I posted a top level comment about this also, but Anthropic has done some research on this. The section on reasoning models discusses math I believe. The short version is it has a bunch of math in its corpus so it can approximate math (kind of, seemingly, similar to how you’d do a back of the envelope calculation in your head to get the orders of magnitude right) but it can’t actually do calculations which is why they often get the specifics wrong.
One of the big AI companies (Anthropic with claude? Yep!) wrote a long paper that details some common LLM issues, and they get into why they do math wrong and lie about it in “reasoning” mode.
It’s actually pretty interesting, because you can’t say they “don’t know how to do math” exactly. The stochastic mechanisms that allow it to fool people with written prose also allow it to do approximate math. That’s why some digits are correct, or it gets the order of magnitude right but still does the math wrong. It’s actually layering together several levels of approximation.
The “reasoning” is just entirely made up. We barely understsnd how LLMs actually work, so none of them have been trained on research about that, which means LLMs don’t understand their own functioning (not that they “understand” anything strictly speaking).
Oh. Where you going?.. Oh, you men are all alike. Seven or eight quick ones and then you’re out with the boys to boast and brag. YOU BETTER KEEP YOUR MOUTH SHUT. Oh… I think I love him.
would the world be a better or worse place if everyone did what I’m about to do?
This is basically another formulation of Immanuel Kant’s categorical imperative, sometimes phrased:
“Act only according to that maxim whereby you can at the same time will that it should become a universal law.”
Which is itself basically another version of the Golden Rule: do unto others as you would have them do unto you.
I’ve heard advocates for the Platinum Rule as well: Do unto others as they would have you do unto them.
My point is that even the great thinkers of history, in conversation with each other over the millennia, have not gotten over the hump of morality being fundamentally subjective yet some lust for justice keeps us arguing in favor of some version of objective, universal morals.
One of the more helpful tools I’ve found is John Rawl’s veil of ignorance, also called the original position argument. Basically, if you were redesigning society around your new rules, but had no idea which position you would hold in the new society (it’s randomly assigned or impossible to predict, in the thought experiment), would you consider your rules successful and your society valid?
This tool allows for an objective evaluation of many subjective points of view, through statistical means.
All of these tools fail in a particular way though, which is that they individualize the search for ideal behavior. They ask: What morals would be best in a perfectly designed society, of which I will be the architect. Perhaps no individual is capable of devising a universal system of behavior?
Locked into their subjective experience of the world, how could any individual operating within such a system gain the conceptual distance necessary to redesign the whole? Rather, we are all shaped and attempt to shape society, aided and resisted by our resources and allies, in a chaotic and turbulent system that we are incapable of existing outside of. Even with a plan for a universal morality, how could you possibly implement it without contradiction?
Ah man, I’d been holding off on setting up another TES game on my PC since going to Linux. Now I’m going to have to go pre-emptive mod shopping for the essentials, and do it for so many games!
It’s a real shame that between crypto currencies, federated servers, and bittorrent we still haven’t come up with a fair and robust system for hosting large content archives.
The need for corporate-funded servers is killing video sites, mod sites, emulators, streaming, libraries, and so much more. Even the “self-hosted” crowd is like 90% dependent on cloud-hosting companies.
Commandos to me is the start of a different lineage of real-time tactical stealth games, which goes on to include Desperados, Shadow Tactics, and Shadow Gambit (yes, most of those were made by the same team).
Outside of the OGRE-alikes (FO Tactics, FF Tactics, Disgea, and so on) some other options for tactical games that are a little different:
This is a good idea, but also working remote frees up time to meet new affinity groups.
Not to dump on people’s relaxation strategies, but even the most introverted person can’t survive on video games and gooning alone.
If you don’t want or like hanging with coworkers, find a local bar to hang out at and meet some folks, go to a community board game night, join a choir, attend an anime viewing night, just do something to take initiative and meet some folks that like what you like.
You never know how many of those comments now aren’t bits and ai also. The malleable human mind sees “people” expressing opinions and wants to take a side, have an opinion. How convenient that all the options, all the feelings and responses you should have, are already laid out for you. Just “Like and Subscribe” to the persons whose opinions most align to your own.
New Orleans also had No Kings in the morning and (corporate sponsored, thanks Shell Oil!) Pride in the afternoon yesterday. Lots and lots of folks made this same silly sign or a variant. “No Kings, Yassss Queens” etc.
I tend to concur with the sentiment that it still promotes monarchism as aspirational, but I think a lot of protest signs are “heart in the right place” kinds of things. Like anyone that snow clones “Make Something Whatever Again” I hate, quit validating his slogans with copies!
Or calling Trump TACO and using tacos as mocking images. I don’t want to associate Trump and tacos, I liketacos! And he doesn’t “chicken out”, if you think his tariff nonsense is about playing chicken, please learn what “market manipulation” is.
“Elected” haha, I know who’s getting my “vote” this year!
I’ll reiterate a comment I made earlier today:
Peaceful protest is only a meaningful choice when violence is also a real option.
Otherwise you’re just holding a pointless rally. Might as well show up with handcuffs on already.
Chuck’s already got ya!
https://www.thriftbooks.com/series/trans-wizard-harriet-porber/200695/
Eh, there’s a curiosity aspect as well. I can’t do work on my car, but I can change the oil, tires, brake pads, and such. I understand the principle of how an IC engine works. I’m a computer programmer but I think it’s because I’m a curious person who likes knowing how things work, and computers offer more chances to learn than anything else on the planet.
It isn’t ignorance that has ever bothered me about boomers, zoomers, or anyone else. It’s that 99% of people you meet are fundamentally incurious. They don’t care how things work, they don’t care if they could work differently.
Yatta! An ancient amusement from the early days of the internet.
“Pencil Break” was the hot shit for about a year in middle school. This was right on the edge of when mechanical pencils started to replace wooden ones in school (not completely, I’m sure kids still use wooden #2s.)
It was stupid, but so many people would take turns slapping pencils together to see which ones would break. Obviously brand new pencils out of the box snap the easiest, while worn down nubs are impossible to break. The school eventually banned it and briefly flirted with banning wooden pencils after parents and teachers complained that all the really dedicated students had no writing instruments left.
Nation of geniuses over here.
Trying to imagine objects in higher than 3 spatial dimensions.
Imagining 2 or more temporal dimensions.
Designing a system of governance that is fair to all constituents, physically realizable, and marketable enough to convince future constituents to follow it.
I have a vague notion of a new color. Success!
Hmm…this repost is really making me regret softening my stance in the original thread yesterday.
Where’s the twin peaks meme community? c\blacklodge? c\rrdiner? c\damnfinecoffee?