Cruise Robotaxis Are Back in Phoenix - But People Are Driving Them

Cruise is redeploying robotaxis in Phoenix after nearly five months of paused operations, the company said in a blog post. The catch? The cars will be in so-called "manual mode," so they won't be driving themselves. From a report: Cruise will resume manual driving of its autonomous vehicles to create maps and gather road information in certain cities, starting with Phoenix, the company said Tuesday. The General Motors subsidiary already had a presence in Phoenix before it pulled its entire U.S.-based fleet last year following an incident in San Francisco that left a pedestrian stuck under and dragged by a Cruise robotaxi. Prior to that incident, Cruise had been announcing launches in new cities -- including Dallas, Houston and Miami -- at a startling pace. Critics accused the company of expanding too fast and cutting corners on safety.

Autonomous driving is the new cold fusion

[memory_register]

Always just over the horizon.

Re:

[st33ld13hl]

Maybe, if Cold Fusion were a technology where we could start using it right NOW, get some amount of benefits from it (more power back than we put into it), and know that it will continue getting better and better over time.

For instance, I have a vehicle with Ford's Blue Cruise. It drives by itself, on some roads, like a 12-year-old, bouncing back and forth between the lane markings (and trying to drive off the freeway when the lane markers disappear). It constantly turns off when the road conditions change

Re: Autonomous driving is the new cold fusion

[Baloo Uriza]

You're putting a lot of trust in something that's roughly as capable as a video game NPC. You're going to kill someone and blame the car some day, and the only reason you'll avoid prison is because we very rarely prosecute deadly drivers in this country to start with.

Re:

[TWX]

The trouble is there are a lot of truly terrible human drivers on the roads as well, and those terrible drivers' issues driving end up manifesting when traffic is most congested. Sure, there are occasional single-vehicle accidents or accidents between vehicles in uncongested areas from time to time, but it's usually when tensions are highest due to the stress of traffic that bad human driving results in collisions.

What they're going to need to work out is how to handle situations when the 'correct' process

Re: Autonomous driving is the new cold fusion

[Baloo Uriza]

You speak for a greater need for investment in quality transportation infrastructure in the first place. The problem is a need for everyone to drive themselves. The solution is better public transportation, better urban land use, better and safer sidewalks and better and safer cycleways. This would also reduce the barrier to entry for transportation and give independence to children, the elderly, people with addiction problems, proven record of inability to operate machinery safely, the poor and the disa

Re:

[LostMyBeaver]

Wow! It seems to me that you feel very strongly about your thorough lack of understanding of this topic. Not only don't you understand it, but you also wish to force your beliefs on others.

Let's start by ignoring GM owned companies. They have never made any product worthy of purchase. They should be simply shut down.

Now, you make a great point. Now that NPCs are employing AI, they are rapidly improving. So much so that game makers will have to intentionally cripple them or human players will never be able t

Re:

[Baloo Uriza]

Not only don't you understand it, but you also wish to force your beliefs on others.

Project much? Opinion discarded.

Not Ready for Prime Time

[cob666]

In my opinion, autonomous vehicles are nowhere near prime time for general roadway use right now.

Re: Not Ready for Prime Time

[Baloo Uriza]

And they never will be. SAE level 2 is probably the likely maximum. Possibly 3. Anything past that is a crackpipe dream.

Re: Not Ready for Prime Time

[LindleyF]

Never say never. They don't have to be perfect, just safer than an average human. That's an achievable bar.

Re:

[fluffernutter]

They also need to never get a human into an an accident they otherwise wouldn't have gotten into. Bevause when you create a program you test it until it doesn't kill people. Anything less and that's what creates companies like Boeng. When all the self driving companies are rolling to take all responsibility for inquiring their cars, we will know when self driving is realistically safe. Because flying safely is a hell of a lot easier than driving and a lot easier to anticipate hazards.

Re: Not Ready for Prime Time

[LindleyF]

"Never" get into an accident isn't a practical standard. The politics of the situation dictate they need to keep the accident rate as low as possible, but zero isn't realistic. Pick some low percentile of human drivers, divide by 10 for politics, and aim for that.

