Hacker News Reader

Comment by dahart | Hacker News Reader

Have you read the product page yet? It does advertise off-grid functionality. I don't know what it does when the servers go down, but the evidence I have appears to suggest they've thought about this, it is an adventure vest ffs, meaning it's expected to be offline at times, so the pure speculation and searching for a reason to continue believing they're wrong seems misplaced to me, fueled by an article that clearly misrepresented the situation.

View on HN

Sure, but I think you are just adding risk by adding an additional way your safety equipment can fail. If you have a way for a non-payment to disable the vest, you have a way for a software bug to disable the equipment.

There should be no way in software to make it not work.

I can totally agree with the general ideas about reducing risk and failure modes.

But with or without a payment switch, this is an electronically controlled airbag. It’s designed to have a way for software to keep it from inflating at all times, just like your car’s airbag has. It has a heuristic algorithm that tries to anticipate a crash based on sensor inputs, and the fundamental operation of the thing is that software is going to decide whether to inflate. Like your car, its default mode of operation is don’t inflate. The software switch is nothing new, nothing out of the ordinary, and you’re already relying on software switches daily.

There’s no evidence or reason to believe the payment system adds any additional risk beyond the fact that all airbags are already electronically controlled. It is possible for the payment system to be isolated so that it won’t affect inflation during a ride, and they explicitly stated their intent to do exactly that. So there’s no reason to think that non-payment can lead to failure during a ride, and there is evidence to suggest otherwise.

All the pure speculation in this thread about payment affecting safety was stirred up by the Vice article suggesting it, despite the product specs.

There's a difference between a black box without any user input that always does what you expect of it - and one with an on/off switch. Where the switch can be set via software - either the official one ... or some unofficial one.

That is true. I think the comment you replied to had already acknowledged that, but in case it wasn’t clear, yes of course there’s a difference. Does that difference contribute meaningfully to the conversation? Switches that can be set via software, for better or worse, are now the norm in vehicles, so why the concern about this vest? Software switches control ABS brakes, airbags, steering, and throttle in nearly all new cars sold today, and online features are mixed into many many brands. Whipping up incredulous concern about this completely optional motorcycle safety vest over software switches isn’t just speculation, it’s being willfully ignorant of today’s standard engineering practice.

Have you ever worked on vehicles? E.g. for Airbags, there's one single box whose whole purpose is to decide whether to pop the airbags or not. This box is separated from the rest of the car's systems and only puts out status messages via the CAN bus. And apart from cutting power to it, there's no way to disable this box - and thus the safety mechanism it provides. Also the software running there is verified by multiple people and certified to always do what it's supposed to do.

Same for the other safety features - there must be no way to disable them. Or take control of them from the outside. (Yes, there was a case where hackers were able to gain brake/steering control of a Jeep, but that's definitely not the norm.)

I think in the end it comes down to trust. Do you trust some startup from 2016 (In&Motion is the company behind the airbag module) to get the software 100% correct?

> Yes, there was a case where hackers were able to gain brake/steering control of a Jeep, but that’s definitely not the norm.

And that’s not because Jeep removed the remote control or the software switch, it’s because Jeep improved the network security. Now, there are a whole bunch more brands of cars that use the same remote access system Jeep is using. And there are a bunch of other systems in other car brands. How would you place a Tesla or other self-driving car into your black-box philosophy? The brakes and steering can run 100% of the time on software.

Are you sure KLIM’s airbag vest doesn’t have a CAN bus and a verification process?

I don’t trust Boeing or Toyota to get it right, so no I don’t trust a startup either. The point I’m making is not that I trust them, it’s that the vast majority of this comment section brought out pitchforks to rail against a straw man idea that Vice wrote about but does not match reality, and haven’t bothered to even read the product description that directly addresses things people here are speculating incorrectly about.

"There’s no evidence or reason to believe the payment system adds any additional risk"

Thats not how safety critical analysis is done. YOU have to prove thats it's physically impossible for the system to fail.

We already had boeing kill 300 people with crappy software and sensors in a plane.

It's only an adventure vest if you pay the extra $8/mo...

( https://www.klim.com/Ai-1-Airbag-Vest-3046-000 down near the bottom)

The cost is irrelevant to my point, which was it was designed to support adventure modes, which I would guess comes with design constraints because adventuring means off road and off grid.

Reply on news.ycombinator.com