39300+ entries in 0.024s
mircea_popescu: asciilifeform,
they're not equally distributed. here's what i propose :
take a rng run, ent ;
then
take out all carmichael numbers from it, run ent.
then see if you can
tell which is which.
a111: Logged on 2019-03-17 03:03 mircea_popescu: "We wonder, and some Hunter may express wonder like ours, when
through wilderness where London stood, holding
the Wolf in chace he meets some fragment, huge, and
there stops
to guess. What powerful but unrecorded race once dwelt in
that annihilated place ?"
mircea_popescu: asciilifeform, i'd say you found
the first meaningful patch
to be put on
the fg
tree.
mircea_popescu: imo item should be baked anyways, as a
tmsr.crypto model object.
BingoBoingo: <mircea_popescu> being pretty much
the ~only
test anyone here in any substantial sense gives a shit about. << Indeed
mircea_popescu: being pretty much
the ~only
test anyone here in any substantial sense gives a shit about.
mircea_popescu: but yes,
the relation you unearth is sound.
the problem or set
thereof i started discussing is exactly homomorphic
to "well, we have no proper rng
tests, "ou'll have
to
take
the girl by
the nose, count, and break out
the abacuses.")
mircea_popescu: your rng working or not is aside
the point ; we're discussing here random numbers as a mathematical abstraction, we're not even counting "well, your set of 4, 4, 4, 4 is not exactly an implementation of
that abstraction"
mircea_popescu: asciilifeform, and
this SHOULD STAY
THAT WAY, is one
thing i'm saying. for
the above reasoning.
mircea_popescu: afgain : just because you've proven
that 3/4 numbers in
the interval 1, n have property p, you have not shown ANYTHING about how many numbers have
the property p in some other interval q, q+k.
mircea_popescu: argument reduces
to "you do not know what random means".
a111: Logged on 2019-03-28 15:05 mircea_popescu: i do not dispute
that
the exercise is not worth doing, unless one has a girl nose fetish. nevertheless, worth doing is a different consideration from
true.
mircea_popescu: asciilifeform, i dun see how works either! but what i'm saying is
that you don't have ~proof~ it doesn't!
mircea_popescu: just because "every other human is female" does not mean
there's 3 males in my harem.
mircea_popescu: statistical means some
things, we do not know how it diverges once we narrow intervals.
mircea_popescu: asciilifeform,
the problem's not necessarily stated as you state it. i am not capable
to know in advance WHAT attack will have
to be faced.
mircea_popescu: asciilifeform, for all you know, all numbers between n and n+k are liars for all primes larger
than 2^l
mircea_popescu: i do not dispute
that
the exercise is not worth doing, unless one has a girl nose fetish. nevertheless, worth doing is a different consideration from
true.
☟︎ mircea_popescu: there, again
there is not a mathematical proof
to permit you
to say -- if i say "the largest contiguous set of consecutive liars before number n will contain k such liars where k = n divided by alf's gf's nosehair count"
there's no formulaic approach you can fall back on. you'll have
to
take
the girl by
the nose, count, and break out
the abacuses.
mircea_popescu: the bound presumes a flat spectrum rng and properties of large sets of random numbers
that ~have not been proven~,
though
they are experimentally VERY reliable.
mircea_popescu: this is not what i'd call "crackpottery". for all you know
there IS a manner
to construct "~always lied about by random witnesses" prime candidates.
mircea_popescu: asciilifeform, as
to your earlier question : not such a large degree of crackpottery needed. consider
that if i affirm
today
that a) given a list of however many ~randomly chosen~ witnesses from (1,n) b)
the number k = x^2 + q x + p is going
to be falsely identified as a prime number while n = q ^ 2 + p ^ 3,
there is c) nothing you can practically do
to give me
the lie other
than ~actually construct such numbers and check~.
a111: Logged on 2017-07-18 22:58 asciilifeform: whaack is quite likely
thinking of
the bulk of
the b00k, which consists of blockcipher liquishit which is complicated for no reason at all other
than
the religion where 'it is confusing
to ME, author, and
therefore Must Be Hard
To Break'
mircea_popescu: it is factually
true
that
the liar occurence in (1, n) is ~1/3 ; i don't know how
to evaluate
the occurence in (n, n+k). it could be 1/3^2 or 1/3 ^ k or 1/3 ^ 1/k^16 or whatever
the fuck else.
mircea_popescu: in any case, it seems
to me
that
the a witnesses MUST be generated as rng(0, 2^4096) rather
than rng (2^4095, 2^4096).
mircea_popescu: asciilifeform, possibly a is actually significantly weaker
than b. i dunno, intuitively i'd readily agree with you, but
the fact is ~we don't know~.
a111: Logged on 2016-09-11 22:50 asciilifeform: it is foolish
to design for 'what if my rng silently fails'. it is a 'jesus bolt' failure
mircea_popescu: atm i am unaware how one would answer
this ; but if any brins are in
the audience by all means, even an unsuccessful ~meaningful~ attempt at
tackling
this question of our
times is liable
to make new math.
☟︎ mircea_popescu: type a has
the disadvantage
that possibly we end up with worse witnesses, and
type b has
the disadvantage
that possibly lists of n
to which our known list of candidate witnesses is blind
to may be also constructed in advance.
mircea_popescu: the ~only question
that actually needs an answer is whether a) a selection of n random numbers b bits long made every
time a candidate prime is checked for compositeness or b) a pre-given list of prime numbers b bits long among which n are randomly picked each
time a candidate prime is checked for compositeness is a better approach
towards checking a candidate prime for compositeness.
mircea_popescu: thinkin gmore about
this whole witness and liar discussion :
diana_coman: I had it in my stash from writing eucrypt
time so it rang a bell instantly
mircea_popescu: diana_coman, nice find, definitely
the source of my memory.
mircea_popescu: "Digg Inc., a social-media pioneer once valued at more
than $160 million, is selling for
the deeply discounted price of about $500,000,
three people familiar with
the matter said." << back in
2012.
mircea_popescu: all part of work
to make silent subs ; even
the exhaust gets bottled.
mircea_popescu: afaik it's internally organized much like
the "car battery" : a buncha half-liter cells inside a lined
tank
mircea_popescu: the engines won't go very far on it (in no small part because it';s almost never
taken
to ~water density, ie, 800 or so atmospheres) ; but
there's still a lot of air
that can be fit in some not-so-huge
tanks.
mircea_popescu: in
truth, never actually looked into
this
too closely.
mircea_popescu: ~everyone does
this now ; im pretty sure soviets also did.