Hacker News new | ask | show | jobs
by scylla 796 days ago
I, and apparently a lot of other people on the Internet find the latest Tesla FSD update pretty incredible. There's a reason why Tesla gave it as a free trial to everyone. I drove it downtown and back without any interventions this weekend.

Is it 100% perfect? Of course not. But this article is the ultimate in low effort. A person who felt that the car might hit another one but probably not. Yes, real value there.

12 comments

> the article is the ultimate in low effort.

What’s low effort in sharing your personal experience? It’s very clearly written as an opinion.

Any reader would be reasonable to assume that there are other opinions out there, and that this opinion may be in the minority.

That said, I’m happy to see a dissenting opinion in the comments here! Would be curious to see others thoughts in the sudden appearance of a free trial.

The author went in with a negative attitude opening with hate for the CEO and then proceeded to diss the product based on how it "felt like it was about to" do something bad rather than what it actually did. It seems that the author has already decided that it was bad before trying it and then decided to confirm their bias based on vague feelings.

I find it extremely biased and lacking in substance.

I get it though... Elon does deserve a lot of hate so I guess the HN crowd is gleefully cheering out of schadenfreude whenever someone says something along the lines of "Tesla bad".

> The author went in with a negative attitude opening with hate for the CEO

I hate to break it to you, but that's going to be a lot of people. He's made himself extremely hateable by the general public.

There are lots of cool things that probably shouldn't used by people operating heavy machinery. Phones for example.

From my experience working in this space though, judging quality is extremely difficult. For one thing, different people have different criteria. I've taken rides with other people where we all had different opinions on the quality of the ride at the end. These systems (FSD especially), can also be highly sensitive to specific details of the situation. Two drives in apparently similar conditions have meaningfully different behavior because the system understands them differently. It's very difficult to compare apples to apples outside statistics and simulation.

All this is to say: don't dismiss feelings or intuition about the danger. These are good caution signs and talking about them can help companies improve the product.

I think everyone using this product should be asking themselves why we don't have an objective measure for the quality (or lack thereof!) of the product.

With every new release of FSD, there are people quick to say it's a breakthrough, and amazing, while others lament it doesn't nearly do what it says on the lid.

Why are we putting safety critical systems we can’t adequately measure on our streets?! There are people who confidently state that Tesla's FSD is already better than humans.

Where's the data?! Why isn't it open? Why can't we build an objective measure for effectiveness and safety of each of these releases?

The data isn't open because it's not required to be. Every manufacturer is either making a good faith effort to collect that data (to their own standards) or they probably shouldn't be on the road. Regulators, NHTSA especially, don't have the expertise, time, or mandate to get it from them, analyze it, and make it available. That could (and should) change though.

As an unrelated aside, measuring safety performance is hard. Doing it accurately without putting vehicles on the street is probably impossible. Simulation is not a fully adequate replacement for road testing and many companies are already looking at cutting the latter because of the expense. It'd be nice for regulators to outline an acceptable road test framework that better balances the goal of public safety from unsafe vehicles against ensuring safety can be demonstrated in real world environments.

Actually, there are some attempt to measure it.

https://www.dmv.ca.gov/portal/vehicle-industry-services/auto...

Of course Tesla is not participating - the real numbers would make them look disastrous.

Regarding objective measures for the quality - the famous standards for level 1-5 etc. Objectively speaking, Tesla's system is Level 2. There are other companies at level 3, and some others operating fully autonomous vehicles.

As to why Tesla chooses to base all their claims on purely subjective and highly cherry-picked numbers and videos ? We all know the answer to this question.

> Tesla's system is Level 2

Which I'm happy with. I do 95% of driving on Autopilot already. Add another 4% for FSD that I'm happy to babysit and I'd be very happy.

For some reason people are so caught up on naming choice and what it implies...

There are several reasons - for some it's about security, for other about truth.

I find it fascinating that a guy become the richest person in the world, by convincing 'investors' that a Level 2 system is 'almost ready' for a Robotaxi operation and will make them infinite money. To give him credit, he also did mention that Tesla is worthless if it doesn't solve FSD.

Now this is an opinion I can get behind.

I'd love to see all cars (including Waymo) compete on a standard set of tests with clear gates as to what conditions they'd be allowed to go to Levels 3, 4 or 5.

> There are lots of cool things that probably shouldn't used by people operating heavy machinery. Phones for example.

I will tell you straight up that a Tesla model Y on FSD 12.3.3 is a safer vehicle operator than a human being using a phone. Period. No qualifications. Under all circumstances.

Yeah, and? Being safer than an objectively dangerous behavior doesn't make it safe.
FWIW, I'm happy to see the "Yeah" there. But regarding the "and", it gets to the sincerity of the discussion. In a world where literally millions of people are texting from the drivers seat of a moving vehicle every day, freaking out on the internet about a handful of robots doing less dangerous stuff seems maybe a little performative.
I have a bit of professional interest here because building these kinds of systems safely is my job, whereas distracted driving isn't something I can meaningfully change as an individual. At the very least, I can explain to others what a safe development process looks like.
> Is it 100% perfect? Of course not.

