103500+ entries in 0.035s
mircea_popescu: more like comparing "ast nodes". not the best and not the worst measure of complexity.
mircea_popescu: nevertheless - the argument as to neuron count is not directly mass ; it is parts. one DOES compare transistor counts.
mircea_popescu: (ftr, the "100bn neurons" figure IS MISTAKEN. yes at any point you slice a brain you find something like that ; but neurons also die - yet their death is not automatically in vain now is it!)
mircea_popescu: which is why we could properly qualify this complex as a memetic virus. it infects, and the host can feel it's infected, but can't prevent the infection.
mircea_popescu: it certainly is MADE to be persuasive for the sort of minds coming out of ivy league "colleges"
mircea_popescu: anyway, i suspect on the face, "the difference between stupid 5k cell brains and smart 5k cell brains is in the yottabytes they looked at" is almost persuasive.
mircea_popescu: it's not even some obscure poetic license. basic vocabulary.
mircea_popescu: it's what "totalitarian" means, after all - everyone is the enemy.
mircea_popescu: it was also said cca 1800s, it read like "si de-aceea tot ce misca-n tara asta, riul, ramul imi e prieten numai mie, iara tie - dusman ti-este."
mircea_popescu: asciilifeform i recall this being said here years ago aha.
mircea_popescu: anyway, the fate of the socialist empire in its eternal historical cyclicity is fascinating. so - nazis put everything on large tanks ; and failed. but then soviets put everything on a particular notion of industrialization (steel-per-capita industrialization, so to speak) and... failed. so the one socialist empire left standing is putting it all on "ai". and...
mircea_popescu: but it's why the two week war isn't done after thirteen.
mircea_popescu: the funny thing is that the approach is trivially defeated in practice. they're relying on a ~fixed~, unchanging enemy. which naivite in the manner of 3yos and their modelling of environment geometries was previously touched upon.
mircea_popescu: historical accident be what it may, it can't really be said it constitutes a proper cause.
mircea_popescu: the criterion is not "optimal efficiency", but "optimal efficiency in the paradigm space".
mircea_popescu: "why would girl go to all thetrouble of pretending on facebook when she could just walk up to every guy in sight and go "hey, wanna feel my tits ?"
mircea_popescu: some comments mp made here last year sent some cold shivers.
mircea_popescu: and they expect to do it by pouring lots of gb in very trivial neural/expert systems.
mircea_popescu: it's in the more general form "detection of edges". they want to identify cars, and go winning strategies, and people who don't like them and everything else.
mircea_popescu: the applications to eg banking are beyond lulzy. but they're willing to destroy the currency (currency ? what currency ? dollars don't pay nothing, it's notes from hitler if you want anything done) to try and prop it up because outside of usg.ai there's nothing
mircea_popescu: they will suffer the same fate, too, but until he is killed the death convict still lives like any other.
mircea_popescu: basically, the ~one hope of the imperial technologistic church of some sort of practical results is a peculiar form of ai. they're doubling down on it exactly in the manner usg doubled down on "alternative fuels".
mircea_popescu: the other expansion of that is of course the whole "mechanical fakenews detection"
mircea_popescu: asciilifeform incidentally that nonsense is only superficially related to snowden. in practice it is an extension of the "artificial intelligence" used in "spam detection", especially as the doomed but desperate attempt at websites to fight the republican bots.
mircea_popescu: most resilient use of self-blown glass was because "it's the only way to get a proper seal"
mircea_popescu is curious if the forum will prefer gabriel_laddel parked in sv start-up over gabriel_laddel parked in park.
mircea_popescu: devwork must be pretty hectic with all this accidental deleting.
mircea_popescu: (the above boolean function K is an evident extension of the dericlet function ftr)
mircea_popescu: they more or less udnerstand they're a formalisation of what the more numerate philosophers call the theory of knowledge.
mircea_popescu: anyway, i'm entirely not up to speed to this, but there actually exists this field in math, of "boolean function sets"
mircea_popescu: the padding pads the message just as the message pads the padding.
mircea_popescu: more than that : there can't be conceptual difference between number x and number y.
mircea_popescu: importantly - there is no need for the payload and the padding to be distinguishable.
mircea_popescu: can be constructed so their values are known besides K0(x) = true for K0(a) = 1 if a=x and 0 if a!=x.
mircea_popescu: but yes, the statement is correct. to restate : 1. padding is defined as a function F (x, y) with x, y and F ∈ N such that 2. F is defined for all (x, y) pairs and that 3. given any set of boolean functions Ki(x) along with their values no Kj(y) can be constructed so their values are known besides K0(y) = true for K0(a) = 1 if a=y and 0 if a!=y and 4. given any set of boolean functions Ki(y) along with their values no Kj(x)
mircea_popescu: specifically 256kb does not even remotely promise for 64kb to be walked. you get a third untouched more than half the time.
mircea_popescu: moreover, to be 99% sure 100% of 64kb were walked, you need millions of steps.
mircea_popescu: this resolves the part where "all have been walked" though not the part where there's structure intrinsic in the method.
mircea_popescu: yes. a point n distance from origin has more chances of being walked over than a point 2n distance from origin by a factor of about 1.4
mircea_popescu: "we have always been at war with fake news such as any item that is politically inconvenient today."
mircea_popescu: ben_vulpes lol that leclerc dude is visiting your blog on a weekly basis ? 2, 9, then nobody gave a shit so he skipped 16 and i guess 23, we see him on the 30th ?
mircea_popescu: good thing there's reddit. now that yahoo "finance" closed the public boards, where'd all the scum go.
mircea_popescu: asciilifeform thinking about it, the walker scheme doesn't seem too great. the theoretical objection is that it ~does~ introduce structure, through the convention that the walker moves from where it last moved. in practice though, run a few simulations over a 16bit message which you can then print as a 256x256 bitmap. your walker makes anthills basically.
☟︎ mircea_popescu: another problem is the observation that 1011011011 is neutral.
mircea_popescu: you prolly want circular tape the size of message tbh.
mircea_popescu: if you make the waltzer start from ~the end~ of the message, even sqrt(n) steps improves rsa enough.
mircea_popescu: a 5x size penalty for full strength encryption is not inconceivable either.
mircea_popescu: i now have to a) generate 4kb of entropy (roughly enough for 8 4096bit rsa keys) ; b) complete 16k operations to pad ; c) execute a 5kb rsa exponentiation. so i'm looking at what, about an hour ?