have small mechanical engineering expertise. But they are not a software company. And they are not fleet operators.
I wonder when the eventual “out of business” announcement will be posted by Cruise. When Ford stopped funding Argo two years ago it was equivalent to death sentence.
while I tend to agree that self driving software is not a core competency for GM, it not really a core competency for anyone...
If you got money and initiative, anyone can do anything. GM had both.
Robotaxis are a flawed concept incongruent with the current state of our infrastructure. It would require a MASSIVE investment, probably including atleast 10s of billions for one city to upgrade its roadways and maintain them.
and at the end of the day... doug driving for uber earning $2.50 an hour after fuel and insurance is ultimately cheaper.
Now human is cheaper. Tomorrow that is not the case. Humans tend to have a knack for work hours and pay raises. Machines do not.
Meanwhile companies with software competency - Google, Tesla, Netflix, Apple, Amazon are pulling ahead in their industries and in the market caps.
>They are integrator - they are god at defining specs for subsystems, having supplier build them, and integrating them. They are manufacturer - they are good at building millions of things. They have small mechanical engineering expertise. But they are not a software company. And they are not fleet operators.
That's what all OEMs are, they only integrate and assemble bits and pieces together. An OEM uses about 100 different suppliers from around the world to assemble a car.
Only niche and exotic companies like Koenigsegesgesgesg are fully 100% vertically integrated. Even boutique builders like Pagani and Bugatti outsource major components like powertrain (Pagani with AMG and Bugatti with Cosworth).
I agree with you, OEMs GM, BMW, Ford etc need to stick with what they do best, manage the massive supply chain and build stuff and let the experts in software development do what they do best (mobileEye et al) and deliver a solid software stack to be integrated by the OEMs into their products.
Auto CEOs hate the idea of being stuck with an insanely capital-intensive business with poor ROIC and worse labor relations while all the tech/software guys get trillion dollar market caps.
I don’t think this is that simple. Core competencies needed by market change as market conditions change. Flexibility is rewarded, and rigidity is punished. They all need to learn software. Why? Because new wave of vertically integrated companies will eat them alive. Their vehicles become more software defined each generation. And system architects who understand that will design a better system. But more than that - system architect who has software developer in house can change software behavior much quicker. It makes for decent cost and time savings….. They do not need to be google level good. They just need to be better than Chinese competition.
> They are integrator - they are god at defining specs for subsystems, having supplier build them, and integrating them.
GM hates that this is the case because if true (it is) it implies they will get eaten alive by other integrators moving into the space from other countries or even industries, such as Foxconn's Lordstown plant.
Ford stopped funding and asked for outside investment that didn't pan out.
GM is planning to fold some of cruise into their L2 and L3 super cruise team to incrementally improve their product like tesla. That'll get you to L3 but GM is out of L4 and L5 self driving now officially.
Not even remotely a shocker.
Waymo will follow in 2-3 years. Will take them some time to admit they will never profit before they give up.
BUT WAYMO IS AHEAD OF TESLA.
Lol. Okay. If you say so. Their cars are crazy expensive, they can't make a lot of them, crazy investment needed for scale and to maintain HD maps, etc. They'll never be able to compete against a Tesla vehicle that costs $20k to make.
Prediction - Tesla will surpass Waymo in terms of square miles availability by the end of 2025. And any place where both are available, Tesla will be significantly faster to get from A to B, even once Waymo starts going on highways.
Tesla can't announce a single product without delaying it by multiple years half the time, and they can't even break level 2 self driving yet, meanwhile waymo is doing over a million level 4 miles a week. tesla would had to have already (magically) gotten their shit together to beat waymo by the end of 2025.
Sure, but 100m at Level 2 is still Level 2. You're required to be attentive and be ready to intervene with FSD. In a Waymo, I'm free to do anything, like sleep, from point a to b, and there is no control available to anyone in the car. And while that may seem like an obvious statement, the confidence and ability to make that jump is huge.
I guess I don't really understand your point. Are you saying from an achievement standpoint waymo is more impressive because they've done 25 rider-only miles and Tesla has done 0? Or from a data gathering standpoint?
I think from either perspective, if it's from an achievement standpoint that means as soon as Tesla releases an equivalent "level 4" release, they will surpass Waymo's distance traveled in about 3 days. From a data standpoint, I don't see a difference between level 2 or 4. The car is still the one driving itself and gathering data. Whether a human is reuiqred to be attentive or not doesn't affect the data gathered.
Tesla only does "easy miles" though. As soon as things get even slightly complicated, they hand things back over to the human.
It's easy to rack up big total miles numbers when you have a large install base and get to tally up the easiest N% of said base's miles.
Waymo will be surpassed this year, and they never make a profit, ever.
> Tesla can't announce a single product without delaying it by multiple years half the time
Model 3, Model Y? Lol. Ok.
Fsd can't even drive between the lines on my 2 mile drive to daycare lol. It can't see more than 100ft ahead it's always slamming on the brakes like an ass hole for every stop sign.
When you drive your EV so you use your brakes at all? I touch my brakes maybe once every ten miles. The Tesla is always slamming the brakes cause it is so slow to detect and react to things.
