Can’t you do that already just by getting a normal taxi? And the SFO to BART connection is pretty good if your final destination is anywhere near BART or MUNI.
Can’t you do that already just by getting a normal taxi? And the SFO to BART connection is pretty good if your final destination is anywhere near BART or MUNI.
The SFO taxis are even worse than the ubers. From the frying pan into the fire. I mostly go for work or to visit friends these days, so I am nowhere near market street or mission.
The SFO taxis are even worse than the ubers. From the frying pan into the fire.
And the SFO to BART connection is pretty good if your final destination is anywhere near BART
I mostly go for work or to visit friends these days, so I am nowhere near market street or mission.
Which is part of the reason why Uber and Lyft were able to get off the ground in the first place. Lower prices were a big factor too, but if the taxi experience had been good beforehand, neither...
The SFO taxis are even worse than the ubers. From the frying pan into the fire.
Which is part of the reason why Uber and Lyft were able to get off the ground in the first place. Lower prices were a big factor too, but if the taxi experience had been good beforehand, neither would’ve grown nearly as quickly.
Ugh, the first time I got an airport taxi in Las Vegas, it was August and well over 100F/37C. The car had no air conditioning and I found out when I arrived that it was cash only. I was VERY hair...
Ugh, the first time I got an airport taxi in Las Vegas, it was August and well over 100F/37C. The car had no air conditioning and I found out when I arrived that it was cash only. I was VERY hair to use a ride share there when they became available.
Santa Clara County will host several major sporting events next year, including Super Bowl LX and the FIFA World Cup, both at Levi’s Stadium.
SJC will be the second international airport in the U.S. — and the first in California — to offer a commercial robotaxi service. In 2022, Waymo began providing service to Phoenix Sky Harbor’s Sky Train stations; that was expanded to 24/7 curbside service last August.
Waymo conducted mapping and testing at SJC in August and September 2024, with safety drivers present in the vehicles.
Waymo is not yet available to the public in San José. The company has gradually expanded its zone of operation this year from San Francisco to the Peninsula and South Bay.
"Car operates as expected." Would you read that news article? Neither would anyone else. I'm not actually saying you should ride in a self-driving vehicle, but consider that frequency with which...
"Car operates as expected." Would you read that news article? Neither would anyone else. I'm not actually saying you should ride in a self-driving vehicle, but consider that frequency with which something is reported on is a bad metric to use for actual risk. More than 40,000 people die each year in regular car crashes in the USA. It's not more safe just because it isn't news.
That's fine, but wasn't my point. You do you, but think about how the media chooses news articles when it comes to reporting novel failure over regular failure.
That's fine, but wasn't my point. You do you, but think about how the media chooses news articles when it comes to reporting novel failure over regular failure.
It is funny how so many people on social media can not accept that someone else can understand their argument and not agree with it at the same time. My point, which I will repeat, is that based...
It is funny how so many people on social media can not accept that someone else can understand their argument and not agree with it at the same time.
My point, which I will repeat, is that based on my own experience and what I've read is that I do not trust new automation technology.
If you understood my argument, perhaps you might have said something at any point that indicated that understanding as opposed to just stating a different, unrelated point? It truly looked like...
If you understood my argument, perhaps you might have said something at any point that indicated that understanding as opposed to just stating a different, unrelated point? It truly looked like you didn't even register what I said. I apologize for reiterating my point.
I did understand your argument and I did not agree with it. I'm sorry those two things coexisting seem inconceivable to you. No disrespect meant. Have a nice weekend.
I did understand your argument and I did not agree with it.
I'm sorry those two things coexisting seem inconceivable to you.
You understood my argument that it's reasonable to have media literacy and think about what they have the incentive to publish. And you disagree with it? I'm fascinated. Can you go into that a...
You understood my argument that it's reasonable to have media literacy and think about what they have the incentive to publish. And you disagree with it? I'm fascinated. Can you go into that a little more?
As someone else who does a lot of automation, no. I do not. You know who wrote that automation that’s driving your cars? Human beings. Worse in my eyes, ones that seem to benefit from claiming all...
As someone else who does a lot of automation, no. I do not.
