Hacker News Reader

Comment by ftxbro | Hacker News Reader

> I believe you can't, at the same time, accuse OpenAI of benefitting of ideas mostly invented elsewhere, and also claim that they are in a predominant position.

There used to be a team at Google called Google Brain and they all left to go to OpenAI after the employee protests against taking military AI contracts in 2018. Now Microsoft has those contracts and funneled $10B to OpenAI from the CIA. OK that's a little bit of exaggeration but not so much; I guess not all employees left, and Google Brain still technically exists. Also some of the brain employees went to other startups not only OpenAI.

View on HN

Microsoft doesn’t have $10 billion from the CIA. They’re splitting military cloud contracts with Amazon and other startups, but it’s just the same thing as corporate contracts

Many of those employees have changed their minds about military contracts in this new cold war era.

America! What is in your military now, eventually ends up in the hands of the civilian police.

*and other countries are following. Fucking, sadly.

What the hell are they going to do with an M1 Abrams?

Having said that, I dread to see what the NYPD manage to achieve with an F-35B.

Trump tried to convince officials to use tanks to combat the Portland protests while he was president.

If I'm not mistaken their opposition was on principle not whether it was needed or not. So the fact that we are in a new cold war era does not change that Equation. The principle is still the same. The only way for this "change of mind" is if their opposition was due to "it's not needed because US has no rivals". Or what most probably happened, they realised that if they want to keep workign on this field there is no escape from those kind of implications and They don't have the big bad wolf Google to blame anymore.

And then Russia invades Ukraine and people change their minds about what is important, and what is possible.

But that's my point. If their opposition was not on principle but based on the naive Idea that "we will live in peace and harmony", I really am afraid what other naive principles are in the bases of their work and what safeguards are being set in place for the AIs.

There is a such thing as being principled and then finding out you were naive.

There is also the substantial effectiveness of propaganda, and the fact that it is essentially impossible to know if one's beliefs/"facts" have been conditioned by it.

That most discussions on such matters typically devolve rapidly into the regurgitation of unsound memes doesn't help matters much either.

Sure, but it is equally likely that your old perspective was the result of propaganda and your new one is a rational adjustment in the face of new information. Or that both are propaganda, I suppose.

That's the standard loop, but there are ways out.

ofcourse but then you still must have some principles and be as loud about how naive you were as you were when you were protesting the thing in the first place.

Why must you be loud about changing your mind? Why not just quietly realise you were wrong, have conversations with friends about it and move on? That’s what I’d do.

My life is a story. I’m under no obligation to share.

Not the op but the quotes from John Maynard etc made me think… if I listen to your original thesis and you convince me (maybe you have some authority) and I go on believing you, is it not harmful to me if you realize your error and don’t inform me? Boss: “why did you do X?” Me: “But sir you told me doing X was good!” Or to put it differently, if you spread the wrong word then don’t equally spread the correction then the sum of your influence is negative.

Maybe for someone you know in person. But the people who quit google in response to defence contracts aren't your boss. And they aren't your friends. You're a stranger to them. So its a bit of a long bow to draw to claim they owe you a followup.

It wouldn't surprise me if some of them did blog about changing their mind after the Ukraine war started. We must still have no idea because it just didn't make the front page anywhere.

Live in peace and harmony is not a good principle for AI?

When I was a young person, I used to deride writings like 1984 on the grounds that the scenarios and stories presented to carry the message were too far-fetched.

Reading your comment has set off an epiphany, I think I get it now, there probably exists some higher-up person who is thinking along these lines: we must always be in a state of war, for if we are, the populace will want to be ready and willing to fund the instruments of war. And we always want to be ready for war, because if ever we are not, we lose our capability to win a potential future war. For our very survival we must contribute our efforts to build these instruments of war. War is constant. War is peace.

The more powerful your military is the less likely you will need to use it to defend yourself.

From near-peer opponents, yes.

But it increases the temptation to use it against weaker opponents.

Offensive use of a military is a very different thing.

In the real world there are no clear cut boundaries like that so it’s military action either way.

Cool, so now they're gonna fight some elites' wars?

>There used to be a team at Google called Google Brain and they all left to go to OpenAI after the employee protests against taking military AI contracts in 2018.

This is blatant revisionism. Military contracts is not the reason Google has been losing AI researchers to OpenAI.

>funneled $10B to OpenAI from the CIA

Without some evidence to support it, this really sounds like a conspiracy theory.

I mean it is a conspiracy theory.

He did not mean it literally. Jeez.

Reply on news.ycombinator.com