empthought

Podcasts (and video) are very low-throughput, low-density information channels. Essays and articles are superior. To demonstrate this, you can just compare the transcript of a typical podcast — even a high-quality, well-researched one — with a typical high-quality, well-researched blog post, essay, or journalistic article.

You can make up different ones that no one cares about if you like. This thread is about a particular thought experiment that is nothing like the one you posted.

The only way to find out would be for this thought experiment to actually occur, in which case I would do nothing, because I would be dead.

Of course not.

Pretty sure there’s some stolen valor here — forward deployed engineers are disarming IEDs, breaching fortifications, and so on.

Not snowing customers with slop.

You didn’t even bother reading what I wrote.

For me, blue is the purely selfish choice. Because I have no interest in living in a world where red wins a majority on this one. It’s not even a tough call.

You read a bunch of things into what I wrote that simply are not there. I like plenty of people who would press the red button without hesitation. Their decision is their decision and my decision is mine. Might they “have a bad time?” Maybe. Maybe not. I’m certain I would have a terrible time, if I joined them in the majority.

I have no idea what you are referring to. This isn’t a question of ingroup or outgroup.

Then I’d say you either really deserve the world you get when red wins, or you don’t really deserve the world you get when blue wins.

It doesn’t guarantee my safety, though. It puts me in a world where the only people left were so afraid of dying that they opted into a completely avoidable mass murder.

> this is irrational and doesn’t likely reflect what people would actually do

People are irrational. I guarantee it’s likely that a lot of people would actually do it.

> The idea behind claiming you’d choose the blue button is to appear noble and altruistic.

Not at all. If the majority of people can’t be bothered to press the button that says “nobody dies as long as half of the other people say nobody dies” rather than “you don’t die,” I’m happier not being around. It’s purely selfish and blue is a win-win.

Of course the framing affects how people vote. The thought experiment demands we use the framing as given. Some people might reason themselves into your analogy, others won’t.

It is rephrased. Any phrasing besides the original exists only in your head and not in anyone else’s.

The only reason I’m in my tomb is because you and people like you voted to kill me instead of voting to do nothing. Luckily for me, I’m dead and don’t care.

Congratulations! Enjoy your life with people who think like you.

You don’t get to reframe the problem with different wording or circumstances to demonstrate your intelligence to others before they choose and you choose. That’s part of the thought experiment.

Yes, when you completely rephrase the problem you will have different sentiments.

The thought experiment demands that the phrasing that was used actually be used, and you don’t get a chance to show the dumb blue people how smart you are before they pick their button.

Yes, but those are different thought experiments from this one.

This one explicitly is not.

One button could kill you — if and only if enough people press the other button.

The other button certainly won’t kill you, but will kill everyone who pressed the first button — if and only if enough people besides you press it.

That’s not the thought experiment.

These types of analyses always treat voting red and ending up in the majority as preferable to voting blue and ending up in the minority. I don’t think that assumption is universally valid.

In the first case you contend with living in a world after a catastrophic population loss — likely including at least some of your loved ones — knowing that had some of you and your fellow survivors voted differently, nothing bad would have happened.

In the second case, you don’t care at all. Because you’re dead.

> Why does it need to say things to itself like “great I have a plan now!”

How else would it know whether it has a plan now?

Your question was about calling for violence.

Yes.

If you want a molotov cocktail thrown so badly, throw it yourself. Don't put it on other people to do it for you.

Of course! If it were interesting, they wouldn't need to pay you. You would pay them!

Yes, I wasn’t clear.

I was offering a reason why dollars “losing value” is not actually a thing to worry about. If they didn’t lose value, people would hold onto them instead of spending them. Which is objectively a loss to everyone, regardless of how you decide to denominate it.

Dollars that people hold onto, instead of using to purchase something else of value, are effectively worthless to everyone.

No one is talking about AGI in this thread except you, though. The post said nothing about it. It's an absolute non sequitur that you brought up yourself.

I said I don’t think it follows, and you certainly gave no support for the idea that it must follow. Logically speaking, it’s possible for improvements to continue indefinitely in specific domains, and never come close to AGI.

Progress in LLMs will not slow down before they are better at programming than humans. Not “better than humans.” Better at programming. Just like computers are better than humans at a whole bunch of other things.

Computers have gotten steadily better at adding and multiplying and yet there is no AGI or expectation thereof as a result.

I don’t think that follows, nor do I think it will keep improving indefinitely. It will certainly continue to improve for a while.

We don’t need anything close to AGI to render the job “software engineer” as we know it today completely obsolete. Ever hear of a lorimer?

I know some proponents have AGI as their target, but to me it seems to be unrelated to the steadily increasing effectiveness of using LLMs to write computer code.

I think of it as just another leap in human-computer interface for programming, and a welcome one at that.