Elon Musk Buys Twitter

[diller]

Why do we need self driving taxis in the first place?

[Jessicka]

Why do we need self driving taxis in the first place?

So Uber can keep surge charging you without having to pay staff.

- - - Updated - - -

Except nobody with any sense ever thought viable fusion was doable in any convenient timeframe.

Hell, hot fusion isn't even workable yet, with the current record of 48 seconds being set earlier this year. Cold fusion is still a pipe dream, especially since most theories for it require another as-yet-undiscovered advancement: room temperature superconductors.

Autonomous cars are at least feasible with current technology, and the rollout would be fairly quick once it's achieved.

But making optimistic promises about it is just dumb, regardless.

Just like AI, VR, AR…

[Belize]

Why do we need self driving taxis in the first place?

To keep the evils of mass transit away. America would literally explode if a single light rail line is built, and trust me bro another highway lane is finally going to solve traffic forever.

[PhaelixWW]

Just like AI, VR, AR…

Well, most things labeled "AI" are simply not.

But VR/AR are probably more apt analogues. 3D TV, too, among many others.

All are technologies that struggle with turning the corner from early adopter to mainstream acceptance due to a seemingly permanent half-realized status. They each have limitations that are honestly too costly and/or time consuming to overcome without significant refinement.

[Shadowferal]

That Las Vegas Loop...was that a success?

[Trifle]

That Las Vegas Loop...was that a success?

Probably depends how you define success. It's not a "hyperloop", it's just a tunnel that a few taxis drive down. It's probably a bit faster than using the road, but I can't imagine it's very cost effective.

[diller]

To keep the evils of mass transit away. America would literally explode if a single light rail line is built, and trust me bro another highway lane is finally going to solve traffic forever.

I don't even mean in the sense that there should be mass transit which there obviously should, I mean how is it an upgrade over a manned taxi, the only reason I can think of is that it will cost less for the big company providing the service since they don't have to pay drivers.

[PhaelixWW]

I don't even mean in the sense that there should be mass transit which there obviously should, I mean how is it an upgrade over a manned taxi, the only reason I can think of is that it will cost less for the big company providing the service since they don't have to pay drivers.

"But that savings will totally be passed on to the user!"

[Belize]

I don't even mean in the sense that there should be mass transit which there obviously should, I mean how is it an upgrade over a manned taxi, the only reason I can think of is that it will cost less for the big company providing the service since they don't have to pay drivers.

FWIW a taxi that doesn't hate you for using it would honestly be an upgrade.

Like, I'm sorry I'm interrupting your smoke, mr. Italian Taxi Dude, but like, you're in the front of the queue so I don't really have a choice.

[Edge-]

That Las Vegas Loop...was that a success?

If the goal was a one-lane tunnel for Tesla cars to be driven by real-ass humans, providing an incredibly inefficient method of transporting people between static hub stations compared to something like a rail system, then it's a smashing success, yes.

[Jessicka]

That Las Vegas Loop...was that a success?

Absolutely, it meant money and plans were diverted away from light rail and other public transportation projects. Exactly what it intended to do.

[Endus]

I am constantly amazed how people can still believe a word that comes of Ol' Musky's mouth when it comes to...well...generally anything...but specifically "fully autonomous cars".

In 2015, He said that Fully Autonomous cars were just two years away
In 2017, He said that Fully Autonomous cars were just two years away
In 2018, He said that Fully Autonomous cars were just 6 months away
In 2019, He said that Fully Autonomous cars were going to be on the road that year
in 2020, He said that Fully Autonomous cars were going to be on the road by the middle of the next year
In 2021, He said that Fully Autonomous cars were going to be on the road by the end of that year
In 2022, He said that Fully Autonomous cars were going to be on the road by the end of that year
In 2023, He said that Fully Autonomous cars were going to be on the road by the end of that year

Any time any techbro types says that any major paradigm shift is "only X time away!", what they really mean is "we can't do it, we have no idea how to do it, we need some kind of major breakthrough and we have literally no idea how to get there, but we still need investors so we're just gonna say it's in X time, which is a made-up number we pulled out of our ass to lie to you to get money. Also, this may be literally impossible but we'll never admit to that."

- - - Updated - - -

Well, most things labeled "AI" are simply not.

But VR/AR are probably more apt analogues. 3D TV, too, among many others.

