←back to thread

614 points nickthegreek | 7 comments | | HN request time: 0.001s | source | bottom
Show context
trinsic2 ◴[] No.39122579[source]
Based on everything I am hearing about all the harmful uses this tech could have on society, i'm wondering if this situation is alarming enough to warrant an inquiry of some kind to determine whats going on behind the scenes.

It seems like this situation is serious enough that we cannot let this kind of work be privatized.

Not interested in entertaining all the "this is the norm" arguments, that's just an attempt at getting people to normalize this behavior.

Does anyone know if the Center of AI Safety acting for the public good and is this on their radar?

replies(3): >>39122627 #>>39123186 #>>39123521 #
JumpCrisscross ◴[] No.39122627[source]
> wondering if this situation is alarming enough to warrant an inquiry of some kind to determine whats going on behind the scenes

OpenAI is making people rich and America look good, all while not doing anything obviously harmful to the public interest. They’re not a juicy target for anyone in the public sphere. If any one of those changes, OpenAI and possibly its leadership are in extremely hot water with the authorities.

replies(2): >>39122752 #>>39123545 #
trinsic2 ◴[] No.39122752[source]
> all while not doing anything obviously harmful to the public interest.

Yeah, gonna have to challenge that:

1. We don't really if what they are doing is harming public interest, because we dont have access to much information about whats happening behind the scenes.

2. And there is enough information about this tech that leads to the possibility of it causing systemic damage to society if its not correctly controlled.

replies(3): >>39122763 #>>39122807 #>>39124666 #
danielmarkbruce ◴[] No.39122807[source]
You don't have access because you aren't supposed to. Nothing about the founding, laws or customs of the US suggest that you (nor the government itself) have access to information about other people/companies any time you/they feel like "finding out what's happening behind the scenes".

As for "too important to privatize"... practically all the important work in the world is done by private companies. It wasn't the government who just created vaccines for Covid. It isn't the government producing weapons for defense. It's not Joe B producing houses or electricity or cars or planes. That's not to say the government doesn't do anything but the idea that the dividing line for government work is "super important work" is wildly wrong and it's much closer to the inverse.

replies(3): >>39123315 #>>39124103 #>>39133059 #
romwell ◴[] No.39123315[source]
>practically all the important work in the world is done by private companies

LOL, another one thinks the US is the entire world.

replies(1): >>39123422 #
danielmarkbruce ◴[] No.39123422[source]
The comment about access is related to a US company. The relevant legal jurisdiction and framework is the US. If it were a French company, the relevant jurisdiction would be... France. You may not realize this, but OpenAI is a US company.

The comment about all the important work in the world being done by private companies was indeed a global comment. You may not realize this, but covid vaccines were made by astrazeneca (UK), BioNTech (Germany), several US companies and others. Defense companies are located in every major economy. Most countries have power systems which are privately owned. Commercial planes are mostly built by one large French company and one large US company. All the large producers of cars around the world are private companies - big ones exist in the US, Japan, various European countries, Korea and China.

replies(1): >>39127509 #
romwell ◴[] No.39127509[source]
It looks like you have a problem understanding the meaning of the word "all".

Specifically, you are confusing all and some.

replies(1): >>39130514 #
danielmarkbruce ◴[] No.39130514[source]
"practically all" was the initial statement. The second mention is a reference to that statement.

A basic education in economics (and manners) might be in order.

replies(1): >>39133263 #
romwell ◴[] No.39133263[source]
>A basic education in economics (and manners) might be in order.

Oh, I apologize. I have idiotism allergies.

>"practically all" was the initial statement

Yeah, and my original statement still stands.

Go open a history book or something. Or find out what "N" in "NSF" stands for, or realize that China exist.

replies(1): >>39137312 #
danielmarkbruce ◴[] No.39137312[source]
NSF budget is 10 bill or so out of a 20+ trill economy, less than a tenth of a %. Even if you thought 90% of work wasn't important... 10 bill out of 2 trill is tiny.
replies(1): >>39138887 #
1. romwell ◴[] No.39138887[source]
So you measure importance of work by the amount of money spent on it?

My do you have some peculiar ideas.

replies(1): >>39139022 #
2. danielmarkbruce ◴[] No.39139022[source]
Re-read it. I'm measuring the amount of work by the amount of money. Total GDP = total work. 10% of GDP = 10% of total work. GDP is a decent proxy for total work... Add reading comprehension 101 to econ 101 and then it will compute.
replies(1): >>39140357 #
3. romwell ◴[] No.39140357[source]
Re-read what you wrote yourself.

You were using the NSF budget as a measure of importance of its work.

You also compared it to the GDP, which only makes sense if you think that all work done is equally important. How very socialist of you.

>Add reading comprehension 101 to econ 101 and then it will compute

What you say "computes" only after adding Dunning-Kruger to the mix.

replies(1): >>39143861 #
4. danielmarkbruce ◴[] No.39143861{3}[source]
Nope, the calcs above assume all NSF work is important, and they show how little work they do as a % of all the work. Then as a % if we assumed 90% of all work (keeping all NSF work important) wasn't important.

To calculate a %, both the numerator and denominator have to be the same units.

replies(1): >>39148941 #
5. romwell ◴[] No.39148941{4}[source]
> and they show how little work they do as a % of all the work. Then as a % if we assumed 90% of all work (keeping all NSF work important) wasn't important.

Great, you're simply saying that pretty much all of science has the same importance as 10% of all other work being done. And you consider budget as a measure of output.

All that in the context of a conversation about technological breakthroughs, mind you.

By that metric, someone like Richard Feynman has produced less important work than your average run-of-the-mill engineer with a slightly higher salary.

Did you time-travel here from the USSR? The leadership there had similar ideas back in the day.

This is becoming very entertaining at this point.

replies(1): >>39150099 #
6. danielmarkbruce ◴[] No.39150099{5}[source]
Econ 101 is your friend.
replies(1): >>39152698 #
7. romwell ◴[] No.39152698{6}[source]
It is. So is the reality.

Highly recommend, A+++, 10/10.

"Impact is hard to measure, so let's take budget as a proxy" has got to be the hottest take of the year, and yes, I'm aware it's January.