Hacker Newsnew | past | comments | ask | show | jobs | submitlogin

> there's no way to be sure that every second of GPS-tagged video shot by a DJI drone isn't going into a giant server farm owned by the Chinese intelligence service.

I checked your source and it didn’t back up this claim.

I’m not a networking specialist but isn’t it possible to detect if something is transmitting a massive amount of data (such as video) to an undetermined destination?

seems like this type of blatant data export would be easy to detect and subsequently ban the device doing it.

I’m just a simple software developer, so the network stuff can go over my head sometimes (heh), but the claim that such a large amount of data is being transmitted in a way that couldn’t be confirmed enough to ban the product seems dubious.



You do not need to transmit the video first. But coordinate or interested person. Even a sport watch can be a security threat as demo by some exercise army personnel. They just know who and when and where these places are.

For the transfer part it is much harder as said and easier to detect in peace time.

But then what happen in war or proxy-war time. You have to know how many senior Russians are killed by using a phone …


> Even a sport watch can be a security threat as demo by some exercise army personnel.

In the early days of the Afghan (or was it Iraq?) war, people used Strava to figure out the locations of American bases. They'd see a bunch of smart watches suddenly wake up in the morning and start exercising at the same time, a dead giveaway.


Recently a Russian commander got killed because he shared his daily running route on Strava which was seen by the killer.

https://edition.cnn.com/2023/07/11/europe/russian-submarine-...


I wonder if killing their dumber commanders might inadvertently be helping them…


Even steelmanning the argument, it probably doesn't. For one, commanders need some level of experience and training, you can't replace them for free. But more importantly, having a lower technically skill, which might open one up to inadvertently sharing their running route, probably does not correlate too much with the skills required to be a successful commander. Now, in general (hah), stupid people will probably both be more open to these kind of mistakes and be worse commanders, but being a good commander doesn't mean are not making minor opsec mistakes like this, so in the end you will still loose strategically valuable people.


We disagree on the severity of the mistake, hiding information from adversaries is a core competency and a big part of training. Someone who failed to learn from that training has a low innate intelligence and even if trained is unable to use the training effectively and is a liability.

The CIA for example drills into their people this same information denial training but they appear to neglect randomness so you end up with a bunch of people with the same peculiar behavioral patterns so they’re ironically rather easy to detect if you have access to click stream data. For example, they’re told not to follow each other on social media, but they still interact so you end up with two people who freely follow lots of people who interact with each other frequently but don’t follow each other. It’s weird, I’m not saying everyone who does this is information hiding but you can extract networks of people who behave in this same weird way with each other.

In order to protect the enigma cracking secret the UK randomly allowed themselves to be bombed with a bias towards less strategic targets. That’s the kind of thing you have to do to hide information, letting yourself be bombed should denote just how serious it is.


what a strange argument to make


Yeah, I'm serious, I'm not making a 'if you kill your enemies they win' argument but part of learning by doing is having the people who make poor decisions suffer the consequence of those decisions so that there are fewer people around making poor decisions. Especially in the military where you're spending other peoples money and other people suffer the consequences of your misadventures. Of course there is quite a lot of randomness in outcomes, but a blunder of this magnitude is inexcusable considering their line of work. Because of the corruption in the Russian army I would assume there is only a weak link between competence and rank and having an actual enemy around to punish mistakes would be helpful in winnowing out the morons. I'm pretty sure Russia knows they're corrupt and have deliberately adopted a learn by doing strategy to improve their warfighting capability for this very reason.

A big part of the process in undermining an opposition is promoting the worst aspects in them. Instead of killing off a moron, perhaps secretly encourage them to run for office and donate to their political campaigns, secretly buy media coverage for them, etc.


Arguments based on natural selection are sometimes unintuitive!


It's fair, but they've been working on improving the quality of their officers since 1904.


Don't know if you were attempting to make a joke, but Stalin purged the army of older, most qualified officers in the late 1930s, because they came from the pre-revolution times and were viewed as a loyalty risk. One of his biggest blunders that severely disadvantaged the country when the WW2 started.


Many white-era officers did serve in the Red Army; including Zhukov and his ex-boss Rokossovskij. Being a cadre in pre-revolution times wasn't the issue itself; willingness to sabotage their own country was.