All are technologies that struggle with turning the corner from early adopter to mainstream acceptance due to a seemingly permanent half-realized status. They each have limitations that are honestly too costly and/or time consuming to overcome without significant refinement.

To spin on the above point, the limit on VR/AR is primarily that, for widespread adoption as a paradigm shift, you need to be able to install this in, at a minimum, regular glasses, including prescription ones. Anything bigger and it's a toy you use for toy shit and that's about it. And I don't mean things like Google Glass, either, I mean seamlessly integrating over your natural lived experience. If not an actual neural implant that overlays the UI in your own visual center somehow.

Needless to say, there's wide gaps in innovation that need to be closed for that to be feasible, let alone affordable. And until it's feasible and affordable, it's going to remain a toy.

[Rasulis]

Any time any techbro types says that any major paradigm shift is "only X time away!", what they really mean is "we can't do it, we have no idea how to do it, we need some kind of major breakthrough and we have literally no idea how to get there, but we still need investors so we're just gonna say it's in X time, which is a made-up number we pulled out of our ass to lie to you to get money. Also, this may be literally impossible but we'll never admit to that."

- - - Updated - - -



To spin on the above point, the limit on VR/AR is primarily that, for widespread adoption as a paradigm shift, you need to be able to install this in, at a minimum, regular glasses, including prescription ones. Anything bigger and it's a toy you use for toy shit and that's about it. And I don't mean things like Google Glass, either, I mean seamlessly integrating over your natural lived experience. If not an actual neural implant that overlays the UI in your own visual center somehow.

Needless to say, there's wide gaps in innovation that need to be closed for that to be feasible, let alone affordable. And until it's feasible and affordable, it's going to remain a toy.

The hardest part with putting robotaxis on the streets is dealing with regulators. That part has tripped many companies including Uber which has pretty much given up on making their own autonomous vehicles. Currently we only have Waymo (backed by Google) and Zoox (backed by Amazon) offering rides in the US. I think Zoox will start taking rides this month. Not sure the exact date.

I suspect that Musk supported Trump hoping that Trump will be able to help him bypass all the regulations. Not only for his Tesla robotaxi, but also with SpaceX and his other endeavors. Otherwise, he won't catch up. Waymo is so far ahead now, it will take a miracle for anybody to catch up.

[Jessicka]

The hardest part with putting robotaxis on the streets is dealing with regulators. That part has tripped many companies including Uber which has pretty much given up on making their own autonomous vehicles. Currently we only have Waymo (backed by Google) and Zoox (backed by Amazon) offering rides in the US. I think Zoox will start taking rides this month. Not sure the exact date.

I suspect that Musk supported Trump hoping that Trump will be able to help him bypass all the regulations. Not only for his Tesla robotaxi, but also with SpaceX and his other endeavors. Otherwise, he won't catch up. Waymo is so far ahead now, it will take a miracle for anybody to catch up.

Hardest part will be contending with real life and kids just not setting them on fire or sending them into rivers. Didn’t we go through all this with undocked bike rentals?

I mean, them not getting stolen for parts is going to be a huge hurdle if there’s any kind of simple way to fuck with them.

[Rasulis]

Hardest part will be contending with real life and kids just not setting them on fire or sending them into rivers. Didn’t we go through all this with undocked bike rentals?

I mean, them not getting stolen for parts is going to be a huge hurdle if there’s any kind of simple way to fuck with them.

I think we are past that already. Parents are sending kids to schools by themselves using Waymo.

Arizona parents sending kids in Waymo cars alone against app policy

Parents’ hush-hush back-to-school hack: Sending their kids off in a Waymo

Yes. I personally have seen it done. Weird. They would not do that with Uber or Lyft. But Waymo is okay.

On the last part of your post. Waymo been aggressively suing people that vandalized or interfere with their cars in San Francisco.

[Jessicka]

I think we are past that already. Parents are sending kids to schools by themselves using Waymo.

Arizona parents sending kids in Waymo cars alone against app policy

Parents’ hush-hush back-to-school hack: Sending their kids off in a Waymo

Yes. I personally have seen it done. Weird. They would not do that with Uber or Lyft. But Waymo is okay.

I saw kids going to school on the old undocked bike rentals, I’ve also seen those bikes in rivers, up lamp posts, in bits strewn around streets, and still see a few about that were robbed.

