imyxh

[imyxh @iimyxh]

@Smirkley if i sprinkle 1 drop of water on a gpu and resell it on ebay is it a zillion dollars of economic output per 5 gallons of water?

[imyxh @iimyxh]

@ug0tpwn3d @jamieantisocial chacha8 is not easily broken

[aryl ⌬⌬⌬ @endless_sine]

someone stop her PLEASE

[imyxh @iimyxh]

@quetzal_rainbow @ChazakielDoremi and anyway, supposing the problem _were_ so tractable, here's something strictly better than one-boxing: flip 100 fair coins. iff ≤49 of them are tails, take both boxes.

[imyxh @iimyxh]

@quetzal_rainbow @ChazakielDoremi then we're at a situation where what you do in the room influences p. imo it's inadmissible to then treat p as an independent parameter to decide your choice. the very knowledge that you might attribute a probability to an event changes the prob of the event? here be dragons

[imyxh @iimyxh]

@quetzal_rainbow @ChazakielDoremi did the predictor know you would flip the coin? p = 0.5 all along? but only for you? what about me, if i always two-box, and predictor knows: p ≈ 1 for me? so p is different for different people?

[imyxh @iimyxh]

@quetzal_rainbow @ChazakielDoremi i don't think "your choice was predicted correctly" is an event with a well-defined probability (p). consider: walk into the room. what is p? flip a fair coin to decide how many boxes to take. what is p now?

[imyxh @iimyxh]

@HeNordlinder @Aella_Girl but since we're trying to identify them with individual rows, we only care about P(is any random row), not P(in released dataset), so the dataset size cancels out, right?

[imyxh @iimyxh]

@HeNordlinder @Aella_Girl i guess my question is in what sense is 3.6% a prior and for what estimate?

[Will Sennett @Senn_Spud]

(Cognitive behavioral therapy enjoyer I just cut off in traffic) Think positively. He is probably in a rush for a reason. Maybe he’s late for a job interview. Maybe his wife is giving birth

Me: I’m da king of da highway

[Gazam @NotGazam]

@DeepDishEnjoyer You forgot the +AI after you integrate

[imyxh @iimyxh]

@HeNordlinder @Aella_Girl i'm confused, how would the prior change for a given row? the probability of being anywhere in the subsample is 3.6%, but for a specific row it's still just reciprocal of original sample size, right?

[imyxh @iimyxh]

@Aella_Girl @sparr0 no lower than the odds of that row being your friend before you took the subset. if i shuffle a deck of cards, then draw 5, is any given ace in my hand less likely to be the ace of hearts than if i had drawn the whole deck into my hand?

[imyxh @iimyxh]

@EigenGender @Aella_Girl i don't think this is true. look at bayes rule: P(row is sue | row has sue's answers) = P(any random row is sue) / P(any random row has sue's answers). neither of those changes when you discard data.

[imyxh @iimyxh]

@crilly_jim @DoingFedTime @JackRhysider sure but proton servers still get the plaintext in the end

[austin petersmith @awwstn]

today at the passport agency:

very serious passport lady: hi sweetie, can you tell me who these two adults that brought you here are?

6yo, deadpan: i have no idea

[imyxh @iimyxh]

@olivine_stan i'm always saying incredibly stupid things to my gf and then 2 weeks later it'll come up and i'll be like that's incredibly stupid there's no way i said that