Considering you are putting your life, that of your family and everyone on the street around you in the hands of this software, I don't get the nonchalant "sure, it's not perfect" attitude like you were trying out the beta version of a new note taking app. It absolutely has to be 100% perfect before being allowed on the streets.

I use it daily and I would be the first person to agree that it isn't 100% perfect, but not in the way you are assuming. The flaws it has are that it errs on the side of being too cautious.
Being overly cautious on the road is as bad as being reckless. The frequent example about 4-way stops is a perfect one. If a car is behaving unpredictably and stopping when it's their turn to go then it is creating a dangerous situation for everyone at the intersection.
Of course, it's a spectrum. It's nowhere near dangerously over-cautious. I'd describe it as a very prudent new driver. It drives the way most people probably should, but that means there's a little hesitation in low-risk scenarios sometimes. All you need to do is tap the accelerator and it instantly gets the message. We're talking about something that's barely an inconvenience, but it isn't the "100% perfection" the people here seem to blindly demand.
You can always disengage and drive faster. Old, overly safe and imperfect drivers exist already. Do you suggest kicking them off road?

How about hardware? Is it all 100% safe? IIRC USA doesn't even require yearly warrant of fitness / MOT.

Are the roads perfectly designed and in perfect conditions? How about weather? Do you stop driving when there's blizzard outside?

Yeah, just like human drivers.
I think the second they used the term "Full Self Driving", they opened themselves up to criticism when it doesn't fulfil that promise.

Hell, I'd argue they shouldn't be able to call it that legally.

My devil’s advocate response for that would be that Tesla as the first mover on the market gets to call their product whatever they want.

Is a “garbage disposal” a device that can dispose of all garbage in your sink? No, all it can handle is light food scraps. But the inventor named it a “garbage disposal” not a “light food waste processing device.”

To me it seems stupid that so much of the debate after all these years is about the name. I think we are beyond the point where the name matters.

They get to call it that when it can drive safely on snow covered roads. Before that happens, it's marketing hype that can't be relied on in all scenarios. When they transitioned to radar-only, they sold "FSD" that needed a whitelist to ignore the sensor input in problematic locations. What other half-baked hacks are in the current implementation?
Here’s a thought experiment: would you turn on traditional cruise control during a snowstorm?

I wouldn’t.

A self driving system that only works during good weather is still useful and has non-zero value.

"Full" is a dangerous word.

Especially for a device that has health impacts.

Typically the FTC regulates truth in advertising by (1) Is it an unequivocal statement? & (2) Does it have a health impact?

Sounds like if that’s true then the FTC will get right on it and has a slam dunk case.
They're a bit busy with Amazon now, I think.
I don't think past mistakes mean we shouldn't demand reasonable names for things.

If you want to use an arbitrary name, use one, if you pick a descriptive name and don't do as described, it's just false advertising.

And yet nobody cares that lottery products have names that imply winning, prizes, pots of gold, etc even though almost nobody wins statistically.

Again, I just think that marketing is marketing and everyone knows that. Everyone knows that a trash bag isn’t indestructible and that a paper towel can’t absorb a tsunami.

I’ve used FSD since the first invite beta and this version has been a step back for me. Multiple break checks on the freeway, incredibly bad path planning through intersections where it just switches lanes mid way through, random turn signals coming on for no reason, list can go on. It’s like they just tweaked some params to reduce the shakiness of the wheel but didn’t actually do much to improve confidence.
Elon is missing a huge opportunity by not hiring his flight attendants from this website. Wouldn't have to buy any horses and probably wouldn't even have to ask.
I don't think it's low effort at all. In fact, I totally see where the author is coming from despite loving FSD and using it often.

Driving is a very personal activity. There are nearly infinitely many variables driving every decision during every drive, and the default experience that people will want from something like this is "someone that mostly drives like me and never makes mistake because computers".

FSD is not that. It will be a while before it is that. Until then, I can totally understand someone "never using this again" (for a while) over a few fumbles that FSD made that they would not have (in most cases).

Tesla has to know this because they also know that if even 10k Tesla buyers decide to subscribe to FSD, that's an extra $2M/month that can be (mostly) shoveled back into training infrastructure!

I drove my car downtown and back yesterday 100% perfectly. So while you're happy with not 100% perfect, what about other road users who have no say in whether you turn on your not 100% perfect death machine?
Don't know which utopia you live in, but are you really happy with the other drivers texting, sleepy, high or just completely incompetent at the wheel?

If anything, FSDs are over-cautious and slow down to well below speed limits if they can't make out conditions. So, yes I'd say other drivers shouldn't complain about a system that still requires a human to constantly pay attention and keep their hands on the wheel.

What's the point then? (serious question)
( serious answer )

It makes driving a lot more relaxing in a way that you don't appreciate until you get used to it. My feet are off the pedals and can stretch, I don't have to steer so my attention is only partly on the road ( the interior camera monitors to see if you're looking ahead and not on your phone for example ) If I have to change the music I don't have to worry about the car in front of me suddenly braking etc.

