Hacker News new | ask | show | jobs
Tesla's self-driving efforts rank last in study of self-driving companies (businessinsider.com)
23 points by Luka78 1205 days ago
7 comments

The market research firm that put this together is in the business of selling expensive reports and custom consulting engagements to auto manufacturers to aid with automotive strategy.

They have a business interest to make lesser known firms appear to be the biggest threat as it makes their services more valuable. Any manufacturer can benchmark self-driving performance on the widely available Tesla, but it will be difficult to know what all the other firms are doing as they're all, comparatively, niche players.

They fact they don't even have Comma.ai on the list suggests this is a garbage tier ranking study.

> They fact they don't even have Comma.ai on the list suggests this is a garbage tier ranking study.

They also don't have my OS self-driving hardware and software plans. Comma.ai is a cool toy, but it's a joke in the world of real self-driving.

> make lesser known firms appear to be the biggest threat

Lesser known firms like their top four Waymo, Baidu, Mobileye and Cruise? Yeah, okay.

> Waymo, Baidu, Mobileye and Cruise?

2 of the 4 don't have anything available for mass market, and they are money losing ventures.

Mobileye is the only player that matters of the lot, given Baidu will effectively be limited to CCP friendly markets.

The ability to take a 1.5K "toy" to make a standard car have partially autonomous is also pretty nice. A great toy that is pushing the industry forward to make sure self-driving is available at reasonable prices vs $15K premium for a "beta" software that will never make it to production.

I dunno. Have you seen the video of the electric cars trying to park?

Edit: here it is https://m.youtube.com/watch?v=nsb2XBAIWyA

I'm underwhelmed by the article's quality. First, it's an article about a consultancy's opinion of the marketplace. The most said about methodology:

"To come up with the list, Guidehouse weighs factors such as a company's vision, go-to-market and production strategies, partners, tech, commercial readiness, and more. Guidehouse sorts the companies it ranks into four categories: leaders, contenders, challengers, and followers."

Oh dear. How much mere opinion and supposition went into this one I wonder. Did the companies evaluated provide information? Where they even requested to do so? Did only some of them do so? What are the relative weightings of different factors? etc.

Another, perhaps more familiar way to this audience, to read this report is to think about the credibility you'd give to a news article with a title like "Company XYZ Comes in Dead Last in Gartner Group Study". Yes, you know others with important decision making power may weigh such a Gartner Group report in their decisions, but that the report tells you more about the thinking, biases, and objectives of the study authors than anything terribly objective or important about the marketplace itself.

Deceptive title and report as this is ranking based on factors like like “company's vision, go-to-market and production strategies, partners, tech, commercial readiness…” Basically everything except for actual driving ability, which makes this highly subjective and biased.
Tesla's camera-only driving is a gamble. It has the _potential_ to be the best in the business.

...but it also takes a ton of computing power to determine how far objects are just from a moving image. A LIDAR unit would give you range + position instantly.

Also assumes the current camera suite is high enough resolution, high enough dynamic range, sensitive enough at night, and positioned in correct locations to provide the data the compute needs.

Arguably that is not true still.

The issue with the camera-only model is it is very non-deterministic.

Personally, every accident I have gotten into is a result of some sort of inattention, distraction, or fatigue. Computers don't experience these things.

Nonetheless, Tesla's occasionally do unpredictable insane things like turn directly into a bollard, cross in front of a moving trolley (that it can see & has rendered on screen), run right into a firetruck, run over child size dummies, etc.

It's worse than supervising a teen driving because a teen will fail in somewhat predictable ways, generally in being either overconfident or over hesitant. You can verbally coach a human as a situation approaches. With FSD you either let it ride, intervening at last moment.. or you need to constantly step in proactively.

"Not hitting things" isn't really the hard part of driving, though. Yeah, it does sound like the absolute minimum level of effort, but it doesn't get you anywhere. It just tells you some of the places not to go.

You need to read lane markings, locate cross streets, interpret signs, stop at traffic lights, etc. All of that requires a camera... and if you can do that, you might as well use it to not hit things.

LIDAR would be great in a belt-and-suspenders way, and it seems as if it would have saved some lives already. The other stuff is all incredibly hard; it's not ready and may never be ready. But if self-driving ever works, it would have to be able to work without LIDAR, because it has to solve the harder problem that also encompasses the things that LIDAR could help with.

> "Not hitting things" isn't really the hard part of driving

Until your tesla hits a stopped firetruck at 60mph, then it's a real hard part

That's just it. If a Tesla can't figure out that it shouldn't drive through the big red thing, then it can't do anything at all.

