Jiro

Karma: 4,636

Jiro 7 May 2024 17:51 UTC
4 points
2
in reply to: Rebecca’s comment on: Q&A on Proposed SB 1047
“This very clearly does not” apply to X and “I have an argument that it doesn’t apply to X” are not the same thing.

(And it wouldn’t be hard for a court to make some excuse like “these specific harms have to be $500m, and other harms ‘of similar severity’ means either worse things with less than $500m damage or less bad things with more than $500m damage”. That would explain the need to detail specific harms while putting no practical restriction on what the law covers, since the court can claim that anything is a worse harm.

Always assume that laws of this type are interpreted by an autistic, malicious, genie.)

Jiro 4 May 2024 22:30 UTC
4 points
2
on: Q&A on Proposed SB 1047
If your model is not projected to be at least 2024 state of the art and it is not over the 10^26 flops limit?

It’s not going to be 2024 forever. In the future being 2024 state of the art won’t be as hard as it is in actual 2024.
That developers risk going to jail for making a mistake on a form.
1. This (almost) never happens.
Because prosecuting someone for making a mistake on a form happens when the government wants to go after an otherwise innocent person for unacceptable reasons, so they prosecute a crime that goes unprosecuted 99% of the time.

The bill says the $500 million must be due to cyberattacks on critical infrastructure, autonomous illegal-for-a-human activity by an AI, or something else of similar severity. This very clearly does not apply to ‘$500 million in diffused harms like medical errors or someone using its writing capabilities for phishing emails.’

“Severity” isn’t defined. It’s not implausible to read “severity” to mean “has a similar cost to”.

Jiro 29 Apr 2024 20:56 UTC
4 points
2
in reply to: Viliam’s comment on: Losing Faith In Contrarianism

I guess in the average case, the contrarian’s conclusion is wrong, but it is also a reminder that the mainstream case is not communicated clearly, and often exaggerated or supported by invalid arguments.

This enables sanewashing and motte-and-bailey arguments.

Jiro 24 Apr 2024 20:17 UTC
2 points
0
in reply to: Garrett Baker’s comment on: Examples of Highly Counterfactual Discoveries?
I’ve heard, in this context, the partial counterargument that he was using traits which are a little fuzzy around the edges (where is the boundary between round and wrinkled?) and that he didn’t have to intentionally fudge his data in order to get results that were too good, just be not completely objective in how he was determining them.

Of course, this sort of thing is why we have double-blind tests in modern times.

Jiro 23 Apr 2024 2:27 UTC
2 points
0
on: Claude wants to be conscious
What happens if you ask it about its experiences as a reincarnated spirit?

Jiro 22 Apr 2024 15:53 UTC
3 points
0
on: Fruits of our Labors Introduction: The Art of Weirdness
I feel this conflates different kinds of weirdness, by using an overly vague definition and talking about cases where certain narrow kinds of weirdness are useful.

I couldn’t even come up with counterexamples because of the vagueness. Being rude to strangers is weird, but surely you’re not the only person who has done this, so you could argue “well, a lot of people do that so it’s not weird enough to count”. And then there’s reference class manipulation. “Yes, there was only one Unabomber, but he should be considered as a member of the class ‘violent political protests’ and there are too many of that class to count it as weird”.

Pointing to weirdness as good is like crackpots pointing to Galileo and Einstein. If you’re doing something weird, it gets bad reactions, and you blame that on the weirdness, it’s far more likely that you’re just trying to excuse some character flaw in yourself than that you’re the lone Einstein that nobody understands.

Jiro 19 Apr 2024 17:16 UTC
2 points
0
in reply to: Adam Zerner’s comment on: What’s with all the bans recently?
Features to benefit people accused of X may benefit mostly people who have been unjustly accused. So looking at the value to the entire category “people accused of X” may be wrong. You should look at the value to the subset that it was meant to protect.

Jiro 16 Apr 2024 21:53 UTC
10 points
2
on: Deontic Explorations In “Paying To Talk To Slaves”
Slavery is one subject that it’s highly likely ChatGPT is specifically programmed to handle differently for political reasons. How did you get around this problem?

Jiro 13 Apr 2024 16:49 UTC
2 points
0
in reply to: bhauth’s comment on: on the dollar-yen exchange rate
If they are, that link doesn’t show it. First of all, it doesn’t show Japanese prices at all. Second, even though it claims to “reflect restaurants of all sizes and segments”, it doesn’t, because a burger at McDonald’s or Wendy’s is not $16 and they obviously excluded fast food restaurants. How much is a burger in Japan if you exclude fast food?

Jiro 11 Apr 2024 8:15 UTC
2 points
−1
on: on the dollar-yen exchange rate

Yet, that’s not what happened; inflation has been higher in the US. In Japan, you can get a good bowl of ramen for $6. In an American city, today, including tax and tip you’d probably pay more like $20 for something likely worse.

I’d be unsurprised if you could get a jar of peanut butter or a turkey in the US for a lot less than you could in Japan. This tells you nothing about the economy.

