Hacker News Board Hacker News, browsed like a classic message board
Live data from Hacker News

Robots eat cars

telemetry.endeff.com

Post reply on HN
Print view View original

71–80 of 125 posts

Re: Robots eat cars

#71

Post by ErroneousBosh »

cyberax wrote: post #6
Terr_ wrote: post #2

> Steer-by-wire Thinking back to case-studies around the Therac-25 [0], I would like to pre-emptively highlight the differences between: 1. Technique X is unsafe. 2. Technique X is unsafe because too much can go wrong even with the best intentions. 3. Technique X is unsafe without strong QA and interlocking safety measures, and there's too much economic pressure for the manufacturer to cut corners. [0] https://en.wik…

Moreover, Technique X does not actually provide any significant value. The whole steer-by-wire in CT happened because Musk wanted a yoke as the control system. And a yoke requires progressive steering which is impractical without steer-by-wire.

> And a yoke requires progressive steering which is impractical without steer-by-wire.

But also look at Citroën's DIRAVI system, used on the CX, SM, and some XMs. There's no direct mechanical link between the steering wheel and the rack when the system is pressurised. When you turn the wheel a kind of dogbone link thing pushes a spool valve one way or the other allowing hydraulic fluid to push the rack along, which pushes the other end of the link back to shut the fluid off again.

So far, so similar to the Danfoss valve in a conventional power-assisted steering system, except that uses a rotary valve and a big torsion spring in the steering rack (that's why your steering wheel feels springy with the engine off).

But DIRAVI is fully powered with no mechanical link, so how do you get increasing resistance with increasing speed? Well, there's a governor on the gearbox that allows hydraulic fluid into a little cylinder that pushes a spring-loaded roller against a heart-shaped cam attached to the steering wheel shaft. This will try to spring back to the middle, and the faster you go the harder it springs back. At 70mph you can barely move the steering wheel, but it will flick a large heavy car from lane to lane with fingertip pressure.

You have to get used to this and for the first few miles you'll be zig-zagging down the road like you're tacking a dinghy, but after that you'll get used to just thinking about your right pinky finger being a gram heavier and going round a corner. I've driven some seriously high-end sports cars with legendary handling and performance and they feel pretty rough and tractory now ;-)

If the pressure fails of course then there's no powered steering (notice I say powered, not power-assisted), although in practice what tends to happen is that the "resistance" part goes first giving you very sensitive steering.

What happens once there's no pressure is that the steering wheel moves about 20° before you run the valve to its end and then the dogbone pushes directly on the drive gear for the rack. So the steering is very loose and wobbly but you can at least steer well enough to get it out of the parking space and into the workshop. You still have brakes for an hour or so if the pump belt breaks, and enough steering to get safely to the side of the road, or at least out of the fast lane.

In the 1960s they had a prototype Citroën DS controlled by a joystick using pretty much the same setup (hydraulic valve to push the rack around, heart-shaped cam to apply resistance). Apparently it was very comfortable and natural to drive but ultimately a bit to weird even for Citroën.

Not a scrap of electronics in it, unless you count the pressure switch and dashboard lightbulb.

Re: Robots eat cars

#72

Post by taffydavid »

yalogin wrote: post #68

Building robots at that scale without any indication that the market wants it is weird. I wouldn’t want to say atupid because with musk there is no rational thought. However this is not cars where the concept exists and we know people spend 100k towards a car. We don’t know if people will even spend on a robot that doesn’t do shit. Figure is looking at 100-150k robot if built at scale, so u less they revised this est…

Well from this article I got the feeling the intended customer is industrial, not domestic. There's a lot of talk about how much a robot can lift versus hydraulic systems.

But I do also get the feeling that maybe Musk is just off his rocker and everyone else is copying what he does just in case he actually a genius

Re: Robots eat cars

#73

Post by lmm »

soopypoos wrote: post #65
SR2Z wrote: post #54

Earlier quoted context omitted.

> For the safety of any person being accidentally hit (the spring handles the retraction)

But if a motor is needed to counter the retracting spring, doesn't that mean any impact must overcome the motor before the spring will engage?

Presumably the motor runs once to extend it and then it locks into position (with some kind of mechanical catch that's calibrated to come loose if you hit anything), rather than being constantly running.

Re: Robots eat cars

#74

Post by BLKNSLVR »

throw310822 wrote: post #66
ipsum2 wrote: post #55

I read the title 2-3 times, and every time I swore it said "Robot cat ears". Wtf.

Then you complain about the Rs in strawberry...

That's exactly why we complain about the Rs in strawberry. We can get funny-stupid human interpretations all day long. What we can't get is cold facts, and isn't that what was promised by AI (at least, before ChatGPT was released in 2022)?

Unfortunately we fed this current iteration of AI with human behaviour (not only that: human behaviour on the Internet...)

Re: Robots eat cars

#75

Post by Modified3019 »

ipsum2 wrote: post #55

I read the title 2-3 times, and every time I swore it said "Robot cat ears". Wtf.

Literally same, saw “Robot cat ears”. Somehow I not only misread the words, but flipped the order as well, multiple times.

For those who need help recovering from the crushing self induced disappointment, here’s some brainwave controlled cat ears: https://www.youtube.com/watch?v=CqSZg0oiYuM

Re: Robots eat cars

#76

Post by psd1 »

mchusma wrote: post #51
vessenes wrote: post #44

Earlier quoted context omitted.

