The Grey Tribe

It increased our error rate while reducing the overall ticket count by two. And redistributed numbers.

It smoothed out our data trend, erasing the actually important-to-report information (that big dip and increase is important):

It erased my security and networking practices and came up with resubmitted tickets out of nowhere and shows an overall downward trend in ticket counts when only one of the five categories is actually trending down.

On another slide it also changed the status of a device from "attention needed" to "healthy," and on yet another slide it removed the captions that I'd grouped with a set of images.

This is your reminder that the robots are stupid and if you work with them you need to check their work.

I don't say this because I think that YOU should be using AI for stuff (and if you hate AI, I get it, I don't want to use it either) I say this because one of the absolute best ways I've found to get people to step back from believing everything the AI tells them is to ask the AI a question about something the person you're speaking to is an expert on, or ask it something that YOU are an expert on and point out where the AI's answer is wrong.

rather than try to keep this from ever happening, i have simply set my heart on raising her to be a main battle tank, or failing that, possibly a telecommunications satellite. i point out that she is already mostly achieving one of the core requirements for the latter, being approximately geosynchronous most of the time.

and yes, i have on multiple occasions had conversations like:

kid: ow, that hurt seebs: sounds like something someone without six inches of tempered steel plating would say

1. Were your poems any good? I mean, looking back now, (A) do you still stand by them, or (B) do you immediately see obvious flaws.

You already answered my question #2. You think (A) the publisher humoured you and decided to publish you because you were 11. I assume they would not (B) have published your poems if they thought they had been written by a 50 year old high school English teacher.

I also assume they didn't tell you they were humouring you.

3. Did other adults in your life take you seriously? I mean (A) were they genuinely impressed or did they (B) know it was political and they were just joining in and humouring you, too?

4. Did somebody (A) explain to you that this was political and not about you, or did you (B) understand this only later, as an adult.

For me it's 1B 2B 3B 4A. People acted impressed, but they clearly didn't take me seriously, because I was just a child. Somebody took me aside and explained it was all political.

the neo-reactionary movement has already given us the Zizian murder cult, the FTX fraud and bankruptcy, Curtis Yarvin’s whole deal, and the worst fanfic that can exist under currently understood laws of nature.

I sense a certain preoccupation on the side of the writer, or dare I say a partisan bias! On a positive note, this time nobody suggested that the fanfic existed first, and the other stuff grew out of that, so that's a positive.

Craft, community, it's all one thing! Neoreaction, effective altruism, Bayes' Theorem: If you ever thought how to best spend your money on charity or if you ever used a spam filter before LLMs, you're basically Peter Thiel.

Just like rationalists in other fields

The accusation is clear: Rationalists are by definition not domain experts. They are autodidacts, outsiders, MENSA types who haven't properly taken the cursus honorum and instead just barge into established fields with nothing but spreadsheets and first-principles thinking.

I think this isn't quite true. There is a lot of academic support for ideas like effective altruism, bayesian inference, compatibilism/materialism, or the Many-Worlds Interpretation.

I also happen to think there is often merit to "barging into" other fields, such as that time when C++ programmers barged into Neil Ferguson's CovidSim. Yes, they weren't epidemiologists. Yes, they were providing ammo to antivaxxers, and some antivaxxers knew C++ and had a field day. But you can't just say "I will guard my cherished conclusions until somebody with the right credentials produces a proof or counterexample, any proof or counterexample or empirical data first produced by an outsider doesn't count!"

That piece just nutpicks, but so can I! I can nutpick and sanewash witht he best of them.

I pulled the story once I started feeling like it simply wouldn’t be possible for me to publish a version with NYMag that didn’t carry a subtle hostility towards Leverage, not to mention affiliated communities in Silicon Valley — and, more importantly to me, hostility towards a core spiritual sensibility that I see in both myself and in the people the story describes.

(Emphasis mine)

I find this relateable and understandable. You can't just tell your story without somebody trying to use it as ammo in the culture wars, trying to draw much broader conclusions about a much broader group of people!

There was a time in the 90s and early 2000s when the tabloid journalist planning a hit piece was a stock character, but when newspapers started dying, it wasn't just journalists who circled the wagons. It became culturally distasteful to say "don't believe everything you read in the papers".

Then Geoff posed a thought experiment: What if Tyler took a pill, then started floating off the ground, and touched down five minutes later — then would Tyler feel that he needed to use a scientific tool in order to trust his own observations? What if Tyler took another pill from the same jar, and the second time he took the pill he floated off the ground, then touched down five minutes later? How long would it take for Tyler to conclude that each pill made him float for five minutes?

I have to admit something. I only wrote this to talk about this bit! What would you do? Would you just keep on taking pills until you come to a conclusion? Obviously not!

First you would ask a friend to observe you and take video. You would ingest the pill in a controlled environment. You would tether yourself to the ground.

Then you would try to lift increasingly heavy weights. You would test whether the pill works differently on an empty or full stomach. You would give the pill to a test subject, without mentioning flight, to see if the effect is different. You could try to prime them about machine elves and try again. Off the top of my head, depending on other factors and the outcomes of the first experiment, there are so many other things to test, and so many follow-up questions.

Like, is the flight caused by a force exerted on the ground, or are you moving the air? Do you just become lighter? Where does your mass go? Does your body become a jet engine? How would you stabilise yourself? Does it make a noise when you fly? Is the energy stored in the pill or taken from your body?

The existence of this pill wouldn't just contradict what we know about flight, or what we know about pills. It would contradict basic biology, basic chemistry, or basic physics. It might violate the laws of thermodynamics. The most remarkable thing about this pill isn't that it makes you fly. Before that, it's more remarkable that you survive, that you land after five minutes, and that the pill doesn't violently dissipate its energy as soon as you open the jar.

So at the very least, I would need a stopwatch, and two video cameras, one for myself, and one for the observer, so I could make sure this isn't a completely subjective thing. There is only one problem: If the pill makes me think I fly, and if it makes it all look completely real, I think I'll be too far gone already. As my confidant sees me violently convulse on the ground for five minutes, I am experiencing an hour, with five minutes of flight, and fifty-five minutes of excited review of the data that feel completely real.

Your answer is, "Yes of course I'd use scientific tools to check your own observations, duh".

Leverage itself obviously deserves a not particularly subtle hostility, because as described by the people who were there it was Peter Thiel giving a guy a lot of money to start a cult.

That's the very clear takeaway from the non-hostile, primary source.

I suppose any large philosophy is subject to cultish thinking but there's something I find particularly odd about it coming from rationalism.

Every homeopath will tell you that people thought Gallileo was crazy, too.

I really don't want to defend Leverage Research in order to defend the sequences, or Slate Star Codex, or Everything Studies, or some really good LW posts by Stuart Armstrong, Luke Muehlhauser, or whoever. I care about ideas.

But he was everywhere. People talked about him. People talked about him missing the point of Squid Game. People talked about the Mr Beast production handbook. People talked about retention editing.

I had watched three Mr Beast videos, just to see what the fuss is about. They are about things that are really expensive, and they feel like trailers, but then the thing that's teased isn't even happening. There is only a trailer, no video. That's retention editing. Also, we blew up a Lamborghini.

The thing that really messed me up happened later. I began to see Mr Beast everywhere on YouTube. Not everybody wanted to be PewDiePie or tried to emulate him, but now everybody was slowly starting to emulate Mr. Beast. Everybody slowly took pieces from his production handbook.

(or if they have, then they're not experiencing the same thing most people mean when they say "sex")

bc it seems that to them, "sex" is a mildly pleasurable activity on par with a walk in the park or eating ice cream, which inspires no passion, elicits no strong feelings, which no one should experience intensely or care very much about, and which matters as much as and should be talked about in the same way as the content of your average HR policy email. all very cool and clean and antiseptic.