Hacker News Reader

Comment by unyttigfjelltol | Hacker News Reader

A phone app that did this would be malware. When Alexa does it, this is a feature?

You mean, like AirTag/Find My?

I think that Amazon's marketing for this is horrible, reminds me of XBox' failed marketing around their online-only plan for XBox One game purchases [1].

But really all they've done is built AirTags as a platform rather than an application, e.g. Tile is supported by Amazon's new platform.

I don't want to give random people with random apps any amount of arbitrary bandwidth; I'd rather that users of the Sidewalk platform would have to register app specific message types publicly (to ensure transparency) and with Amazon (who should enforce message type and payload restrictions).

[1] https://arstechnica.com/gaming/2013/06/xbox-one-eighty-micro...

> You mean, like AirTag/Find My?

Can you explain your comparison?

Any AirTag owned by anybody can and will communicate with your iPhone over Bluetooth if you have the right features enabled (which are default). Your phone will contact Apple and tell them about the interaction. I haven’t seen protocol details, but last week a security researcher demonstrated that it is possible to use the AirTags protocol as a covert channel (with very low bandwidth).

not great, but not the same as open connectivity to unknown devices.

No, it isn’t a great comparison. But it is a similar slippery slope.

Amazon Sidewalk also is not "open connectivity".

I recently read up on Bluetooth low energy. The AirTags are most likely periodically broadcasting announcement packets, and folks' iPhones are passively receiving them. It therefore seems comparable to using 802.11 SSID beacons for geo-location, except that folk intentionally bought the AirTag to be position tracked.

Just by modulating whether or not the AirTag is transmitting, it seems like you could communicate at a low data rate. You could even use hundreds or thousands of AirTag IDs to get a useful data rate. Just because you could do it, though, it seems like there are plenty of other more practical covert channels.

> I don't want to give random people with random apps any amount of arbitrary bandwidth

It uses a maximum of 500 megabyte per month.

Regarding payload type registration – it seems like it would be pretty difficult to monitor that, given that payloads are encrypted.

That actually seems more like a feature than a problem to me: If my IP address does not show up in any logs and I can't even see what's being transmitted, I really doubt there would be any liability for the user here.

Wait till your IP address gets banned from something for being a source of spam

Sidewalk does not allow generic access to the Internet. It connects to Sidewalk servers only.

Source: https://www.amazon.com/gp/help/customer/display.html?nodeId=...

Or a crime gets committed via your IP address and you get thrown in jail for it.

500 megabyte a month is a lot if you are loving in a relatively poor country.

Are people in poor countries likely to own a lot of Amazon devices?

That's a fair point, and I do hope that regional caps will take that into consideration.

Generally speaking, are volume caps on fixed line internet a thing in many countries? I'd imagine that it would be hard to use something like a smart speaker on such a tight data budget in the first place, as I can't imagine them to be very bandwidth-conscious in general (music streaming having no "low data usage" option on them usually, etc).

It would be trivial to design a protocol that allowed client side contact inspection.

It would be possible to design an encrypted protocol that allowed Amazon server side content inspection.

What I've noticed is that the blackhat behaviors of the past have become business models.

I remember when going around logging locations of wifi access points was "wardriving". Now apple and google do it as a matter of course using everyone's phones.

You mean like HolaVPN or Geosurf the “free VPN” applications that quite literally install a backdoor proxy on your phone or PC? The parent companies then sell access to the proxies running on your devices as “residential proxies” and also sell access to your browsing history?

That sounds pretty close to malware to me. Are you saying they aren't?

I absolutely think that qualifies as malware, yes. Apps and browser extensions harvest PII all day and the owners of those platforms (apple, google, microsoft, mozilla) do absolutely nothing to prevent it.

I mean, is it just the obfuscation of the privacy policy? If the user was told up front would that still make it malware?

Silent auto-update would also be normally classified as a deliberate RCE vulnerability or a backdoor, but for some reason most people are okay with that.

Corporate broad spectrum wire tapping has been greenlit legally for a long time. An employer can listen to private and public communications that occur on and off of company owned resources. This includes:

- emails

- text messages

- microphone arrays

- shared networks

- keyboard input

- file storage

- contact and communication history

It's a logical next-step from there to expand these services under the guise of subscriptions. That's why you owning a device is problematic for them. That said, owning the platform is the next best thing because they can now do many but not all of those things in the name of protecting their platform from "abuse".

Sources:

https://www.jstor.org/stable/27673175?seq=1

https://www.npr.org/templates/story/story.php?storyId=123024...

Reply on news.ycombinator.com