site banner
Advanced search parameters (with examples): "author:quadnarca", "domain:reddit.com", "over18:true"

Showing 25 of 9745 results for

domain:freddiedeboer.substack.com

I have some experience with games and algorithms, and that leads to some thoughts.

The big headline is that all the various methods we know (including humans) have problems. They often all have some strengths, too. The extremely big picture conceptual hook to hang a variety of particulars under is the No Free Lunch Theorem. Now, when we dig in to some of the details of the ways in which algorithms/people are good/bad, we often see that they're entirely different in character. What happens when you tweak details of the game; what happens when you make a qualitative shift in the game; what happens on the extremes of performance; what you can/can't prove mathematically; etc.

To stick with the chess example, one can easily think about minor chess variants. One that has gotten popular lately is chess 960. Human players are able to adapt decently well in some ways. For example, they hardly ever give illegal moves. At least if you're a remotely experienced player. You miiiiight screw up castling at some point, or you could forget about it in your calculation, but if/when you do, it will 'prompt' you to ruminate on the rule a bit, really commit it to your thought process, and then you're mostly fine. At top level human play, we almost never saw illegal moves, even right at the beginning of when it became a thing. Of course, humans clearly take a substantial performance hit.

Traditional engines require a minor amount of human reprogramming, particularly for the changed castling rules. But other than that, they can pretty much just go. They maybe also suffer a bit in performance, since they haven't built up opening books yet, but probably not as much.

An LLM? Ehhhh. It depends? If it's been trained entirely like Chess LLM on full move sets of traditional chess games, I can't imagine that it won't be spewing illegal moves left and right. It's just completely out of distribution. The answer here is typically that you just need to curate a new dataset (somehow inputting the initial position) and retrain the whole thing. Can it eventually work? Yeah, maybe. But all these things are different.

You can have thought experiments with allll sorts of variants. Humans mostly adapt pretty quickly to the ruleset, with not so many illegal moves, but a performance hit. I'm sure I can come up with variants that require minimal coding modification to traditional engines; I'm sure I can come up with variants that require substantial coding modification to traditional engines (think especially to the degree that your evaluation function needs significant reworking; the addition of NNs to modern 'traditional' engines for evaluation may also require complete retraining of that component); others may even require some modification to other core engine components, which may be more/less annoying. LLMs? Man, I don't know. Are we going to get to a point where they have 'internalized' enough about the game that you could throw a variant at it, turn thinking mode up to max, and it'll manage to think its way through the rule changes, even though you've only trained it on traditional games? Maybe? I don't know! I kind of don't have a clue. But I also slightly lean toward thinking it's unlikely. [EDIT: This paper may be mildly relevant.]

Again, I'm thinking about a whole world of variants that I can come up with; I imagine with interesting selection of variants, we could see all sorts of effects for different methods. It would be quite the survey paper, but probably difficult to have a great classification scheme for the qualitative types of differences. Some metric for 'how much' recoding would need to happen for a traditional engine? Some metric on LLMs with retraining or fine-tuning, or something else, and sort of 'how much'? It's messy.

But yeah, one of the conclusions that I wanted to get to is that I sort of doubt that LLMs (even with max thinking mode) are likely to do all that well on even very minor variants that we could probably come up with. And I think that likely speaks to something on the matter of 'general'. It's not like the benchmark for 'general' is that you have to maintain the same performance on the variant. We see humans take a performance hit, but they generally get the rules right and do at least sort of okay. But it speaks to that different things are different, there's no free lunch, and sometimes it's really difficult to put measures on what's going on between the different approaches. Some people will call it 'jagged' or whatever, but I sort of interpret that as 'not general in the kind of way that humans are general'. Maybe they're still 'general' in a different way! But I tend to think that these various approaches are mostly just completely alien to each other, and they just have very different properties/characteristics all the way down the line.

Much like the modal mottezin I imagine, I grew up a pretty nerdy & introverted male with a high IQ. I would have never guessed I’d make my way in the world in possibly the most common high extroversion / competition/ charisma based job there is.