Incidentally, the blunder in the 1941 in the Red Army was an issue of loyalty indeed. Navy didn't experience the same problem.



I mean, I assume it's mostly a joke, but if you assume that their system of selecting high officers isn't merit-based (which you would tend to assume given that it is Putin's Russia) then assassination which preferentially kills off the more incompetent officers would indeed be beneficial to the military as a whole.


Neither. The US’ wars in both places predate general availability “smart” devices by quite some time.

There were devices like black berries and tmobile sidekick, and gps trackers from garmin (etc) but I think the incidents you were referring to happened much later (2018 vs 2001/2002).


but the us only left Afganistan two years ago


The issue with the watch is the data is publicly available (social aspect) and requires no effort on the other side to figure out.


I believe the incident you’re referring to happened in 2018 [1].

Also happened across a few American bases across Syria, Yemen, Niger, Afghanistan, Djibouti and more. Some British and Russian bases were highlighted too.

[1] https://www.bbc.co.uk/news/technology-42853072



[flagged]


The point GP is making is that DJI does not indiscriminately upload every video. The moment it has the chance to associate a VIP to a drone, only then the tracking could start. It’s not like you can detect that during random tests on a new unit.


> there's no way to be sure that every second of GPS-tagged video shot by a DJI drone isn't going into a giant server farm owned by the Chinese intelligence service.

-GP

> The point GP is making is that DJI does not indiscriminately upload every video.

-You

That seems to be exactly what GP is claiming could be happening.

They were concerned about “every second of video” which would fit the definition of “indescribably”.


That's not a straw man. It's a valid threat assessment.

Lots of things get shut down for potential misuse.

One of the many jobs of the security apparatus is to predict which surfaces can be exploited, determine how bad those exploits could be, then firewall off the riskiest threats.


Given the volume of data, they can just box up the hard drives and ship via container ships or planes.


I assume he means at the device level.


I did. Thanks!


I don't know how DJI works, but presumably it ships the video out to a service that you then log in to to view?

If so, it's on their servers and there's no "networking" you can do to know whether they forwarded it on from there.


"I don't know" - so why leave a comment?

Because you can doesn't mean you have to, especially when you don't actually have any knowledge on the topic at hand. (And, as people have pointed out, this is both a weird and incorrect assumption, adding nothing to the discussion other than confusion.)


Except that they do offer exactly that. No need to be so condescending.

https://www.dji.com/lightcut


There's certainly no requirement to use that (it's literally a separate app from the DJIFly app you use the control the drone), and I do know as I fly a DJI mini drone.

What's worse is that I don't think it actually works how you've assumed. The drone has no internet connection itself, the software is instead pulling the files down over wifi, which is something the standard software supports if the drone is close enough. This is the "no need to export from your DJI device" - LightCut can presumably access the drone files directly. None of this requires uploading the videos to anywhere, and doing that wouldn't even make sense - these are large video files, people would notice their data plans being ravaged by multi-gig uploads every time they flew their drone.

As far as being condescending, I think that's less of a negative trait than offering unbacked "I don't know, but" comments which add no value to the discussion.

I did consider the value of my own comment at the time, but I think there is a big problem in tech discussion with people with no actual experience or relevant knowledge feeling that their off-the-cuff suppositions are as welcome and useful as meaningful input from people with direct experience, and highlighting this behaviour as negative and unwelcome is worth risking the inevitable backlash in response.


>What's worse is that I don't think it actually works how you've assumed.

Holy irony man


Yes, based on my experience with the device I can extrapolate how that software could work. What experience did you base your assertions on? Nothing at all? Right.


15 years of working on SaaS products and a "how I would build it" guess. So basically exactly what you're doing.


The DJI Mini Pro 3 works without Wifi and using the expensive controller, without a phone app, so the opportunities to upload captured data to China are very limited.


Unless you use their recommended software. https://www.dji.com/lightcut


What a strange assumption to make. No it doesn't work like this at all. The video is saved to the SD card in the aircraft, which you then remove and insert it into your computer to download the files.


And potentially use their recommended software to work with from there: https://www.dji.com/lightcut

So not that many jumps from what I described.


If that’s true then you’re right but I don’t know if that’s true.




Guidelines | FAQ | Lists | API | Security | Legal | Apply to YC | Contact

Search: