Self Driving Cars are Going Nowhere

Investment Analysis Clubs Macro Economic Trends and Risks
61

[quote=“OrmontUS, post:60, topic:79202”] Given enough time, effort and money these things are possible - and desirable
[/quote]

That’s rather my point. It took legislation from one of the biggest trading blocks in the world to make it happen. The magic of the market didn’t.

We all lived through the BetaMax vs VHS era, that was a time when the market sorted it out, but only after consumers had spent billions guessing on which format would survive. That seems a rather poor model for automobiles which need to be recharged - or abandoned by the side of the road.

Sometimes standards need to be set by fiat, other times it just happens. (Stop lights took a long time to be standardized, and the UN has a whole division which tries to rationalize street signs among nations, which is why is most countries stop signs are red and octagonal, even if they don’t say “STOP”.)

Even now some websites work better on Firebox, some on Safari, some on Explorer, etc. The differences are usually trivial, but there is never going to be a standardization of all things driving: fireplugs, location of drives, painting at the edge of roads, guard rails, look and feel of speed bumps and crosswalks, and obviously temporary things like construction crews, warning cones, road flares, downed power lines, etc.

The amount of data required to process all these edge issues and the amount of time needed makes full Level 5 unlikely any time soon, in my view. (Even in slow motion: I am amused by the story of the Tesla rolling slowly but deliberately into the rear of a private jet at a billionaire-ish airport because, apparently, it had never been trained to recognize a jet. And of course, why would it? Heh.)

1 Like
62

I write short posts. I assume you read them. I am not sure you understand them. But this was about your own posts. You left off a comment or study really that FSD cars have fewer accidents than humans in most situations. Not sure how you write that and then say otherwise.

63

Not that they have “fewer accidents than humans in most situations.” A Level 2 car might have fewer accidents in some situations (we don’t yet know) as long as they are continuously monitored by a human. But that Level 2 car cannot drive itself ever, and they cannot even drive themselves even with human monitoring in many other situations. So if you were to try to let that car drive itself as if it were a Level 5 car (ie. all the time), it would be extremely dangerous.

Which is why there are no Level 5 cars out there. The Level 2 cars aren’t secretly Level 5 cars that can drive more safely than humans - they’re just Level 2.

2 Likes
64

While Level 5 may not be achieved in our lifetime, the overall introduction of EV’s will certainly be expedited by an industry/government agreement standardization of all “non-competitive” aspects (say, similar to a product adhering to the DIN standard in Europe). While each manufacturer can be as proprietary as they want on their guidance system, maximum GPS resolution, response time, understanding of traffic signs/laws and so on should be standardized. So should “refueling” plugs and possibly even physical aspects of the car’s “envelope”.

Jeff

65

I think the key distinction is if Human+level 2 is safer than Human only, it is likely worth adding level 2. Same argument as adding ABS (1994+) or adding rear view cameras (2018+), seat belts (1968+) etc.

I assume Human+level 3 will be better than Human+level 2, and so on. So each level will be added as it matures.

Level 5 is, of course, a special case because it has to fully operate, at all times, under all conditions, without a human. And because we are human, it is possible that even if level 5 is safer than Human+level 4, it may still not be accepted until it is MUCH safer. In fact, it may not be accepted until it is 100% safe, and even then, those rare cases of failure (even if due to something external) will get really bad press.

66

My MIT trained nephew in Computer Science agrees with him…

Going back and forth over outdated information is not productive.

67

My guess is that the main selling point for levels 1-3 is ease of driving rather than driver safety. At those levels the driver is still responsible for the driving of the vehicle but the automated stuff, like cruise control, makes driving more relaxing.

I think level 4 will be widely available in five years. This is typically defined as autonomous driving under restricted conditions or geography, i.e., Operating Design Domains (ODD) and is currently being tested in Austin, Detroit, and San Francisco, as well as several areas in Europe. At level 4 the AI will signal when a driver is needed and safely park the car if a driver doesn’t take over in time. One can easily imagine driverless taxis in retirement communities, like the Villages in Florida, or in suburbs where the taxi is limited to 40 mph. If for a thousand bucks/year you could contract for unlimited used of the taxi, that may reduce the desire for second and third vehicles.

1 Like
68

If you want people to pay for a discussion board you need to have a few celebrities posting a few times a week. Perhaps Peter Lynch is bored with retirement or Robert Reich might want a new sounding board. Even a Bernie Madoff would get you some traffic.

69

I don’t even think they need to be celebrities. Look what Saul has done with his board. Think about Mungofitch on the BRK board. Look at MeTAR with Wendy. Once upon a time it was Mish.

The model is there, I don’t think the Fool has ever gone about trying it in any significant way. (It may also be that its not economically viable to “hire” this level of board leader, I’ve not really thought about the economics of it.) I recall in pre-history the Fool hired one of the more prominent posters on the @Home board. He was a cheerleader rather than a thoughtful analyst, but the board brimmed with activity, although that was probably as much a function of the time as well as his “curation.”

No, it’s been done, it’s been shown. On Twitter, celebrities, yeah. They make Twitter bigger, Twitter makes them bigger. I’m not sure what the magic glue would be for people like that at the Fool.

1 Like
70

We are the freemium.

The fool does its premium stuff in-house. But I will guess the fool is very aware of MarketWatch’s marketing formula.

Regardless of some folks here having growing pains, this new format is open to growth for the fool. The old format was not. The next wave in here will be Millennials giving us ideas while we cherry pick based on experience.

71

I’ve been driving Tesla’s Autopilot from 2015, and more recently Tesla’s Full Self Driving.

Level 2 systems, nothing more. They make driving easier under the circumstances where they work all the time, and make driving more stressful in situations where they fail much of the time.

For a long time they’ve effectively been level 4 on limited access highways, and are a huge improvement over driving yourself in stop and go traffic on same. But they are still level 2. This is mainly because Tesla’s vision is grander, so they aren’t going to claim level 4 until they can list the few exceptions.

You can see Tesla’s statistics here (Tesla Vehicle Safety Report | Tesla). While the latest numbers show about half the per mile crash rate for Teslas driven with Autopilot on vs. Autopilot off, you could argue that obviously people use Autopilot only for the easiest driving tasks. No way to prove this isn’t the case. It’s certainly not the case for me, but I’m not the typical driver.

-IGU-

6 Likes