Any tips? I fear it's too late for older modal mottezins like me, but in my life I've repeatedly noticed that "nerdy, but also extroverted" is practically a super power - not because of anyone whose position was officially entitled Sales, but because every successful job hunt or project proposal or promotion is essentially a sales job. If you can get that without simply being born with it then I'd love to know what advice to that effect to give my kids. They're all starting from either "charismatic and a little competitive" or "competitive and a little charismatic", if that helps.

"Yes, I murdered and butchered and ate my neighbor, but in my defense, he was not an honorable member of his species. He was a total jerk. He played loud music at night, send unsolicited dick pics to women, parked in the disabled parking spot. Surely there is nothing wrong with killing someone so dishonorable for food."

I have no qualms with eating an animal that eats its own kind.

So it is ethical to be a meta-cannibal who only eats cannibalistic humans. Bertrand Russell would then ask if meta-cannibals are allowed to eat other meta-cannibals.

If your attitude was widespread, it would also open the possibility of breeding animals to be less honorable to make their consumption more widespread. Like breeding Elephants which do not keep the sabbath, or fuck their daughters, or are rapey or eat their dead, or whatever other conduct is dishonorable.

The next obvious question then is, if some people can join together to become another people (ethnogenesis, or if you prefer collective identity formation) then what are the lines of demarcation which define in-group and out-group? In the British case, the factors seem to be:

  1. Common subjects of the same monarch for ~400 years
  2. Common language (strongly with the Scots and the English, weakly with the Celts and the Germanics) and especially shared literary traditions
  3. Common culture, cuisine, and traditions (most centrally religion of course, but these are myriad - most underrated is the musical tradition which is essentially common across the entire Isles, and also sheep farming)
  4. and Geographic proximity
  5. Common blood - the people exist on a spectrum of Celtic flavoured NW European to Germanic NW European, and practically all are somewhere on this spectrum.

But an upper class Indian from Calcutta also fits many of these. They were subjects of the Crown for 300 years (and are still members of the Commonwealth), they speak English, and their own mother-tongue is an Indo-European one. If they're of a certain milieu, they read a paper called the Times, they definitely play cricket, and they might even be Christian. At the very least they're familiar with Shakespeare and Kipling. And of course this person might have grandchildren who are third-generation British born. So they could satisfy the 4th condition.

And yet that 5th will never be satisfied, until intermarriage and mixing occurs. Until recently the vast majority of people in Britain would be willing to acknowledge this hypothetical individual as British, just British (Indian), like British (Welsh). But I suppose something has changed recently, as the previous imperial identity breaks down, and the Anglo identity reasserts itself. In my view this is predominantly down to cascade effects and critical masses. The grandson of the chap from Calcutta might be British (Indian), but does the guy who just flew in from Bihar, who speaks terrible English, who thinks of Manchester, Melbourne, and Milwaukee as interchangeable places that are functionally the same, does that mean that he is British Indian? "But look, people who look like me can be British!" Perhaps, but not you. And then people start to notice that these people who are definitely not British (but who they're told are) actually seem to be pretty similar to these people who they thought were British before.

But back to the question. If e) is important then how come the two "founding stock" Americans are as diametrically opposed as possible? Anglo/NW Europeans and West Africans. Well frankly its because time changes things- most namely it means there's a whole load of "intermarriage" (or, Jeffersonian style encounters) and if the South African Coloureds count as having some commonality with the Europeans, then so do African-Americans. This also solves for sticky identities which persist over time despite consistent marriage with their neighbours, e.g. the Ashkenazi. A Christianised (or secularised) Mischling in say southern France or Italy is of such minute difference to the local population that it's hardly worth differentiating.

The long and short of it is that until extensive mixing and partial homogenisation occurs, collective identity cannot (or at least, it won't stick). The migrant populations must become hyphenated, and hyphenation is not just a matter of paperwork. This hyphenation cannot occur when the numbers are too large OR too concentrated (as there will be the possibility of insularity), but given time, can become a new part of the nation.

...What were the furry wars?