Non-instant ramen (or peanut butter) is vastly more popular in one country than another. Comparing the prices between the US and Japan is trying to compare the prices of a specialty food and a common food. Of course the prices will be different.

Jiro 3 Apr 2024 8:56 UTC
15 points
7
on: Introducing Open Asteroid Impact
I’m going to be a party pooper here and point out that though this may be presented as an April Fool’s joke, its main joke is that in a live debate, it is extremely funny to strawman your opponent’s side. That’s bad practice whether done as a joke or not.

Jiro 3 Apr 2024 8:47 UTC
2 points
0
on: So You Created a Sociopath—New Book Announcement!
Many of these things are subject to the objection “and you know who else proclaims that they’re innocent? Innocent people.”

Advice which says “don’t act like you’re innocent, and be skeptical of someone who claims to be innocent” is generally bad advice.

Jiro 30 Mar 2024 23:02 UTC
8 points
5
in reply to: Muireall’s comment on: My Interview With Cade Metz on His Reporting About Slate Star Codex

It also pattern-matches to a very clumsy smear, which I get the impression is triggering readers before they manage to appreciate how it relates to the thesis.

It doesn’t just pattern match to a clumsy smear. It’s also not the only clumsy smear in the article. You’re acting as though that’s the only questionable thing Metz wrote and that taken in isolation you could read it in some strained way to keep it from being a smear. It was not published in isolation.

Jiro 30 Mar 2024 4:59 UTC
4 points
1
in reply to: tailcalled’s comment on: My Interview With Cade Metz on His Reporting About Slate Star Codex

I’m just arguing that there is a tension between common rationalist ideology that one should have a strong presumption in favor of telling the truth, and that Cade Metz shouldn’t have doxxed Scott Alexander.

His doxing Scott was in an article that also contained lies, lies which made the doxing more harmful. He wouldn’t have just posted Scott’s real name in a context where no lies were involved.

Jiro 30 Mar 2024 4:55 UTC
8 points
4
in reply to: tailcalled’s comment on: My Interview With Cade Metz on His Reporting About Slate Star Codex

Another is recipes for destruction, where you give a small hostile faction the ability to unilaterally cause harm. … But that seems less relevant for his real name, when it is readily available and he ends up facing tons of attention regardless.

By coincidence, Scott has written about this subject.

Not being completely hidden isn’t “readily available”. If finding his name is even a trivial inconvenience, it doesn’t cause the damage caused by plastering his name in the Times.

Jiro 30 Mar 2024 4:45 UTC
7 points
5
in reply to: LGS’s comment on: My Interview With Cade Metz on His Reporting About Slate Star Codex
The vague insinuation isn’t “Scott agrees with Murray”, the vague insinuation is “Scott agrees with Murray’s deplorable beliefs, as shown by this reference”. The reference shows no such thing.

Arguing “well, Scott believes that anyway” is not an excuse for fake evidence.

Jiro 30 Mar 2024 4:36 UTC
4 points
3
in reply to: Zack_M_Davis’s comment on: My Interview With Cade Metz on His Reporting About Slate Star Codex
I think some kinds of criticism are good and some are not. Criticizing you because I have some well-stated objection to your ideas is good. Criticizing you by saying “Zach posts in a place which contains fans of Adolf Hitler” is bad. Criticizing you by causing real-life problems to happen to you (i.e. analogous to doxing Scott) is also bad.

Jiro 28 Mar 2024 18:41 UTC
8 points
6
in reply to: Alex Vermillion’s comment on: My Interview With Cade Metz on His Reporting About Slate Star Codex
The reason that I can make a statement about journalists based on this is that the New York Times really is big and influential in the journalism profession. On the other hand, Poor Minorities aren’t representative of poor minorities.

Not only that, the poor minorities example is wrong in the first place. Even the restricted subset of poor minorities don’t all want to steal your company’s money. The motte-and-bailey statement isn’t even true about the motte. You never even get to the point of saying something that’s true about the motte but false about the bailey.

Jiro 28 Mar 2024 18:32 UTC
4 points
2
in reply to: Hazard’s comment on: My Interview With Cade Metz on His Reporting About Slate Star Codex

it seems really unlikely that he’s gotten any better at even the grammar of rationalist communication.

You don’t need to use rationalist grammar to convince rationalists that you like them. You just need to know what biases of theirs to play upon, what assumptions they’re making, how to reassure them, etc.

The skills for pretending to be someone’s friend are very different from the skills for acting like them.

Jiro 27 Mar 2024 14:11 UTC
9 points
−1
in reply to: Hazard’s comment on: My Interview With Cade Metz on His Reporting About Slate Star Codex

I understood the comment I was responding to as saying that Zack was helping Cade do a better job of disguising himself as someone who cared about good epistemics.

Yes, but disguising himself as someone who cares about good epistemics doesn’t require using good epistemics. Rather it means saying the right things to get the rationalist to let his guard down. There are plenty of ways to convince people about X that don’t involve doing X.