I don't have any idea why that happened. It seems odd. I'm afraid I'm always a little suspicious of such stories, since so many turn out to be people lying about the auto-driving.

It's entirely possible that Tesla failed to train the things on fire engines. I rode in a Tesla the other day and was a bit surprised that it didn't note ambulances as different from other trucks. If they've failed to train them on any emergency vehicles, that's a serious problem: there are laws like "pull over for fire trucks" that it can't follow.

And again, that's not a problem LIDAR can solve. Something is very wrong if it identified a big red object as a thing it was allowed to drive on. LIDAR would solve the accident, but not the fundamental question of whether self-driving would work at all.

> And again, that's not a problem LIDAR can solve.

Tesla are scared of clouds and shadows and don't see big metal boxes

It's 100% something a lidar would help with. It wouldn't hit shadows or clouds but it definitely would hit a big metal box early enough to at least brake instead of accelerating

It seems biased. Tesla exaggerated claims feel scammy, but at least they have a product out there. None of these “leaders” have something you can test yourself.
Perhaps because none of them are actually ready for usage in public?

That would say a lot more about Tesla's practices than about the ability of the others to deliver a product.

Charging $15k for software that doesn't work yet, and encourages dangerous behavior by untrained testers on public roads, for years, is not brave.
They make it clear that it's not ready yet at time of purchase and there is no definitive launch date. You are by no means required to pay for it and most people don't.

>encourages dangerous behavior by untrained testers on public roads

There are idiots everywhere. The functionality has been locked behind their driver safety score to ensure only responsible drivers can beta test. There's a delusion that it is dangerous based on the amount of press coverage any Tesla involved incident gets.

Do I trust that their car is better at driving than me? No. But is is better than most people on the road. I'd rather be driving near a Tesla than a Prius.

Great, how do I opt out of having other people's Tesla FSD near my kids?

> But is is better than most people on the road.

So says Elon. Yet he won't release disengagement data. Data speaks louder than blow hards.

Sure, as long as your kids are tall enough to ride.. for some height that Elon hasn't definitively disclosed.. with some probability curve.

* 90% won't kill your kid if 4ft+

* 70% won't kill your kid if 3.5-3.99ft

* 50% won't kill your kid if 3-3.49ft.

* All bets are off below 3ft.

Also make sure they are moving enough for the car to identify them, but not too fast for the car to be unable to react.

Make sure they wear high contrast clothing and it's not too dark (or too bright & glaring) out.

> They make it clear that it's not ready yet at time of purchase and there is no definitive launch date. You are by no means required to pay for it and most people don't.

Elon has been loudly and repeatedly saying it's a year away for what, a decade?

> The functionality has been locked behind their driver safety score to ensure only responsible drivers can beta test

Pretty sure that's no longer true.

100% They unlocked it for all buyers without safety score requirement in December so they could book all the revenue finally
I owned one for 4 years and all of this is in the eye of the beholder, and constantly moving goal posts.

A lot of what FSD does & doesn't support is left intentionally vague.

At various times the FSD product page, the car sales page, and what Musk tweeted were completely out of sync as to what mode does what.

Lots of people bought the car thinking it handled situations it very much did not, and have attempted to use it in unsupported ways.

For example - poll 10 Tesla users about Autopilot on non limited-access highways use and you will get 4 different answers. Then ask them about what objects Autosteer is capable of avoiding. LOL.

You are never testing as self-driving car only by yourself. Any time you take a self-driving car for testing, you are also including other cars and pedestrians in your experiment. (unless you are doing it on a closed track )
You mean the others don't make public beta tests.
At least we are not beta testing a vaccine.
Baidu, Waymo, and Cruise all offer driverless taxis right now. Tesla is far from offering such a service.
Anyone can get in a Waymo right now.
If you are in Phoenix, AZ. And it only operates in a well-defined area.
Phoenix, AZ is one more place than Tesla can claim. There are exactly zero places on this plane of reality where your driverless Tesla will come pick you up.
ok
Maybe it's because to do cutting edge self-driving you need cutting edge software and machine learning engineers, and software engineers are less likely to put up with Elon's crap because they have more options?

Tesla and Space X get heavy talent (despite the downsides of working for Elon Musk) because in their primary fields they are the leader by a large enough margin that if you're one of the best there aren't a lot of other places to go that are going to really challenge you. That's just not the case for the software involved.