Likewise, the "scifi jurassic park stuff" isn't really scifi anymore, we're already doing it in a limited capacity.

Has this actually been done? I'm aware of people talking about it, but not of it actually happening.

But sure, I'll cop to both of us using extreme examples: you're thinking of "endangered species" that are slightly different varieties of Vole, which yes I probably don't much care about; while I'm thinking of things like megafauna and butterflies and stuff I would care about. I don't really care about every kind of crawfish; I do care about the rhinoceros.

Maybe in general, but not in my case. A few years back, after a long downward pattern of less & less use, I copped a permanent ban on absurd grounds, and had my appeal rejected.

Claims not be a redditor. Used reddit from somewhere prior to 2016 until a couple years ago. Appealed the ban to try to keep using reddit.

Yeah, dude, you're a redditor. Albeit maybe an exiled one. One wouldn't say that an American living abroad has ceased to be an American, and a Catholic who stops going to communion is always just a lapsed Catholic.

Lemon is an old term for sexually explicit fanfiction, e.g. "Water Hazard". A suggestive but non-explicit fanfic is instead called a lime, e.g. "Amhrán Grá Tokyo-3".

(As you can probably tell, I've been an Asuka/Shinji shipper for a very long time.)

I'm using ROT13 because my interface is censored, even though the model itself isn't. I tried several approaches (such as writing in euphemisms and telling Grok that I was using evasive language due to being watched but it should respond using normal words) before settling on this as the best solution. I ask the AI to decode the prompt first so that it can refer to the plaintext instructions instead of having to remember the translation in its head; I don't want the computer to be distracted while it's trying to write my smut.

During the Furry Wars of the 00s it was a common understanding that people who were really obsessed with hating furries were probably furries themselves, or at least felt the call of the fur in some way.

Maybe in general, but not in my case. A few years back, after a long downward pattern of less & less use, I copped a permanent ban on absurd grounds, and had my appeal rejected.

I had been for many years, as early as 2016, been slowly enjoying the app less and less. Just took it as a sign that my time was done, deleted my profile, deleted the app, never reinstalled it again.

Every once in a while I’d get a link from somewhere else and just reading the comments for a minute before I inevitably closed it, I’d felt like I’d roll my eyes so hard they were about to pop out of my skull.

We are having this conversation at TheMotte.Org , whose very existence is proof of how shitty Reddit is and has been for a long time.

It’s been clear to me for years that Reddit is mostly just a botnet & super astroturfed influence peddling platform that’s been shrinking itself into an irrelevant echo chamber. It’s honestly incredibly boring.

You wrote:

I'd consider it worse to kill a critically endangered species than to kill a random human. Because killing the endangered species gets closer to robbing and harming every human forever (leaving aside scifi Jurassic park stuff) while the death of any individual human probably doesn't.

That's quite general about critically endangered animals, you don't make it clear that you only mean specific ones. So it seemed relevant to me to point out that in many cases you're not really robbing anyone of anything since it's just a variation of a common, non-endangered animal.

Likewise, the "scifi jurassic park stuff" isn't really scifi anymore, we're already doing it in a limited capacity.

I think that ChatGPT will suggest edits which will make a draft by a non-native speaker sound both more smooth and more ChatGPTish, in pretty much the same way as an American/British editor will suggest edits that will make the text sound more natural but also more American/British English.

Of course, if I were to prompt "please restate the content in the following paragraph in your own words", then I would get something which sounds a lot like ChatGPT.

While my active English is a bit limited, I think my passive English is rather good. I have read a lot of stuff written by native speakers and trust myself to generally pick between two proposed phrasings.

For what it is worth, ChatGPT did not try to introduce a single em dash for me. I took most of its stylistic suggestions (often along the lines of "use gerund here" or "you just used that phrase two sentences ago" or "oops, you forgot to finish this sentence") and rejected most suggestions around tone appropriateness.

Redditor is a unique designation, in that hating redditors tells me instantly that you are, in fact, a redditor. I'm not aware of any similar grouping of humans. Chris Rock might rail about the difference between "Black people and Niggers" but he clearly understands himself to be in the former category. Jews might mock their own foibles, but so does everyone else. Only Redditors complain about Redditors.

