Hacker News Reader

Comment by H8crilA | Hacker News Reader

Tesla has been selling level 5 autonomy since 2016. I can see nothing wrong with that. Definitely not vaporware /s

https://www.tesla.com/blog/all-tesla-cars-being-produced-now...

I realize this is a sarcastic comment, but I think this illustrates the problem.

Tesla vehicles cannot operate fully autonomous as-of now. Tesla has no idea if they'll need to replace or upgrade hardware either. They simply have not solved the problem yet, but want everyone to buy into the hype. And, the hype machine is working, unfortunately.

It's a scam. They're implying level 5 capability is just around the corner, when none of the current hardware will ever get past level 2. The current cameras can't even stay clean enough to keep autopilot reliable, it'd be pretty foolish to rely on them for any kind of real self-driving.

I love my car but the FSD debacle is embarassing. A lot of people are finding out that they'll never get anywhere near $8K value from their pre-purchase, and they bought a license that expires when they sell the car. I expect a class action lawsuit eventually (I'm surprised it hasn't happened already).

> Tesla has no idea if they'll need to replace or upgrade hardware either.

They have upgraded the hardware 3 times already.

Maybe the hardware in the last generation of Tesla's suffices, but right now, nobody knows.

Sometimes it seems hype is the main fuel driving Elon's businesses. Self driving and self driving taxis for Tesla, Starlink for SpaceX,...

But I have to admit, marketing wise he is a genius.

I'm not sure it should be called level 5 autonomy until Tesla is willing to take responsibility for any accidents it causes.

Not sure about that, but I'm sure it's not "full self driving capability". It has not been the case ever since they started advertising in 2016.

When a plane auto-pilot fails (e.g. Boeing 737-MAX), the manufacturer is responsible for the failure and its consequences.

That has always been the definition of auto-pilot.

Tesla might be able to redefine the term in the US, but this ruling is proof that other countries disagree with that.

Whatever Tesla is shipping isn't an autopilot, and calling it that is just deceptive marketing.

I mean looking at it it would be a good litmus test for how confident any automaker is in their tech, if the driver is still liable for crashes it implies the driver isn't fully out of the loop which is what I'd expect to be true if we're actually living up to the Autopilot name.

This is correct. The driver is responsible when their actions—in contravention of the road rules—cause an accident. If a vehicle is L5, the only "driver" is software. This software would need to attain a driver license just like a human would. Thus software becomes the vehicle's legal driver for the purpose of accountability in an accident.

Tesla is absolutely willing to take responsibility. You'll have to use Tesla as your car insurance provider, but it'll be cheaper than anyone else's insurance so that's fine.

Why would it be cheaper? Because they're insuring a driver that never gets angry, never gets distracted, doesn't drink and has high resolution dash cams pointing in every direction and capturing every moment. Even as the software is imperfect, it's going to have so many fewer accidents that the statistical risk of insuring the entire fleet is going to be minuscule compared to random humans.

I wonder why Tesla even needed to go that far to make the product sell. It is already a damn good car, a leader in its space. Also, Waymo which was closest at that time to Level 5 autonomy was far off from a product.

It’s never been about selling product, although it helps. The hype around autonomy and full self driving has fuelled their share price growth for some time.

Just selling cars alone does not justify Tesla’s current valuation (how could it? Making cars is a notoriously tricky and oftentimes unprofitable business).

There is a reason no new manufacturers have succeeded in the past 100 years. Auto industry is tough and costly.

Autopilot almost certainly gave them press and coverage that has helped sell cars and keep the company from collapsing.

If a future promise is what it took to drive the battery revolution, it was a reasonable cost.

It's not just for press, they directly make money selling "FSD" add-on for ever-increasing thousands of dollars which currently does a couple of gimmicks, but, trust him: will soon be driving you around effortlessly (and other passengers for $$ when you're not using it)

Nobody said 'just'

This is exactly how I see autopilot as well. You can debate the ethics of this approach and people will forever (I certainly think its been misleading the public), but it so far seems to have worked for Tesla and has arguably forced industry in the BEV direction sooner than otherwise might have.

The bigger issue (for Tesla at least) is what happens to Tesla if they don't deliver the promised autonomy now in a reasonable timescale. Company itself still isn't all that independently secure, yet to turn an _annual_ profit etc. It feels to me like the recent surge in their value is driven by over-enthusiastic investors hopeful that one day all those Model 3s on the road become autonomous revenue generating entities (driverless cabs etc), which I still can't personally foresee happening anytime soon.

I own a Model 3 - the cabin alone, while nice enough, wouldn't stand up to more than a few months "shared" use IMO, ignoring all the other technical barriers. If these things really become shared, we are going to have to seriously rethink cabin design and materials. The kind of cabin a private buyer expects when spending 35k plus on a car today simply isn't going to last a bunch of drunk kids autonomously Ubering, even more so when you consider a long range Model 3 is approx $50k today.

Despite my own misgivings about how autopilot is conveyed to users, I still find it to be a really competent auto steer system on highways, which of course is a relatively simple use case for these systems to excel in. Ford's decision to brand their competing system in the forthcoming Mustang Mach E "CoPilot" feels far more appropriate. Ford aren't as dependent on hype for short term survival though.

Because they can’t sell the car for what it cost to produce. FSD has added ~$3,000 of essentially pure profit

Nope, that's not profit. That's positive cash-flow. They are investing heavily into developing FSD.

I'm doubtful whether that racket will pay off. I guess they will have to pay more and more back to disillusioned buyers. It certainly hasn't turned into profit yet and rather looks like a massive risk.

They can recognize some of it as pure profit.

Reply on news.ycombinator.com