No, I'm saying that the original content is low-effort shitposting, and that Tesla has the ability to scale industrial production to over 1mm 'things' per year, as evidenced by production last year. I did the OP the mild courtesy of asking him to open up a useful conversation. For instance, "Is there going to be demand for 1mm robots, and if so, when?" Or "How much actual retooling is necessary in Fremont for this?"…

I think teslas issue is that they need the AI5 chip for robotaxi ops, the current chip just doesn’t cut it. So if they have batches end of 2026 and start optimizing the models, by mid 2027 volume production you might have robotaxis coming online at about 100k per month. Waymo currently has less than 10k cars on the road. Lots of ifs here. If they can enable hardware 4 for robotaxi ops then they can have 3m+ cars read…

My understanding is superficial, so do knock it down, not it seems to me that tesla insists on vision-only hour self driving, which vastly increases the requirement for ML. Whereas Waymo has a lower sum technology requirement by using both lidar and vision, and have moved faster. So when you say "tesla needs the AI5 chip", i hear the rider "...to avoid a public volte face".

I suppose that bulky lidar modules are undesirable in premium consumer goods, but i don't see that downside for taxis.

What am I missing?

Re: Robots eat cars

#77

Post by psd1 »

taffydavid wrote: post #72
yalogin wrote: post #68

Building robots at that scale without any indication that the market wants it is weird. I wouldn’t want to say atupid because with musk there is no rational thought. However this is not cars where the concept exists and we know people spend 100k towards a car. We don’t know if people will even spend on a robot that doesn’t do shit. Figure is looking at 100-150k robot if built at scale, so u less they revised this est…

Well from this article I got the feeling the intended customer is industrial, not domestic. There's a lot of talk about how much a robot can lift versus hydraulic systems. But I do also get the feeling that maybe Musk is just off his rocker and everyone else is copying what he does just in case he actually a genius

Musk has a very spiky character sheet. He is, in some dimensions, extraordinarily stupid, and I believe his ego makes a lot of big decisions. But something that might fall into the genius category is this: building things speculatively, primarily for the capabilities that you anticipate developing along the way, the nature of which are not yet known. But this increases your odds of having capabilities in the future that others lack, which looks a lot like a venture capital oeuvre.

To condense that, i might use a phrase like "blind-buying future option space"

Whether Musk deserves that credit is a moot point. I haven't trusted a thing he's said for years, and studying him for revealed intent can't get past "clown on drugs" without violating occam's razor.

Re: Robots eat cars

#78

Post by psd1 »

jillesvangurp wrote: post #41
Terr_ wrote: post #2

> Steer-by-wire Thinking back to case-studies around the Therac-25 [0], I would like to pre-emptively highlight the differences between: 1. Technique X is unsafe. 2. Technique X is unsafe because too much can go wrong even with the best intentions. 3. Technique X is unsafe without strong QA and interlocking safety measures, and there's too much economic pressure for the manufacturer to cut corners. [0] https://en.wik…

Safety is a great reason to not do something. Utility and enhanced safety are great reasons to override that reflex. A lot has happened since the Therac 25 incident in the medical world with AI, machine learning, robotic neuro surgery, all sorts of computer aided diagnostics, etc. This stuff undeniably saves lives. The incident did inspire some level of scrutiny of course. But compared to modern medical equipment, th…

All analysis should also keep in mind the "who", no matter how logically separable it is.

I put this less strongly since boeing contracted MBA cancer and yolo'd the 737-max, but that aside, the civil aviation engineering field controls risks to a fault. Commercial pilots are selected to follow checklists without deviation. I allow them the grace to implement steer-by-wire.

Ford kept selling Pintos with exploding fuel tanks, Toyota sold priuses with runaway acceleration defects, and depending on region maybe the worst twenty per cent of drivers ought to be operating nothing more dangerous than shirt buttons. No matter how good the plan is, those people shouldn't be anywhere near it.

Re: Robots eat cars

#79

Post by psd1 »

ErroneousBosh wrote: post #71
cyberax wrote: post #6

Earlier quoted context omitted.

Moreover, Technique X does not actually provide any significant value. The whole steer-by-wire in CT happened because Musk wanted a yoke as the control system. And a yoke requires progressive steering which is impractical without steer-by-wire.

> And a yoke requires progressive steering which is impractical without steer-by-wire. But also look at Citroën's DIRAVI system, used on the CX, SM, and some XMs. There's no direct mechanical link between the steering wheel and the rack when the system is pressurised. When you turn the wheel a kind of dogbone link thing pushes a spool valve one way or the other allowing hydraulic fluid to push the rack along, which p…

TIL! I love old Citroëns, but i never knew about that steering design.

Have you experienced that failure mode yourself? How alarming was it? Do you think it's a reasonable trade-off for the benefits?

Re: Robots eat cars

#80

Post by psd1 »

abcde666777 wrote: post #7

One million robots to be manufactured in a year - one million robots which will likely be obsolete within five years (if that, I wouldn't be surprised if they're dead on arrival). I don't know the figures for Earth's resources and their sustainability, so this may be a naive take, but I'm always left with the impression that these organisations want to speedrun the depletion of the planet.

You're anthropomorphising humanity.

Humans are sapient; industrialists and politicians have intent. But the incentive structure is an evolved system, and that's what selects these people. The result is that humanity, collectively, is amoebic. We are probably doomed to expand until we have a population crash or until another species arises to keep us in check.

Post reply on HN