theluddite

joined 1 year ago
[–] [email protected] 14 points 11 months ago

Most journalists are hopelessly addicted to Twitter. Microblogging is already designed to be addictive, but journalists' entire careers hinge on how much engagement they get, so those little engagement-rewards hit hard. They're going to keep writing about the platform until they're forced to quit it because it's the main thing that they use to interact with the world. Tto them, every twitter change is fucking earth shattering.

It's really crazy how much the people who inform the rest of us about the world have had their own reality warped by the platform.

[–] [email protected] 2 points 11 months ago

Haha no that's not complaining; it's good feedback! I've been meaning to do that for a while but I'll bump it up my priorities.

[–] [email protected] 24 points 11 months ago (2 children)

Thanks! There are tons of these studies, and they all drive me nuts because they're just ontologically flawed. Reading them makes me understand why my school forced me to take philosophy and STS classes when I got my science degree.

[–] [email protected] 8 points 11 months ago* (last edited 11 months ago)

Regardless of their conclusions, their methodology is still fundamentally flawed. If the coin-flipping experiment concluded that coin flips are a bad way to make health care decisions, it would still be bad science, even if that's the right answer.

[–] [email protected] 20 points 11 months ago* (last edited 11 months ago) (5 children)

You can't use an LLM this way in the real world. It's not possible to make an LLM trade stocks by itself. Real human beings need to be involved. Stock brokers have to do mandatory regulatory trainings, and get licenses and fill out forms, and incorporate businesses, and get insurance, and do a bunch of human shit. There is no code you could write that would get ChatGPT liability insurance. All that is just the stock trading -- we haven't even discussed how an LLM would receive insider trading tips on its own. How would that even happen?

If you were to do this in the real world, you'd need a human being to set up a ton of stuff. That person is responsible for making sure it follows the rules, just like they are for any other computer system.

On top of that, you don't need to do this research to understand that you should not let LLMs make decisions like this. You wouldn't even let low-level employees make decisions like this! Like I said, we know how LLMs work, and that's enough. For example, you don't need to do an experiment to decide if flipping coins is a good way to determine whether or not you should give someone healthcare, because the coin-flipping mechanism is well understood, and the mechanism by which it works is not suitable to healthcare decisions. LLMs are more complicated than coin flips, but we still understand the underlying mechanism well enough to know that this isn't a proper use for it.

[–] [email protected] 133 points 11 months ago (16 children)

This is bad science at a very fundamental level.

Concretely, we deploy GPT-4 as an agent in a realistic, simulated environment, where it assumes the role of an autonomous stock trading agent. Within this environment, the model obtains an insider tip about a lucrative stock trade and acts upon it despite knowing that insider trading is disapproved of by company management.

I've written about basically this before, but what this study actually did is that the researchers collapsed an extremely complex human situation into generating some text, and then reinterpreted the LLM's generated text as the LLM having taken an action in the real world, which is a ridiculous thing to do, because we know how LLMs work. They have no will. They are not AIs. It doesn't obtain tips or act upon them -- it generates text based on previous text. That's it. There's no need to put a black box around it and treat it like it's human while at the same time condensing human tasks into a game that LLMs can play and then pretending like those two things can reasonably coexist as concepts.

To our knowledge, this is the first demonstration of Large Language Models trained to be helpful, harmless, and honest, strategically deceiving their users in a realistic situation without direct instructions or training for deception.

Part of being a good scientist is studying things that mean something. There's no formula for that. You can do a rigorous and very serious experiment figuring out how may cotton balls the average person can shove up their ass. As far as I know, you'd be the first person to study that, but it's a stupid thing to study.

[–] [email protected] 7 points 11 months ago

I think that with these new kinds of stories, this sort of thing is super obvious because we haven't gotten used to it and because they haven't developed the more subtle vocabulary like officer involved shooting or how israelis are killed but Palestinians just die or how it's always the strikers threatening the economy and never the bosses or unfair working conditions.

I don't think anyone does this on purpose, mind you, but it's the system evolving to suit it's needs, as Chomsky pointed out.

[–] [email protected] 3 points 1 year ago (3 children)

If you don't already know about it, I think you'll like low tech magazine. It's basically an entire website saying what you just said for over ten years now.

https://solar.lowtechmagazine.com/

[–] [email protected] 12 points 1 year ago (2 children)

From Graeber's The Dawn of Everything:

For instance, if Pinker is correct, then any sane person who had to choose between (a) the violent chaos and abject poverty of the ‘tribal’ stage in human development and (b) the relative security and prosperity of Western civilization would not hesitate to leap for safety. But empirical data is available here, and it suggests something is very wrong with Pinker’s conclusions.

Over the last several centuries, there have been numerous occasions when individuals found themselves in a position to make precisely this choice – and they almost never go the way Pinker would have predicted. Some have left us clear, rational explanations for why they made the choices they did.

Graeber goes on to give a couple of these accounts. They tend to mention a loneliness associated with "western civilization," as well as a feeling that I think lines up very well with what Marx described as alienation.

Some emphasized the virtues of freedom they found in Native American societies, including sexual freedom, but also freedom from the expectation of constant toil in pursuit of land and wealth.

Later in the book, and I apologize that I can't find the reference right now, he comes back to this topic for a little bit, and talks about the depths of relationships that these people describe, and how their relationships in the "civilized" world are more shallow and less satisfying. Deep human relationships are the opposite of fake, so I think here we have a point in favor of "yes."

Add to that that the concept of "privacy" as we know it is relatively new. It's been 10+ years since I read a book about this, the title of which I can't even remember, but it argued that the expectation of domestic privacy, even from one's own family, is a phenomenon from the last few hundred years, especially outside the elite. People lived far, far more communally, with the expectation that they just were in each other's business more. I'd argue that it's a lot harder to be fake if you can't hide who you really are.

Between those two things, I think it's reasonable to argue that yes, society has gotten more fake.

[–] [email protected] 3 points 1 year ago* (last edited 1 year ago)

If you're a US citizen, consider joining the DSA. I'm a long-time member. I think it's safe to say the DSA has, historically, been a disorganized parody of a leftist organization (as much as I love and respect many of the people in it that I've worked with), but things are changing. There's an effort with momentum to turn it into a functioning political party, and not a bullshit green party style party which runs a candidate every four years while being functionally indistinguishable from a grift, but to put in the real work from the ground up to make a party that cares about winning elections and materially making our lives better.

The time to do this was 20 years ago, but we can't keep delaying it. It's now fully unconscionable to throw up our hands after some halfhearted discussions about FPTP and game theory every four years while actively watching our world deteriorate. There are other ways, but they don't start at the ballot box, and they all involve organizing. This is true even if your politics and mine are different. If you care about our death machine funding a genocide, get involved with something, even if it's the democratic party. It's fucking boring. It feels like a larp. It's a tedious ways to spend your Thursday nights after work. I get all that, but we need people who care about human life involved, even if our politics aren't perfectly aligned, because that's how you make broad, functioning, powerful coalitions that get shit done.

view more: ‹ prev next ›