Live data from Hacker News

OpenAI drops ban on military tools to partner with The Pentagon

semafor.com

131–140 of 509 posts

Re: OpenAI drops ban on military tools to partner with The Pentagon

#131

They haven't changed their principles page yet: https://openai.com/safety-standards ``` Minimize harm We will build safety into our AI tools where possible, and work hard to aggressively reduce harms posed by the misuse or abuse of our AI tools. ``` And I think they can still do that. If they corner every big usecase while they are the top ai models - they get to set the standard and define safety for all upcoming ai…

This is exactly the reasoning that every entity involved in doing this will say and is what leads to an arms race.

Re: OpenAI drops ban on military tools to partner with The Pentagon

#132
post #97

Earlier quoted context omitted.

I can tell you're not in Ukraine or Israel right now. Would you like to know how?

He’s talking about the US military. Why would they be in Ukraine or Israel?

Neither Ukraine nor Israel have reached this hypothetical technological era in which a draft is not required, and those facts are inescapable to any American in those countries observing it.

One reason the US military doesn't need a draft is because its proxy forces which are supplied by US equipment do have a military draft.

This is not a complex argument nor are there any facts which are not mutually agreed upon by all parties here.

I'm not able to follow the ideological contortions that would be required to not understand something as simple as:

"The US supplies the equipment, and the local US-backed government uses a draft to supply its army with the number of soldiers it needs to use the equipment."

Re: OpenAI drops ban on military tools to partner with The Pentagon

#133
post #94

The future of warfare is AI powered drones, zero doubt in my mind. Imagine the micro from Starcraft Go on a real life drone with a gun and bombs.

Imagine once they hallucinate and attack friendlies. Although if Machine Vision counts as AI then it's already widespread.

human soldiers already do that, if they can get it down to lower levels than humans it would be an improvement.

Re: OpenAI drops ban on military tools to partner with The Pentagon

#134

I was brought up in a household that was anti military. Only as an adult was I able to form my own opinion. Without a strong military in the past we would certainly not be here today and without a strong military now, who knows what other nation might decide to attack our sovereignty. It's childish to be anti military - it shows you do not understand the interconnectedness of our society - good and bad.

So you're in favor of every US government agency, no matter how flawed or harmful, as long as it theoretically serves some important purpose?

Marijuana criminalization is necessary and it's childish to criticize it, because without drug controls, cocaine would flood the streets.

Life imprisonment without chance of parole is necessary and it's childish to criticize any use of it, because some specific criminals merit it.

Government censorship of any kind of speaker for any kind of content is necessary and it's childish to criticize it, because specific types of information (classified national security information, for example) require it.

Even worse than taking this kind of children's storybook view of the world is applying it to the US military among all militaries, with an incredibly long undeniable track record of incompetence and war crimes.

Re: OpenAI drops ban on military tools to partner with The Pentagon

#135

Earlier quoted context omitted.

I personally welcome cooperation with the Pentagon for OpenAI and other American companies. The disdain that seems to infest so many when it comes to working with our national defense organizations is both annoying and bewilderingly naive and that's even when we take into consideration all of the bad things that happen. Strengthening partnerships with our government certainly doesn't make me mistrust OpenAI.

Say that again when they use AI to spy on you even more than they already do

That's a political problem, and requires a political solution.

Re: OpenAI drops ban on military tools to partner with The Pentagon

#136

Earlier quoted context omitted.

Unless you’re willing to help police or let your kids do it, I wouldn’t be so quick to forget the lessons of armchair patriot’ing wrapped up in saving the world mythologies as to give the US’s international military policy carte blanche, was the main lesson from Iraq. War sucks, but what sucks more is a public giving itself license to disengage bc of an easy moral justification. Approaches like this are why the AUMF…

The US is the most benign empire one could wish for. History makes that plain, and just peeking over the fence at contender's tendencies and track record further reinforces that. Easy moral justification backed by actual moral justification.

Sorry, I guess I'm wrong for thinking the empire that literally installed a brutal military dictatorship in my country is evil. It was actually benign the whole time. Please torture me, it's actually good to be tortured.

Re: OpenAI drops ban on military tools to partner with The Pentagon

#137
post #104

> "Fire!" > "I'm unable to help, as I am only a language model and don't have the ability to process and understand that."

It will "hallucinate" and claim it hit the target successfully when it didn't actually happen.

And in the event that a human being accidentally attacks a hospital[1], the AI will successfully argue that it didn't happen, as doing so would violate the acceptable use policy.

1: https://www.msf.org/kunduz-hospital-attack-depth

Re: OpenAI drops ban on military tools to partner with The Pentagon

#138

I was brought up in a household that was anti military. Only as an adult was I able to form my own opinion. Without a strong military in the past we would certainly not be here today and without a strong military now, who knows what other nation might decide to attack our sovereignty. It's childish to be anti military - it shows you do not understand the interconnectedness of our society - good and bad.

I could make literally the opposite argument. Other countries antagonize the US because of its military interventionism. If the US had lower military spending, other countries wouldn't feel the need to be aggressive. It's childish to be pro military. It shows you think arbitrary lines on a map determine who's good or bad.

Yeah, I'm sure Russia and China would be far less aggressive if the US had lower military spending.

What you're suggesting is the geopolitical equivalent of "ignore the bully and they'll get bored and go away". And you call others childish?

Re: OpenAI drops ban on military tools to partner with The Pentagon

#139

I was brought up in a household that was anti military. Only as an adult was I able to form my own opinion. Without a strong military in the past we would certainly not be here today and without a strong military now, who knows what other nation might decide to attack our sovereignty. It's childish to be anti military - it shows you do not understand the interconnectedness of our society - good and bad.

I could make literally the opposite argument. Other countries antagonize the US because of its military interventionism. If the US had lower military spending, other countries wouldn't feel the need to be aggressive. It's childish to be pro military. It shows you think arbitrary lines on a map determine who's good or bad.

That's the world view that my family came from - very leftist and peace loving.

"If only everyone put down their guns there would be no need etc etc".

But it's not the world we live in - we live in a military world - that's real and not a hope/fantasy about a different civilisation without war or weapons. We have to have a society suited to the world we live in.

And by the way I remain very much left wing and at the same time extremely pragmatic about the need for a nation to defend itself.

Re: OpenAI drops ban on military tools to partner with The Pentagon

#140

Slightly off-topic, but it's absolutely shocking to me how quickly OpenAI has managed to do a meteoric PR nosedive. I think within 6 months they've gone from one of the most exciting (yet controversial) non-profits in the world to mistrusted company that it seems most people are rooting against.

Has OpenAI done any PR/outreach explaining the recent changes? It'll be much harder (or impossible) to undo the mistrust after the fact.
Post reply on HN