Aransentin
p ≥ 0.05 zombie
No bio...
User ID: 123
Aransentin
p ≥ 0.05 zombie
8mo ago
In Demolition Man (1993), there's a scene where they mention Taco Bell. In the international version, this is dubbed over to Pizza Hut, presumably because it's more recognizable outside of the US. You can still kinda tell that the audio had been changed though, as their mouths don't match.
I'm pretty sure I wouldn't have spent much time reading the Taco Bell wikipedia page if that hadn't been done.
Aransentin
p ≥ 0.05 zombie
8mo ago
No, but it was enough for the down payment of the loan.
Aransentin
p ≥ 0.05 zombie
9mo ago
Well, I'm of the opinion that the 'intended meaning' of a poem or piece of media is totally worthless and uninteresting
If that is so, you could very well just entertain yourself with a random word generator like Terry A. Davis, no?
The very best art taps into aesthetics or concepts that are to some extent transcendent
I don't necessarily disagree! Those concepts should in principle be discernible however, and above the "noise floor" of the random associations I'd label "schizo". I don't deny that there might be interesting unconscious features of works, just critiquing the tendency of critics to find signals in random noise.
Aransentin
p ≥ 0.05 zombie
9mo ago
In Sweden to pay our bills you can have a thing called "autogiro". This means you grant a company the right to pull a certain sum from your account each month, so you never have to bother with any administration for your bills. Pretty convenient, except I've been using it for all my monthly expenses so I've never had to log in to my bank at all.
Early this year was the first time I've logged in to my bank for years. I've had constant anxiety to see how much money I actually had, and finally I had to bite the bullet and actually look.
Turns out, I had MUCH more money than I expected. The relief I felt from that was pretty much indescribable; I've now proceeded to buy a condo much closer to the city centre, and have been going out to eat more or less every weekend.
Aransentin
p ≥ 0.05 zombie
9mo ago
Is anyone else interested in poetry here?
A thing I've noticed in poetry analysis that annoys me is what I've come to think of as "schizo" interpretations.
On the one hand, you have symbolism that was likely put into the text intentionally; e.g. in "Ozymandias" (which I assume you are familiar with) the famous "Look on my works, ye Mighty, and despair!" can be very reasonably interpreted as the onlooker ironically despairing that even the greatest ruler will eventually decline and be forgotten.
On the other hand you have stuff like "Scholars such as professors Nora Crook and Newman White have viewed the work as critical of Shelley's contemporaries George IV, with the statue's legs a coded reference to the then Prince Regent's gout". How reasonable is this interpretation? I think not very; if the poem had instead referred to the statue's arms or something, would there have been another possible tenuous interpretation to some other person or concept? Probably. You would need some sort of Bayesian intuition for this, as there is a "base rate" of possible random associations you could make – and for any connection less credible than that you're basically finding patterns in random noise.
It reminds me of how the famous schizophrenic programmer Terry Davis would "speak to God"; he had a random number generator that spat out words, and he'd do free associations between them. Textual criticism is rife with this. I suspect it's because there is really no incentive to find out the "truth" of a text, just finding cool associations that makes the reader look smart, and since there's no ground truth to verify anything it easily gets disassociated from reality.
To "fix" this, I propose a calibration game of sorts. One would write a text with actual symbolism and poetic devices, then publishing both the text and a canonical explanation for everything in it. Readers could then interpret it, and afterwards find out how much their interpretation missed the mark. If anyone wants to try this, I have done so with one of my old poems here.
(For the unfamiliar with meter, it's written in straightforward iambic tetrameter, i.e. each line consists of four pairs of unstressed-stressed syllables; with an ABAB rhyme scheme.)
Aransentin
p ≥ 0.05 zombie
10mo ago
The SuperNova Early Warning System (SNEWS) is a network of neutrino detectors designed to give early warning of a near-earth supernova. They have a mailing list where you can sign up to be one of the first people to be notified when it happens. I signed up for it; just imagine the party trick you could pull off by getting that email one evening and pointing to Betelgeuse before it blows up!
Aransentin
p ≥ 0.05 zombie
10mo ago
Thanks! If you're wondering why on earth we named it SWAGGINZZZ, it's due to us messing up the execution a bunch of times, and since we created a new account/player every time we pretty quickly exhausted all the good names we could think of. The final run was named that because we were just testing things and didn't expect that one to actually succeed..!