You know who wrote that automation that’s driving your cars? Human beings. Worse in my eyes, ones that seem to benefit from claiming all their stuff is perfect to the tune of billions.
I think Waymo is miles better than Tesla, but the only thing worse than a single human driver in my eyes is one owned by a corporation running on code and tools built by it
As someone who does a lot of automation, isn't it pretty apparent that humans are highly fallible and generally bad at things? Not just in general, but in this specific case:...
As someone who does a lot of automation, isn't it pretty apparent that humans are highly fallible and generally bad at things? Not just in general, but in this specific case:
About as apparent as AI being generally awful and bad at things. Theres no reason to think a car, which is not executing a constant set of instructions but instead required to adapt to its...
About as apparent as AI being generally awful and bad at things.
Theres no reason to think a car, which is not executing a constant set of instructions but instead required to adapt to its environment, will be any better
See the fact you think such evidence exists is concerning. The VERY small amount of data we have is under very specific conditions and absolutely cannot be used as an apples to apples to human...
See the fact you think such evidence exists is concerning. The VERY small amount of data we have is under very specific conditions and absolutely cannot be used as an apples to apples to human drivers in all situations. 99% of it is also solely provided by developer and there’s fuck all for audits and standards. It’s somewhat akin to pointing to the low conviction rate for cops and saying they must be fine because the data says so.
Show me safety records for automated cars on the highway in the snow? A useful comparison point because one area where humans are absolutely weak at. How about rain on badly maintained roads? How are they handling construction?
Right now most of the data shows what you’d expect. Machines don’t get tired, don’t get drunk, and don’t push it. They’re also limited to the conditions still where they’re at their best. Namely highly important pilot programs.
How does a Waymo that hasn’t been properly maintained for 2-10 years perform? Because most cars, busses, and trains in the US are hardly held to high standards.
What company is spear heading the call for Congress to pass laws mandating repair of this VERY expensive tech before it’s on the road. How detailed are the standards provided? Are we now going to think that these corporations will be benevolent instead of cost cutting?
We’re at the very begging still of programs like this and people act like there’s a mountain of independent and reproduced evidence. Make insurance companies liable and start building actuary tables and I’ll start to care because at least another corporation will have to make their money off of the data.
Self reported highly controlled test cases are not anywhere close to scientifically rigorous. If Waymo wasn’t at least trying to do parts of this right I’d be even more harsh because they’re basically the only ones in the passenger sector worth a damn, and if you gave them the green light tomorrow I’d bet it’d be a Google situation in 5-10 years where everyone wonders why the multibillion dollar corporation didn’t turn out benevolent.
This is of course a big part of why they have a good safety record. Waymo can just decide not to drive in places or under conditions they aren’t ready to tackle. (Such as driving on the highway.)...
They’re also limited to the conditions still where they’re at their best. Namely highly important pilot programs.
This is of course a big part of why they have a good safety record. Waymo can just decide not to drive in places or under conditions they aren’t ready to tackle. (Such as driving on the highway.) This isn’t a pilot program, it’s how they operate in cities where they’ve been in production for some time. They’re not going to drive in the snow until they’re ready. When they do decide they’re ready, that’s a sign that it will probably just work.
How does a Waymo that hasn’t been properly maintained for 2-10 years perform?
This is asking for speculation about a system that doesn’t exist. There are no signs that Waymo is going to sell self-driving cars or otherwise allow this to happen.
What company is spear heading the call for Congress to pass laws mandating repair of this VERY expensive tech before it’s on the road. How detailed are the standards provided? Are we now going to think that these corporations will be benevolent instead of cost cutting?
For Waymo at least, there is no sign yet that this is necessary. As long as they continue to have a good track record, there’s unlikely to be a push to solve a non-problem based on speculation. What incentive do company managers or legislators have to mess with success? It might just make things worse.
Make insurance companies liable and start building actuary tables and I’ll start to care because at least another corporation will have to make their money off of the data.
This would be a weak incentive at best and stronger ones are sometimes ignored. As we saw with Cruise, bad accidents (and a coverup) can cause the entire effort to be cancelled. They had external incentives not to behave like that, but they did it anyway, because internal company culture somehow wasn’t right.