Re:

[Baloo Uriza]

Vision Zero is a practical standard. Vision Zero is the goal. If your autonomous car kills people, you've created a killbot or a juggernaut, not a practical car. Drivers who crash and kill people belong in prison.

Re: Not Ready for Prime Time

[LindleyF]

It's a great goal. The problem is that it's not a good yard stick. Let's say autonomous vehicles have a safety rate 10x better than 99% of humans. That's really good. But what do you do if one incident happens-----do you take all autonomous vehicles off the road? Under this hypothetical replacing those with human drivers would make the roads more dangerous. I'm not saying any incident is acceptable or that we shouldn't keep trying to improve. But an all-or-nothing approach isn't practical.

Re:

[fluffernutter]

Hey, as long as the company making the AI gets prosecuted to the fullest extent of the law then I don't care, it will take care of itself.

Re:

[Baloo Uriza]

You know what's already available and safer than the average human driver? Professional human drivers, public transportation, real taxis, bicycles, and walking. Oddly enough, this also reduces the number of cars on the road, easing traffic congestion and making cities quieter and less polluted. Why not invest in that instead of flushing money down the toilet to let three companies sell more cars?

I don't think you know what that word means

[sdinfoserv]

Um... "autonomous" doesn't mean manually driven.

Why did they stop in the first place?

[quantaman]

Hiring a bunch of cabbies as safety drivers would be a pretty trivial expense compared to their R&D costs. And in addition to improved safety you eliminate all of those 'car blocks the street' PR disasters.

I can understand having the car handle tricky situations is better data than having the driver take over, but it seems like going driverless adds a big unnecessary risk.

Re:

[TWX]

To pay a living wage to a cab driver would cost around $75,000 per year accounting for payroll taxes, benefits, and other employer-born costs, if the drivers are employees of the taxi firm rather than true independent contractors that basically rent the cabs and then receive dispatch from the company as part of the cab rental.

If the cab company can do away with the drivers then that's a huge amount of money that they're giving up. That's why they're pushing for autonomous vehicles.

The whole point of reintr

Re:

[quantaman]

To pay a living wage to a cab driver would cost around $75,000 per year accounting for payroll taxes, benefits, and other employer-born costs, if the drivers are employees of the taxi firm rather than true independent contractors that basically rent the cabs and then receive dispatch from the company as part of the cab rental.

If the cab company can do away with the drivers then that's a huge amount of money that they're giving up. That's why they're pushing for autonomous vehicles.

The whole point of reintroducing the Cruise cars with human drivers is to get us used to seeing them operating again, where we're not instantly thinking of them as dangerous road hazards. Likely the intent is to try to shift back towards autonomous driving again, slowly as their developers actually get the software to work properly.

In September 2022 they were operating 100 taxis with a plan to go to 5000 [wikipedia.org]. Of course it's not 1 to 1, but if we do that then those 100 cabs would cost $7.5m, if they went to 5,000 is would be $375m.

Considering that GM cut $1B from their budget [reuters.com] the $7.5m would have been easily affordable, and even the hypothetical $375m would be manageable.

Money isn't the factor unless they have the world's dumbest CFO, there was a deliberate choice not to include safety drivers that probably had more to do with PR than cash

Re:

[Baloo Uriza]

Money isn't the factor unless they have the world's dumbest CFO

Carmakers on this continent rigged the system against being able to travel within your own city without a car to sell more cars, and then still are so incompetent they need the government to save their ass every few years. They seem to compete with airlines for this. Next time it happens, we should listen carefully to what they need, tally it up to the penny, then give the USDOT TIGER find ten times that amount and tell the carmakers to go fuck themselves. When the airlines do that, do the same but give

Re:

[Baloo Uriza]

Hopefully this time cities expect them to have taxi medallions and enforce it so you don't have the streets clogged with empty taxis anyway. Problem in Tulsa and Portland where Uber and Lyft send taxis out unlicensed and the city just doesn't enforce it very well.