I've been using FSD for years now at day and night and I think it's been ready for Level 3 going straight on Interstates/highways for a while now. The latest update significantly improves taking turns along a route.

I commented on this thread elsewhere that I wish there was a standard suite of requirements for all automakers to certify the self-driving levels on their cars. Otherwise, I feel too much of the discussion is being polluted by people who agree or disagree with Elons politics.

> my attention is only partly on the road

I hope you don't hurt someone.

If paying full attention to driving is so tiresome to you please don't drive. My impression of this description is that FSD is getting good enough to lull drivers into not paying enough attention. There shouldn't be a product like that either Tesla takes responsibility like Waymo or makes a clearly L2 system that is good at limited things. Blurring the line like this is going to get even more dangerous as FSD gets better
Thanks for the answer!
And yet, I still get phantom braking daily with autopilot and auto wipers that don't actually seem to understand what rain is. Call me EXTREMELY skeptical that FSD is ever going to be viable when they can't even get basic blocking and tackling right that every other vendor has had solved for a decade.
And everyone else gets blinded by your headlights which fail to auto-dim like every other manufacturers'.
^^it's horrible, I end up turning them off most of the time because I feel bad. The auto-dim when you're behind another car is about half the distance of every other car I've ever driven.
calling opinions "low effort" seems to suggest that everyone needs to be an expert in anything they talk about. I keep coming across this in technical circles and I find it discourages honest sharing and harms psychological safety
Is it a complicated route or pretty straight forward?

At daytime or ar night?

What about the weather conditions?

My least favorite part of the article was where the author complained about the “financial mistake” of the car, about how they were underwater on their car loan.

Depreciation is a normal and expected part of buying a car. If you are financing a car you should be prepared to keep it for a long period of time. While many automakers offer car loans of 5+ years, ideally you should only finance a car for 3 or 4 years and/or put more money down to ensure you are never underwater.

At no point did the author say that the car didn’t do its job as a reliable piece of transportation, but somehow it was a “financial mistake.”

The fact that cars depreciate and cost too much is a feature of all cars.

> Depreciation is a normal and expected part of buying a car

The problem is that Tesla owners are spoiled. I bought my Model Y just before the supply crunch in June of 2021, and within a year used models were selling for 20% more than I paid. They've since come down a ton as demand levelled off, but even now my car has depreciated only about 18% (including inflation!) from the cost of a brand new Y with the same configuration.

It's easy enough to imagine a young Tesla buyer over the last few years getting a very weird picture of the financials of owning a car.

On top of that, this phenomenon applies to anyone who bought a car in 2021.
Most car manufacturers take some efforts to try and ensure residuals stay reasonable, otherwise customers will be less keen to repeat buy for obvious reasons. Setting prices correctly at launch is key for this.

What Tesla did, which historically is not a "normal and expected part of buying a car", is repeatedly slash list pricing after launch, thus meaning the used market does not trust valuing them, as so many used car dealers got very badly burned on inventory they paid too high a price for when the new car prices came down, putting them all underwater.

These aren't small or predictable market moves either; they are random swings. The Model S and Model X have been reduced new by as much as 24% at times. This Simply Does Not Happen at most car companies, most of the time, and residuals are more stable as a result.

Imagine you had just taken stock of a used model S, valuing it based on purchase price "n", then tomorrow the manufacturer slashes the price of a new one by 24%. That's very, very hard not to end up underwater.

Similarly, for existing owners - that 24% cut has just shifted the depreciation curve massively, and not in your favor. Existing owners frankly have every right to be pissed.

+ Technology additions in early models

+ Parts availability woes

In short, buying a Tesla is a terrible decision from a financial perspective.

Maybe from a fun or cool perspective!

But if you're interested in maintaining $-value-over-time, there are much better options (Corollas, Accords, Tacomas, Subarus, etc).

The flip side of all this is a used Tesla can be a really great option right now for some folks - bulk of depreciation is over, maintenance/running costs are some of the lowest in the industry (ignoring EV insurance costs I guess, but thats another topic...). The traction battery/motors/drivetrains regularly get to 300k miles. I'd have no problem at all recommending a used 3/Y at the right price to someone if EV fits their lifestyle etc.
> Most car manufacturers take some efforts to try and ensure residuals stay reasonable

lol, walk me through the process at Maserati.

Probably why I used the word "most" here. Residuals matter, even to the manufacturer.

If residuals plummet too much, their lease accounting can go red too. Lease rates have to account for a future predicted value of the vehicle, and manufacturers strangely like making money on leases.

Also, the huge depreciation on a Maserati isn't a surprise, most of the time. A mass produced vehicle selling in the hundreds of thousands annually seeing double digit list price shifts at random did come as a surprise, and has put car dealers underwater on Tesla inventory before.

Would an example of putting in effort to maintain residuals be providing software updates and infotainment system hardware upgrades to owners?

Because those are two things that Tesla is doing that nobody else is doing at all or very well.

https://www.tesla.com/support/infotainment