That’s a pretty good incentive to make sure it doesn’t happen, if you can see it. Whether this cautionary tale is heeded at another company depends on that company’s culture.
So far it seems that Waymo is not rushing things. It’s true that it might not last. Can they preserve a safety culture as they scale up? Hard to say, but the threat of bad press and more government regulation is incentive to try.
Contrast with Tesla, which has a CEO constantly pushing employees to move faster than they believed was possible. The difference in safety culture is dramatic, and it comes from the top. It’s gotten people killed. I do think governments should be doing more about their nonsense, but the US government has bigger problems.
The data we have is under the conditions that Waymo actually operates in. Why are we talking about snowstorms in San Francisco, Phoenix, or Los Angeles? Yeah, maybe we should be hesitant about...
The data we have is under the conditions that Waymo actually operates in. Why are we talking about snowstorms in San Francisco, Phoenix, or Los Angeles? Yeah, maybe we should be hesitant about taking a Waymo in a place that Waymo does not operate but that seems like a bit of an irrelevant concern.
Right now most of the data shows what you’d expect. Machines don’t get tired, don’t get drunk, and don’t push it. They’re also limited to the conditions still where they’re at their best. Namely highly important pilot programs.
This sounds like an argument for using Waymo today.
The proper, honest spin would be "study shows Waymo has 99.9% safety rate, 300% safer than human driver" (you know, assuming studies are being done). It's not eye popping, but it is exciting the...
Would you read that news article?
The proper, honest spin would be "study shows Waymo has 99.9% safety rate, 300% safer than human driver" (you know, assuming studies are being done).
It's not eye popping, but it is exciting the day we can see that headline. If they can't show that, or at least properly show who's accountable for a self driving service, I won't trust it.
Waymo is releasing new research that shows its fleet of driverless cars are dramatically better than humans at avoiding a wide range of collisions on public roads.
According to the peer-reviewed paper, which is set to be published in the Traffic Injury Prevention Journal, Waymo’s self-driving cars experienced 82% fewer crashes with bikers and motorcyclists compared to human drivers, and 92% fewer crashes with pedestrians.
Do you ride on airplanes? Most of the time they’re in the air they are on autopilot. Like flying, robotaxis require giving up an understanding of what’s going on. You don’t know what happens in...
Do you ride on airplanes? Most of the time they’re in the air they are on autopilot. Like flying, robotaxis require giving up an understanding of what’s going on. You don’t know what happens in the 747’s cockpit and you wouldn’t comprehend it even if you could see. Same for a robotaxi. It’s largely a black box. But in both cases the injuries per mile are much lower than driving yourself.
It’s important in the modern world to be able to trust engineers most of the time. Trust the civil engineers that built the bridges you cross. Trust the engineers that built the skyscrapers you walk within or beneath. There are always examples of catastrophic failure. But you’ll oversample on freak accidents if you don’t make an effort to avoid the natural bias.
Planes were made as safe as they are through extremely stringent regulation and strong government oversight. While that has been eroded lately, the baseline level of safety in aviation is still...
Planes were made as safe as they are through extremely stringent regulation and strong government oversight. While that has been eroded lately, the baseline level of safety in aviation is still quite high, especially for the older airframes. Even when they operate in autopilot, these systems are also overseen by two highly trained pilots and further supported by an infrastructure of ground support, even to being able to troubleshoot problems in flight with the engineers who built the system. Furthermore, the autopilot basically flies in an empty sky (getting within several hundred feet of another aircraft is already considered a near miss). Aviation autopilot systems are vastly simpler and have a vastly simpler task to perform than even the simplest AV use cases.
The US automotive market is almost completely governed by voluntary self-regulation. Compliance with the major software safety standards (ISO 26262, ISO 21448, UL 4600) is voluntary and there is no government body that oversees it. The only mandatory standard is FMVSS, and it only mandates certain items on the car, not the correctness of their function. NHTSA can only act after the fact through recalls, and even there, their ability is very limited. The only other control on the industry is lawsuits, which, unfortunately require someone (or many people) to die or have serious injuries before they can move forward. By the time they are settled, the models in question might be a decade old.
There may be reasons to accept riding in an AV taxi, but it's not the same as flying by any stretch of the imagination.