Always the question lol as if the small incremental changes make a difference. I have had two free trials. Latest wide releases. It's not good. It's not generalized. Only shitty drivers think FSD is good lol. You tell on yourself every time you say it's a smooth driver. Is it safer than the average road rage psycho watching tik Tok? Maybe.
I'm a av test engineer. Ive ride in or personally know test engineers at just about every AV company or OEM releasing driver assist. Tesla FSD is a joke and the only people that think it's good don't understand much about driving behaviors or safety.
Whats a joke is car that you can only make 1000x a year of that costs 100k+ per car, requires HD maps, and is geofenced with hard coded logic. That's a fucking joke
Cruise was supposed be operated like a Silicon Valley startup, that's why it was a separate entity, and employees were receiving Cruise "stock" rather than GM stock.
Didn't work out so well.
GM just doesn't have the cash like Google or Amazon to fund the company until it can reach critical mass. They are a capital intensive manufacturing company that doesn't have a lot of margin to play with.
Yep, from what I've heard from friends who used to work there, they were operated like a true SV startup right down to the horrific working hours and toxic management practices. Turns out that, indeed, "move fast and break things" isn't a good idea when your company is building safety critical software, as opposed to say serving thirst traps to incels. Waymo operated like the adults in the room and that's why they're winning.
It feels like Ford funded Argo just enough for them to develop BlueCruise (which is actually pretty good, better that Leon Skum's shit "FDS") and once they had the program fully deployed, killed Argo since there was no further use for them.
Still no, Ford developed BlueCruise in house. BlueCruise launched in 2021. Argo fully developed their own SDS and folded late 2022. Latitude absorbed parts of Argo’s IP after Ford created the wholly-owned subsidiary. From that point Latitude efforts began to contribute to BlueCruise.
Good. Cruise sucked. I saw their shit ass cars stuck in SF all the time blocking up traffic, getting cones out on them, and a variety of other annoying stuff, none of which I’ve seen with Waymo
Yeah cruise had such a higher appetite for risk both not at fault collision risk as well as progress risk. Waymo was much more cautious about progress risk aka being a stone in the river like an idiot.
I'm fortunate enough to be employed and have a retirement fund, so in a roundabout way, I guess so. Not sure what that has to do with Waymo though.
Whatever plans they have with Hyundai are just plans at the moment.
Wut? Waymo is just diversifying their fleet because the Zeekyr they custom designed is now twice the price thanks to EV tariffs...
Waymo has had so many vehicles platforms at this point, two of which have gone fully driverless. They can take the ioniq5 driverless once the manufacturing meets requirements.
I'm a regular Waymo user here in Arizona, and I had been wondering why I haven't seen Cruise cars driving around lately. They've been testing them for years, but never started actually carrying passengers as far as I could tell. Meanwhile Waymo has improved *dramatically* since my first ride 6 years ago.
Honestly Cruise served its purpose as development hotbed that enabled the true revenue generating software “Supercruise” once GM is able to scale it. Excellent investment, additionally the company retains most of the talent onto further development of Supercruise.
75 Comments
stav_and_nick@reddit
gumol@reddit (OP)
MSTmatt@reddit
bestselfnice@reddit
rugbyj@reddit
lee1026@reddit
Striking-Bluejay-349@reddit
stupidzoidberg@reddit
rockomeyers@reddit
roman_maverik@reddit
rockomeyers@reddit
massada@reddit
triplevanos@reddit
hi_im_bored13@reddit
MachKeinDramaLlama@reddit
hi_im_bored13@reddit
MachKeinDramaLlama@reddit
ZeroWashu@reddit
lostboyz@reddit
PontiacMotorCompany@reddit
Legend13CNS@reddit
ARHANGEL123@reddit
Chrodesk@reddit
ARHANGEL123@reddit
stupidzoidberg@reddit
Salty-Dog-9398@reddit
ARHANGEL123@reddit
ARHANGEL123@reddit
Salty-Dog-9398@reddit
LibatiousLlama@reddit
Shmokesshweed@reddit
grizzly_teddy@reddit
EndlessGame8161@reddit
mocoyne@reddit
EndlessGame8161@reddit
mocoyne@reddit
A_Right_Proper_Lad@reddit
grizzly_teddy@reddit
LibatiousLlama@reddit
grizzly_teddy@reddit
LibatiousLlama@reddit
grizzly_teddy@reddit
LibatiousLlama@reddit
grizzly_teddy@reddit
Imperial_Eggroll@reddit
gumol@reddit (OP)
clownpirate@reddit
AcrobaticNetwork62@reddit
clownpirate@reddit
movingtobay2019@reddit
clownpirate@reddit
Drzhivago138@reddit
piggybank21@reddit
TechnicianExtreme200@reddit
TechnicianExtreme200@reddit
e___r___s@reddit
stupidzoidberg@reddit
e___r___s@reddit
stupidzoidberg@reddit
e___r___s@reddit
e___r___s@reddit
gumol@reddit (OP)
e___r___s@reddit
nerdpox@reddit
KingKontinuum@reddit
LibatiousLlama@reddit
akkawwakka@reddit
Ancient_Persimmon@reddit
TimeTravelingChris@reddit
Ancient_Persimmon@reddit
LibatiousLlama@reddit
SassanZZ@reddit
4a4a@reddit
PontiacMotorCompany@reddit
TechnicianExtreme200@reddit