Even with all the car security that exists today, car theft is still a massive problem. I just can’t see these things not being called somewhere and stripped down. And if there’s an opportunity for arson, again, they seem like a prime target. I’ve already seen plenty of videos and photos online of these things stopped and trapped with all manner of stupid tricks. And I really wouldn’t want to be in one that was.

There’s a clear race between security engineering and bored teenagers that bored teenagers will always always win.

[Rasulis]

I saw kids going to school on the old undocked bike rentals, I’ve also seen those bikes in rivers, up lamp posts, in bits strewn around streets, and still see a few about that were robbed.

Even with all the car security that exists today, car theft is still a massive problem. I just can’t see these things not being called somewhere and stripped down. And if there’s an opportunity for arson, again, they seem like a prime target. I’ve already seen plenty of videos and photos online of these things stopped and trapped with all manner of stupid tricks. And I really wouldn’t want to be in one that was.

There’s a clear race between security engineering and bored teenagers that bored teenagers will always always win.

Maybe. I do think robotaxi is here to stay. I heard Waymo is already testing their cars in Seattle and Miami. I am willing to bet that within 10 years robotaxi will be ubiquitous in most, if not all, US cities.

[Endus]

The hardest part with putting robotaxis on the streets is dealing with regulators. That part has tripped many companies including Uber which has pretty much given up on making their own autonomous vehicles. Currently we only have Waymo (backed by Google) and Zoox (backed by Amazon) offering rides in the US. I think Zoox will start taking rides this month. Not sure the exact date.

I suspect that Musk supported Trump hoping that Trump will be able to help him bypass all the regulations. Not only for his Tesla robotaxi, but also with SpaceX and his other endeavors. Otherwise, he won't catch up. Waymo is so far ahead now, it will take a miracle for anybody to catch up.

The other big barrier is that currently they require extensive mapping data for the specific location they're operating in. You can't just take a Zoox taxi in a city it's enabled in and plunk it down in a new city and have it work properly. It won't have sufficient mapping to operate safely. That mapping is expensive and needs constant updating. There's a potential that these companies could use those taxis themselves to constantly update the mapping data themselves, but as far as I'm aware, they're not currently doing so and it's a two-stage thing. But even then, you need starting data to operate from.

The current nearly-functional systems aren't able to operate independently; they need that mapping data. And it needs to be way more accurate than Google Maps.

[Jessicka]

Maybe. I do think robotaxi is here to stay. I heard Waymo is already testing their cars in Seattle and Miami. I am willing to bet that within 10 years robotaxi will be ubiquitous in most, if not all, US cities.

They might get around some inner city areas, where people being about is security in itself but in Europe those are becoming increasingly pedestrianised and things like light rail and pathed buses are increasingly common. I don’t expect it’ll ever be an option for commuting from suburbia over here at least. Maybe in the US where everything is built for vehicles and vehicular access they might catch on. But a proximity from main hub of use will almost certainly be a limiting factor.

[Rasulis]

Tesla is having major issue with its self-driving computer inside new cars

We have been investigating the problem, and based on insider sources and documents that we obtained, we can now explain the issue in more detail.

The problem is linked to a new version of Tesla’s HW4 (sometimes called AI4) onboard self-driving computer. Internally, some refer to the new version as AI4.1. The computers are short-circuiting. The cause is still being investigated, but one source told Electrek that one of the possible causes is the low-voltage battery short-circuiting the computer during the camera calibration process.

Tesla drivers are reporting computer failures after driving off with their brand-new cars over just the first few tens to hundreds of miles. Wide-ranging features powered by the computer, like active safety features, cameras, and even GPS, navigation, and range estimations, fail to work.

Insider sources told Electrek that the problem is quite wide-ranging. It affects vehicles built over the last few months with the new computer.

Two sources said that Tesla is currently receiving a high number of complaints about this issue, and it has yet to release a service bulletin about it. One source said that Tesla service is being told to play down any safety concerns related to this problem to avoid people believing their brand-new cars are not drivable.

The issue arose at the end of the year when Tesla is trying to deliver a record number of vehicles to avoid its whole year down in deliveries – a first in a decade.

It’s unclear if Tesla reported the issue to NHTSA despite the fact that the broken rear-view camera goes against federal safety regulations, which should force a recall.

At the moment, the main remedy being discussed is a computer replacement, but Tesla is also looking to push a software patch to help temporarily.

Tesla service is currently being overwhelmed by the issue, and Tesla is pushing service appointments to next year.