I think it's important to make a distinction between Waymo and the other attempts at making self-driving car services. Lumping together Waymo and Tesla's nascent effort wouldn't make sense since...
I think it's important to make a distinction between Waymo and the other attempts at making self-driving car services. Lumping together Waymo and Tesla's nascent effort wouldn't make sense since they're so different.
It takes a long time to collect enough data to prove it, but Waymo seems to be headed towards airline-level safety.
Planes literally must tell every other plane where they are and plan their route with multiple governments in advance. It’s not remotely apples to apples and still has 2 humans at all times for...
Planes literally must tell every other plane where they are and plan their route with multiple governments in advance. It’s not remotely apples to apples and still has 2 humans at all times for emergencies
And lacks any form of autopilot like was being described in something like 99% of cases. And you still file a flighplane and coordinate with an airport and have a traffic controller you interact with.
And lacks any form of autopilot like was being described in something like 99% of cases.
And you still file a flighplane and coordinate with an airport and have a traffic controller you interact with.
Cessnas do often have autopilots, but this is more like cruise control in a car - it holds altitude and heading. The more modern cockpits have air collision avoidance systems, but when flying...
Cessnas do often have autopilots, but this is more like cruise control in a car - it holds altitude and heading. The more modern cockpits have air collision avoidance systems, but when flying visually, it’s still up to the pilot to see and avoid.
In the US at least, pilots can fly without filing a flight plan. Some small airports don’t even have towers! It’s up to the pilots to announce themselves on the radio and take turns.
It’s required to coordinate with air traffic control in restricted airspace such as around major airports.
(This is from taking flying lessons over a decade ago, so my knowledge may be out of date.)
My dad was a pilot for a good length of time, even longer ago, flying things like twin otters for a company. I have a coworker who flies personally out of Vegas every now and then. I've known...
My dad was a pilot for a good length of time, even longer ago, flying things like twin otters for a company. I have a coworker who flies personally out of Vegas every now and then. I've known quite a few pilots in a personally and professional capacity throughout my life and often converse with them because of my interest in flight.
To clarify with that context,
Cessnas do often have autopilots, but this is more like cruise control in a car - it holds altitude and heading. The more modern cockpits have air collision avoidance systems, but when flying visually, it’s still up to the pilot to see and avoid
Right, and that's another point I don't want to get too far into, but the majority of a flight is the easiest thing in the world to automate. My dads plane had "autopilot" which was mostly a physical lock of the heading. A flight from a small landing strip to another small landing strip going no where near major traffic would be very easy to automate, yes. So would having a car take an empty straight road with no traffic. But that's not really the context of the topic or the discussion.
In the US at least, pilots can fly without filing a flight plan. Some small airports don’t even have towers! It’s up to the pilots to announce themselves on the radio and take turns.
Lets be specific about this because that's objectively not true depending on context. Over 18,000 ft you must have a flight plan, end of discussion. If you're under that and in an uncontested area, yep go nuts (again, this is mostly the straight road analogy in my eyes, with landing and takeoff always being the hard part, absolutely being the hard part here).
Finally even though you can, doesn't mean you should. If there is a tower or air traffic control, you at the very least should coordinate with them, and may very well NEED to (see flying near any major commercial airport).
So going back to the original point of "well planes are automated.". Planes are also some of the most controlled devices in world, and the automation is NOT what people think. Yes there are edge case exceptions, but you don't just plug in coordinates whenever you feel like it on the ground at the airport and arrive where you wanted.
This is relevant because that IS how people drive. You don't have to log jack, there is no one trying to make sure you're 2 miles apart at all times, no one orchestrating the congested areas and directing traffic every step of the way, and you've no idea the quality or training of anyone else around you. It's nothing alike.
Granted one of the big reasons for that is that a vast majority of traffic accidents end in financial damage and maybe health problems. When a plane crash does occur (rare as they are) they are much much much more likely to be lethal.
Ok. Thanks for the answer. We don't have these cars here, I'd be hesitant too. I'm distrustful of the companies running this, not so much the tech in general, but I won't be the test case that...
Ok. Thanks for the answer.
We don't have these cars here, I'd be hesitant too.
