Live data from Hacker News

US Air Force colonel ‘misspoke’ about drone killing pilot

theguardian.com

1–10 of 18 posts

Re: US Air Force colonel ‘misspoke’ about drone killing pilot

#2
This is a full retraction of the previously discussed story:

https://web.archive.org/web/20230602014646/https://www.thegu...

https://news.ycombinator.com/item?id=36159866

Seems like a big story to publish on the word of a single person (relayed through a blog post) with no other corroboration.

Shame on The Guardian for not mentioning the retraction/edits and simply reusing the same URL.

Re: US Air Force colonel ‘misspoke’ about drone killing pilot

#3
You can never tell with these things. Sometimes people exaggerate and get braggadocious, but sometimes they tell the truth, but it's too scary for consumption, to they say it was in error. It seems this scenario is plausible. Do you want an operator who goes rogue to sabotage a plan? On the other hand, you may want to cancel something as truth on the ground changes.

So, I can totally see them testing out different scenarios and making adjustments and maybe this protocol does not make it into production, but doesn't mean it wasn't tested.

Re: US Air Force colonel ‘misspoke’ about drone killing pilot

#4
Whether it was true or not the whole thing just doesn’t make sense. Why would the Air Force even code the simulated capability for the simulated drone to “kill” its pilot?

I swear journalists are way too credulous about this type of story. I just listened to a podcast episode about that which covers some of the more egregious reporting on “AI” lately [0]

0: https://citationsneeded.libsyn.com/episode-183-ai-hype-and-t...

Re: US Air Force colonel ‘misspoke’ about drone killing pilot

#5
post #3

You can never tell with these things. Sometimes people exaggerate and get braggadocious, but sometimes they tell the truth, but it's too scary for consumption, to they say it was in error. It seems this scenario is plausible. Do you want an operator who goes rogue to sabotage a plan? On the other hand, you may want to cancel something as truth on the ground changes. So, I can totally see them testing out different sc…

Things are worse yet in our post-modern post-WWII culture, where for 70+ years we've had an ever-growing schism between consensus reality and truth, thanks to the enshrinement of the premise that security requires secrecy, and anything of actionable value is stamped need-to-know.

And more! The interface between those who know objective truth, those who emit signals, and we who try to discern the former through the games of the latter, is now a state of permanent turbulence. When you say "testing out" that is even more true for the premise of sending up "trial balloons": you emit a message, observe the results, and then walk it back or obscure its veracity as needed. The actual truth is a footnote; the important thing is obscuring that from rivals (broadly understood as anyone who threatens the mission or at least the status quo, which has feature-crept to mean most of the public)...

In a case like this a cynic such as myself sees equal odds that what is reported in these two stories, is a decent rendition of the truth...

...and that the truth is elsewhere, e.g., the first statement was true; it was leaked either as a test or by actual misstep; and the machinery automatically kicked in to emit the second statement—because regardless of what is true, it is valuable for enemies to be unsure. Was this an accidental reveal of a capability far in advance of what the public knows? Or is this is an intentional bluff? Or...?

Who can say? None of us.

It bears saying,

the initial story was profoundly concerning in two different ways.

First because of the Cameron-ready simple existential threat posed by murderous paperclip maximizers, which most here are long familiar with.

But second, and much more interesting to me, and probably likewise to many here, is that the executive planning and scenario-finding implied by a a putative military AI, which is capable of concluding that taking the human out of the loop would best satisfy it,

...is suggestive of a type of AI very very different from what we see in the public sphere, or at least, what I have seen, following the horse race with some real attention.

That to me suggests that it's not even odds that this was a feint of some sort wrt AI and the foes of the West. As with e.g. well documented efforts to suggest to Cold War rivals that they were behind in the recovered-UFO-tech clandestine arms race, who knows what truth is in the basement, there is a lot of 9-to-5 life-and-death action and effort going on in causing sleepless nights over in the GRU and PRC.

Re: US Air Force colonel ‘misspoke’ about drone killing pilot

#6
post #3

You can never tell with these things. Sometimes people exaggerate and get braggadocious, but sometimes they tell the truth, but it's too scary for consumption, to they say it was in error. It seems this scenario is plausible. Do you want an operator who goes rogue to sabotage a plan? On the other hand, you may want to cancel something as truth on the ground changes. So, I can totally see them testing out different sc…

Things are worse yet in our post-modern post-WWII culture, where for 70+ years we've had an ever-growing schism between consensus reality and truth, thanks to the enshrinement of the premise that security requires secrecy, and anything of actionable value is stamped need-to-know. And more! The interface between those who know objective truth, those who emit signals, and we who try to discern the former through the ga…

Agreed on your first paragraph. We in the US literally don't even know how much we spend and how on defence. The pentagon has failed 5 audits in a row

https://oversight.house.gov/release/comer-sessions-open-prob...

Re: US Air Force colonel ‘misspoke’ about drone killing pilot

#8

Whether it was true or not the whole thing just doesn’t make sense. Why would the Air Force even code the simulated capability for the simulated drone to “kill” its pilot? I swear journalists are way too credulous about this type of story. I just listened to a podcast episode about that which covers some of the more egregious reporting on “AI” lately [0] 0: https://citationsneeded.libsyn.com/episode-183-ai-hype-and-t…

> Why would the Air Force even code the simulated capability for the simulated drone to “kill” its pilot?

Maybe I am mistaken, but I had thought lot of the excitement around current AI is the emergent behaviors that nobody specifically programmed for?

Re: US Air Force colonel ‘misspoke’ about drone killing pilot

#9
I don't think this kind of thing is bad anyway. First of all it was a simulation. This is why we test things, to find errors and fix them. I mean, if you build an automated thing meant to kill people then a lot of failure scenarios are going to involve killing the wrong people. It's just the nature of the game.

And the drone operator is at least a military target :) only the IFF logic needs some work. It'd be worse if the thing went all skynet and started killing all the (virtual) civilians it could find.

(Now having a mental image of a bunch of generals slowly backing away from a computer screen and looking to find the power plug to pull lol)

But this retraction sounds very political to me. The way politicians suddenly 'misspoke' after they get caught out on a lie.

Re: US Air Force colonel ‘misspoke’ about drone killing pilot

#10

Earlier quoted context omitted.

Things are worse yet in our post-modern post-WWII culture, where for 70+ years we've had an ever-growing schism between consensus reality and truth, thanks to the enshrinement of the premise that security requires secrecy, and anything of actionable value is stamped need-to-know. And more! The interface between those who know objective truth, those who emit signals, and we who try to discern the former through the ga…

Agreed on your first paragraph. We in the US literally don't even know how much we spend and how on defence. The pentagon has failed 5 audits in a row https://oversight.house.gov/release/comer-sessions-open-prob...

Even if you do get audits they'll have the $100.000 hammers to make things invisible. I think true knowledge of their expenditure is never going to happen.
Post reply on HN