Aransentin
p ≥ 0.05 zombie
10mo ago
Do roguelikes count? I've played a whole lot of them, like Dwarf Fortress, Cataclysm DDA, ADOM, and of course NetHack (for which I have to shamelessly brag that I hold the shared realtime speedrun world record in, from back when tool-assisted runs were allowed)
Aransentin
p ≥ 0.05 zombie
11mo ago
Yeah, but presumably scammers don't care about that wherever they are.
Aransentin
p ≥ 0.05 zombie
11mo ago
Before 2022 it was really easy, you could just buy a prepaid card with cash in a store and it'd have nothing tying it to yourself. Nowadays you need an ID. I checked a random telephone provider (Telenor) and they require you to use the national e-ID to buy a SIM card (or apparently upload a photo of your passport for international customers).
But yes, it's probably the limited market size and language barrier; especially since the prime target for scammers are old people who don't speak English very well.
Aransentin
p ≥ 0.05 zombie
11mo ago
I see Americans complaining online about how many spam telephone calls they are getting, to the point they don't even pick up the phone when the caller is unknown.
This is pretty alien to me; I live in Sweden and have literally never gotten a spam call in my life. (Maybe the reason for this is that the scammers naturally won't bother learning Swedish?)
So, potential silly lifehack: why not get a foreign telephone number from e.g. Denmark or something, and then never get spammed again? Presumably your calls will be more expensive as you'd be paying the international rate constantly, but this can be ameliorated by getting a plan where that's cheaper.
Edit: Oops, you said "apolitical" which this hardly is; at the very least not back when it happened!
I'd say that there's a somewhat reasonable chance that the Red Army Faction didn't commit suicide but where executed by the West German police, as the circumstances are pretty suspicious:
They managed to smuggle guns into their cells in one of the most secure prisons in the world at the time, in a wing built especially for them. The guards a few meters away didn't hear any gunshots, Baader fired thrice, missed the first two, and finally died from shooting himself in the neck from behind using his right hand, but he was left-handed.
There was no fingerprints found on Raspe's gun and no gunpowder residue, even though his gun was lying in his hand when he was found. Meinhof's hanging was suspicious as well (wiki article, too many things to list here). Möller stabbed herself four times in the chest, survived, and claims to this day that somebody tried to assassinate her.
Aransentin
p ≥ 0.05 zombie
11mo ago
I am partial to the theory that advertising works by establishing a shared understanding of what consuming the product signals to others. It explains many puzzling behaviours, e.g. why Coca Cola would bother with ads – they want to steer how other people will percieve you when consuming their product, not necessarily create awareness or an immediate desire.
In that framework, the Mulvaney backlash makes perfect sense; the blog post I linked even talks about how brands generally don't have different messaging on different platforms since it would be directly counterproductive to the idea that you want to establish a common product image. InBev didn't understand it, tried different messaging for different groups, and is now suffering the consequences.
Aransentin
p ≥ 0.05 zombie
11mo ago
In the corner of the ChatGPT replies there's a little profile picture; if it's a blueish-green it's running a 3.X model; black means it's 4.
You might need to click on the image on the Caplan Twitter post to view the entire thing, it gets cut off in the preview window for me.
Aransentin
p ≥ 0.05 zombie
11mo ago
You really need a bit of familiarity and prompt wrangling skill to get the most out of GPT. I've seen plenty of people testing it and fucking up basic stuff which makes them conclude that it's overhyped.
A common issue is not pressing the "GPT-4" button in the ChatGPT interface, and getting GPT-3.5 when you think you're running 4; and not being familiar enough with the interface to recognize that the blue-green icon means it's a lot dumber. E.g. Bryan Caplan fell for that.
A second issue is that GPT is at it's core a predictor, not a simulator. If you ask it something and it messes up the response, asking it something else in the same chat context means it will now start predicting that it's a character that makes mistakes. Resetting the chat fixes that. In the same vein, if you are giving it a quiz, having it predict a student response is suboptimal. Students make mistakes, you want it to predict the answer key.
Also, giving it few-shot examples improves it a great deal. If you have an example of a perfect query+response to a similar question, stuff it in the prompt itself! Then the LLM will know what kind of response you are looking for.
Aransentin
p ≥ 0.05 zombie
11mo ago
Huh! As an European I have always assumed it was the same all over the western world – a down-filled duvet in a removable bag akin to a big pillow case, and nothing else. Americans have multiple layers? Weird. Does the top layer not slide off or bunch together during the night?
Yeah, there's a lot of text on the internet. With a pretty cursory Bayesian analysis, even with a 99.9% accuracy you're looking at a thousand false positives if you are combing through a million posts. Without some other thing to narrow it down, it seems reasonable that it'll not be possible from writing patterns alone.
More options
Context Copy link