I'm distrustful of the companies running this, not so much the tech in general, but I won't be the test case that goes wrong.
It looks promising and Waymo seems to be far less about going fast and breaking things like Tesla seems to be doing. We'll see in a bit if it stays and remains consistently safer than humans.
Hopefully it expands to SFO soon. At this point I would pay more just to avoid SF uber/lyft drivers.
Can’t you do that already just by getting a normal taxi? And the SFO to BART connection is pretty good if your final destination is anywhere near BART or MUNI.
The SFO taxis are even worse than the ubers. From the frying pan into the fire.
I mostly go for work or to visit friends these days, so I am nowhere near market street or mission.
Which is part of the reason why Uber and Lyft were able to get off the ground in the first place. Lower prices were a big factor too, but if the taxi experience had been good beforehand, neither would’ve grown nearly as quickly.
Ugh, the first time I got an airport taxi in Las Vegas, it was August and well over 100F/37C. The car had no air conditioning and I found out when I arrived that it was cash only. I was VERY hair to use a ride share there when they became available.
From the article:
If I am at an airport and I have a choice, I'm taking the taxi with a human driving it.
Can I ask for the reason why you would? Just curious.
I've seen many articles about self driving vehicles malfunctioning. I wouldn't want to take the risk.
"Car operates as expected." Would you read that news article? Neither would anyone else. I'm not actually saying you should ride in a self-driving vehicle, but consider that frequency with which something is reported on is a bad metric to use for actual risk. More than 40,000 people die each year in regular car crashes in the USA. It's not more safe just because it isn't news.
I work in I.T.. I regularly see automated things go wrong. Human driven taxis for me.
That's fine, but wasn't my point. You do you, but think about how the media chooses news articles when it comes to reporting novel failure over regular failure.
It is funny how so many people on social media can not accept that someone else can understand their argument and not agree with it at the same time.
My point, which I will repeat, is that based on my own experience and what I've read is that I do not trust new automation technology.
If you understood my argument, perhaps you might have said something at any point that indicated that understanding as opposed to just stating a different, unrelated point? It truly looked like you didn't even register what I said. I apologize for reiterating my point.
I did understand your argument and I did not agree with it.
I'm sorry those two things coexisting seem inconceivable to you.
No disrespect meant.
Have a nice weekend.
...lol. It seems like a little disrespect is meant.
You understood my argument that it's reasonable to have media literacy and think about what they have the incentive to publish. And you disagree with it? I'm fascinated. Can you go into that a little more?
Can you not say much worse things about human drivers? You’ve lived on the Earth for how long? And yet you trust their abilities that much.
As someone else who does a lot of automation, no. I do not.
You know who wrote that automation that’s driving your cars? Human beings. Worse in my eyes, ones that seem to benefit from claiming all their stuff is perfect to the tune of billions.
I think Waymo is miles better than Tesla, but the only thing worse than a single human driver in my eyes is one owned by a corporation running on code and tools built by it
As someone who does a lot of automation, isn't it pretty apparent that humans are highly fallible and generally bad at things? Not just in general, but in this specific case:
https://www.nbcbayarea.com/news/local/waymo-driverless-cars-safety-record/3858460/?amp=1
Waymo seems to be quite a bit better at driving than humans are.
About as apparent as AI being generally awful and bad at things.
Theres no reason to think a car, which is not executing a constant set of instructions but instead required to adapt to its environment, will be any better
What about the published evidence above showing that it is?
See the fact you think such evidence exists is concerning. The VERY small amount of data we have is under very specific conditions and absolutely cannot be used as an apples to apples to human drivers in all situations. 99% of it is also solely provided by developer and there’s fuck all for audits and standards. It’s somewhat akin to pointing to the low conviction rate for cops and saying they must be fine because the data says so.
Show me safety records for automated cars on the highway in the snow? A useful comparison point because one area where humans are absolutely weak at. How about rain on badly maintained roads? How are they handling construction?
Right now most of the data shows what you’d expect. Machines don’t get tired, don’t get drunk, and don’t push it. They’re also limited to the conditions still where they’re at their best. Namely highly important pilot programs.
How does a Waymo that hasn’t been properly maintained for 2-10 years perform? Because most cars, busses, and trains in the US are hardly held to high standards.