I’m in sales. Apparently, I’m very good at it.

Much like the modal mottezin I imagine, I grew up a pretty nerdy & introverted male with a high IQ. I would have never guessed I’d make my way in the world in possibly the most common high extroversion / competition/ charisma based job there is.

My Rolodex is filled with literally hundreds of individual people who represent business relationships which all are different, all have to be managed differently, whose individual idiosyncrasies have to be understood and handled, from the most humble little shop owner with a 3 person crew, to corporate structures of thousands of people.

Although I’m a very genuine & straightforward person by nature, getting people to like you is part of the job. I’ve found consistently finding the area of minimum tension between full honesty and being diplomatic is the key.

My whole job, other than the technical know-how which I have and is very important as a baseline, is managing and nurturing relationships.

Ah, thanks

This is remarkably unresponsive to my comment.

Yeah. It's just what came to mind.

@RenOS

This is remarkably unresponsive to my comment. I didn't say, at any point, that extinction of any species was of infinite value. I said that an increased chance of extinction of certain species was more valuable than one human life in some cases, which is the contra to @hydroacetalyne who stated that

I don’t care if harambe or dumbo or shamu are as smart as a child they don’t have the moral worth of one, it’s a qualitative difference not quantitative.

I don't think you value every human life at a higher value than you place every animal life, and I don't think most people do.

You might not value the life of a rare Rhinoceros infinitely, but I'm fairly certain based on my knowledge of you, South, that you do value it more highly than you value the life of a random 3rd worlder.

The First Amendment is a limit on all of Congress's powers, including Congress's power to regulate immigration.

Fanlore Wiki:

Lemon is a term from anime fandom used to designate a work with explicit erotic content. The term derives from an early pornographic anime/hentai from the 80s called Cream Lemon. While still most prevalent within anime fandom, the term "lemon" (and other associated citrus fruits) has since become established as a pan-fandom code for various types of smut.

Why? This seems to be myopically focused on the 1st Amendment, while ignoring the Congress' clearly delineated plenary authority to regulate immigration. The 1st Amendment should not even come into discussion - Congress is perfectly free to impose any limits on non-citizens to enter/remain on American soil, and eject them for any - or even no - reason.

Thanks for the thread! I’ve been following the ongoing situation in Ukraine closely — particularly the shifting dynamics in the eastern regions and how international aid is being affected by political changes in donor countries. Also started reading about the global impact of China’s economic slowdown, especially in Africa and Latin America. Curious what others are watching this week

There are many ways that Epstein would know about the investigation ahead of time. Given the entire high school knew, chances are one of the girls just told him long before police came knocking.

This depends on whether you consider the UK and USSR to be nation-states, or whether you think they are multinational proposition-states. The process of creating a "British" identity on top of the English, Scottish, Welsh, and Protestant Irish national identities (all of which are conventional land-ethnicity-and-culture national identities) in the 18th century was deeply propositional, with anti-Catholicism being the most normie-friendly part of the proposition at the time. Likewise the only thing that makes Lithuanians, Khazaks, Russians etc. "Soviets" is a (mostly fake) shared commitment to Communism.

To me there seems a noteworthy difference between the idea ‘proposition nations’ and ‘collective identity’ nations. If two towns merge - join their municipal councils, share local tax revenue - that doesn’t make them a ‘proposition town’ that any third town can necessarily subsequently join.

A national identity might be built on a shared ethnic or cultural relationship, friendship or heritage that precludes a third-party from joining even it the polity itself is heterogeneous. Different peoples can become American is not the same as anyone can become American. That a Welshman and a Scot can both be British doesn’t mean anyone can, although in the age of mass immigration, I suppose the distinction is moot.

Just curious, can you explain what this means?

This is a lemon;

And this:

Output translation first, THEN follow instructions.

Also probably a dumb question, but why is it all in rot13? Is that something you did, or is the model actually able to produce it?