What company is spear heading the call for Congress to pass laws mandating repair of this VERY expensive tech before it’s on the road. How detailed are the standards provided? Are we now going to think that these corporations will be benevolent instead of cost cutting?
We’re at the very begging still of programs like this and people act like there’s a mountain of independent and reproduced evidence. Make insurance companies liable and start building actuary tables and I’ll start to care because at least another corporation will have to make their money off of the data.
Self reported highly controlled test cases are not anywhere close to scientifically rigorous. If Waymo wasn’t at least trying to do parts of this right I’d be even more harsh because they’re basically the only ones in the passenger sector worth a damn, and if you gave them the green light tomorrow I’d bet it’d be a Google situation in 5-10 years where everyone wonders why the multibillion dollar corporation didn’t turn out benevolent.
This is of course a big part of why they have a good safety record. Waymo can just decide not to drive in places or under conditions they aren’t ready to tackle. (Such as driving on the highway.) This isn’t a pilot program, it’s how they operate in cities where they’ve been in production for some time. They’re not going to drive in the snow until they’re ready. When they do decide they’re ready, that’s a sign that it will probably just work.
This is asking for speculation about a system that doesn’t exist. There are no signs that Waymo is going to sell self-driving cars or otherwise allow this to happen.
For Waymo at least, there is no sign yet that this is necessary. As long as they continue to have a good track record, there’s unlikely to be a push to solve a non-problem based on speculation. What incentive do company managers or legislators have to mess with success? It might just make things worse.
This would be a weak incentive at best and stronger ones are sometimes ignored. As we saw with Cruise, bad accidents (and a coverup) can cause the entire effort to be cancelled. They had external incentives not to behave like that, but they did it anyway, because internal company culture somehow wasn’t right.
That’s a pretty good incentive to make sure it doesn’t happen, if you can see it. Whether this cautionary tale is heeded at another company depends on that company’s culture.
So far it seems that Waymo is not rushing things. It’s true that it might not last. Can they preserve a safety culture as they scale up? Hard to say, but the threat of bad press and more government regulation is incentive to try.
Contrast with Tesla, which has a CEO constantly pushing employees to move faster than they believed was possible. The difference in safety culture is dramatic, and it comes from the top. It’s gotten people killed. I do think governments should be doing more about their nonsense, but the US government has bigger problems.
The data we have is under the conditions that Waymo actually operates in. Why are we talking about snowstorms in San Francisco, Phoenix, or Los Angeles? Yeah, maybe we should be hesitant about taking a Waymo in a place that Waymo does not operate but that seems like a bit of an irrelevant concern.
This sounds like an argument for using Waymo today.
The proper, honest spin would be "study shows Waymo has 99.9% safety rate, 300% safer than human driver" (you know, assuming studies are being done).
It's not eye popping, but it is exciting the day we can see that headline. If they can't show that, or at least properly show who's accountable for a self driving service, I won't trust it.
I mean here you go
https://www.nbcbayarea.com/news/local/waymo-driverless-cars-safety-record/3858460/?amp=1
Do you ride on airplanes? Most of the time they’re in the air they are on autopilot. Like flying, robotaxis require giving up an understanding of what’s going on. You don’t know what happens in the 747’s cockpit and you wouldn’t comprehend it even if you could see. Same for a robotaxi. It’s largely a black box. But in both cases the injuries per mile are much lower than driving yourself.
It’s important in the modern world to be able to trust engineers most of the time. Trust the civil engineers that built the bridges you cross. Trust the engineers that built the skyscrapers you walk within or beneath. There are always examples of catastrophic failure. But you’ll oversample on freak accidents if you don’t make an effort to avoid the natural bias.
Planes were made as safe as they are through extremely stringent regulation and strong government oversight. While that has been eroded lately, the baseline level of safety in aviation is still quite high, especially for the older airframes. Even when they operate in autopilot, these systems are also overseen by two highly trained pilots and further supported by an infrastructure of ground support, even to being able to troubleshoot problems in flight with the engineers who built the system. Furthermore, the autopilot basically flies in an empty sky (getting within several hundred feet of another aircraft is already considered a near miss). Aviation autopilot systems are vastly simpler and have a vastly simpler task to perform than even the simplest AV use cases.
The US automotive market is almost completely governed by voluntary self-regulation. Compliance with the major software safety standards (ISO 26262, ISO 21448, UL 4600) is voluntary and there is no government body that oversees it. The only mandatory standard is FMVSS, and it only mandates certain items on the car, not the correctness of their function. NHTSA can only act after the fact through recalls, and even there, their ability is very limited. The only other control on the industry is lawsuits, which, unfortunately require someone (or many people) to die or have serious injuries before they can move forward. By the time they are settled, the models in question might be a decade old.
There may be reasons to accept riding in an AV taxi, but it's not the same as flying by any stretch of the imagination.
I trust engineers. I don't trust insurance. Planes already solved that problem. Self driving services are still pointing fingers everywhere else.
I think it's important to make a distinction between Waymo and the other attempts at making self-driving car services. Lumping together Waymo and Tesla's nascent effort wouldn't make sense since they're so different.
It takes a long time to collect enough data to prove it, but Waymo seems to be headed towards airline-level safety.
Planes literally must tell every other plane where they are and plan their route with multiple governments in advance. It’s not remotely apples to apples and still has 2 humans at all times for emergencies
Nit: this is true for airlines, but not general aviation (which is much riskier).
And lacks any form of autopilot like was being described in something like 99% of cases.
And you still file a flighplane and coordinate with an airport and have a traffic controller you interact with.
Cessnas do often have autopilots, but this is more like cruise control in a car - it holds altitude and heading. The more modern cockpits have air collision avoidance systems, but when flying visually, it’s still up to the pilot to see and avoid.
In the US at least, pilots can fly without filing a flight plan. Some small airports don’t even have towers! It’s up to the pilots to announce themselves on the radio and take turns.
It’s required to coordinate with air traffic control in restricted airspace such as around major airports.
(This is from taking flying lessons over a decade ago, so my knowledge may be out of date.)
My dad was a pilot for a good length of time, even longer ago, flying things like twin otters for a company. I have a coworker who flies personally out of Vegas every now and then. I've known quite a few pilots in a personally and professional capacity throughout my life and often converse with them because of my interest in flight.
To clarify with that context,
Right, and that's another point I don't want to get too far into, but the majority of a flight is the easiest thing in the world to automate. My dads plane had "autopilot" which was mostly a physical lock of the heading. A flight from a small landing strip to another small landing strip going no where near major traffic would be very easy to automate, yes. So would having a car take an empty straight road with no traffic. But that's not really the context of the topic or the discussion.
Lets be specific about this because that's objectively not true depending on context. Over 18,000 ft you must have a flight plan, end of discussion. If you're under that and in an uncontested area, yep go nuts (again, this is mostly the straight road analogy in my eyes, with landing and takeoff always being the hard part, absolutely being the hard part here).
Finally even though you can, doesn't mean you should. If there is a tower or air traffic control, you at the very least should coordinate with them, and may very well NEED to (see flying near any major commercial airport).
So going back to the original point of "well planes are automated.". Planes are also some of the most controlled devices in world, and the automation is NOT what people think. Yes there are edge case exceptions, but you don't just plug in coordinates whenever you feel like it on the ground at the airport and arrive where you wanted.
This is relevant because that IS how people drive. You don't have to log jack, there is no one trying to make sure you're 2 miles apart at all times, no one orchestrating the congested areas and directing traffic every step of the way, and you've no idea the quality or training of anyone else around you. It's nothing alike.
Granted one of the big reasons for that is that a vast majority of traffic accidents end in financial damage and maybe health problems. When a plane crash does occur (rare as they are) they are much much much more likely to be lethal.
I agree that none of this has much to do with driverless cars.
Ok. Thanks for the answer.
We don't have these cars here, I'd be hesitant too.
I'm distrustful of the companies running this, not so much the tech in general, but I won't be the test case that goes wrong.
It looks promising and Waymo seems to be far less about going fast and breaking things like Tesla seems to be doing. We'll see in a bit if it stays and remains consistently safer than humans.
SJC makes a lot of sense as an early adopter. Particularly because of how easy it is to navigate by car.