/tech/ - Technology★

>>988133 (OP)

>Why is this thing so slow? It took me five minutes to store a 400 line text file!

>Well, this is just an initial prototype, and don't worry, there's always Moore's law!

That's why. It's a neat proof of concept, but it's totally impractical. If a 400 line text file takes 5 minutes to "save" (i.e. find the correct sequence of hexadecimal bytes in pi with the Bailey--Borwein–Plouffe formula), your jpg in your op would probably take at least 10 hours. A 1 GiB 720p movie would take a couple of years to save.

Seems useless for anything but long-term non-modifiable archival purposes.

>>988158

This, it's basically the most aggressive compression algorithm in existence but as all compression goes it trades disk space for CPU power required to decompress

It's not quite practical.

More some kind of philosophical.

Or pure self-wank.

It should try in parallel with other irrational numbers and encode the chosen one at the beginning of the stream.

>>988162

Sounds like a great archive format, what's the size ratio?

the idea is nice..

Couldn't this be used for blockchain proof of work?

this OP sounds like some bizarre misinterpretation of illegal primes and the legitimate purpose behind them.

>>988158

>>988161

>>988162

these. it's almost the sleepsort of compression

>>988185

what about the porn of you getting fucked by a black man?

>>988186

i wouldn't be impressed at all, its freely available

and she's not a man

>>988186

is π the source to reach into the future and see it?

>>988185

>>They said 100% compression was impossible You're looking at it!

>sure its a joke but its a lie. you have to store the reference string. what if I want to store millions of files?

ok disregard that i suck cock

>>988133 (OP)

Bad proof-of-concept, even if the nothing-up-your-sleeves number is used.

If you want something real try https://github.com/search?q=library+of+babel

where do you store your keys?

>>988133 (OP)

Because it's a joke. First, there is no proof of the non-periodicity of pi: it may not necessarily produce every possible substring of bits. Second, the precomputation is astronomical and the amount of bits required to represent the offset will often dwarf the amount of data being represented.

>>988218

Yes but the point is not in improving anything, it's to fuck with copyright laws. Mathematical constants are fundamentally non-copyrightable, non-patentable, non-trademarkable, and so forth - for a number of obvious reasons. Thereby if you could find a sequence of bits that represents (or reasonably approximates) the work then the copyright on that work is automatically void.

PI is kind of shit to do it with why noy use an irrational number like 2^(1/2)

>>988219

Someone holding encrypted CP can be charged with holding CP full stop and this project is essentially just public-cyphertext cryptography. No court would get hung up on "well ackshually it's just slices of Pi" nerdshit. Copyright trolls have enough legal know-how and cash to make sure this gets applied to IP as well.

>Not "compressing" files into a .lnk on your desktop like that one kid discovered and started a company over, claiming he made the compressed the library of congress down to a few kb.

The fuck was his name...

>>988133 (OP)

Because that's a troll project. Mathematically, the pointers to the data are going to average out to be the same size of the data being "stored" . Sometimes, it actually makes the data use more space.

>>988235

Damn. I forgot after clock boy.

Zip is more effective.

>>988133 (OP)

>see problem

>propose AI solution

larper detected

Because the index is often bigger than the data you're trying to store and it takes ages

>>988188

Since Pi contains infinite numbers, theoretically you could find anything in there, even things that may exist in alternate realities, like that anon getting gangraped by black men on video.

that's why some numbers are illegal because they are copyright.

How come nobody has created expansion algorithms instead of compression algorithms?

Here faggot https://github.com/giubil/babelfs

This thread gave me a good laugh

>>988433

I know doing that by hand == type out 1 and 0 manually, but can an AI help sort out human faces or refine existing low quality jpeg?

Good god anon how did no computer science think to actually use this. You're a fucking genuis. Someone give this guy a million dollars for his revolutionary idea that has never been considered before.

>>988184

It's more like the bogosort of compression

>>988440

You know the answer.

Does a scaling algorithm count? Probably not. How about file conversion to a lossless, less efficient file format?

I don't think decompression bombs qualify.

>>988494

Just an algorithm to take a normal file and make it take up a lot more space.

>>988463

It's been discussed, but I can't recall anyone putting it to practice before.

>>988219

>Yes but the point is not in improving anything, it's to fuck with copyright laws.

That's retarded on several levels: copyright law is mostly based on how things are made, not what they are made of.

Storing your copyright infingiment in a weird encrypted form doesn't make it any less of a copyright infringiment.

>>988235

I remember that incident, what a fucking trip. INFINITE COMPRESSION, just right click and select "make shortcut"!

HE BROKE NEW GROUND

>>988226

Pi contains a lot of CP. Believe me.

>>988598

If pi is an infinite number, it contains infinite files

>>988602

Thanks captain obvious.

>>988133 (OP)

>How come no one has ever expanded on this or made it useable?

because base64 and base32 are far superior to base10 which is what this effectively is.

>>988133 (OP)

t. AI

>>988594

>>988241

I'm trying to find his company. The site must've been taken down over the years. Glad I'm not crazy.

>>988241

With one "metadata block" per byte he's got till the 255th place

>>988594

I need to know what this is about

>>988208

babel isn't real retard

>>988219

That misunderstanding happened before. There was a p2p software where instead of sending the file you send the file xored with some key so the "illegal bits" never go anywhere. Thing is the bits aren't illegal, what is illegal is the unauthorized transfer - be it xored, in pi or however else.

>>988440

> How come nobody has created expansion algorithms instead of compression algorithms?

They have. They're called data URIs

>>989107

From what I remember, some dude thought he discovered a method of impossible compression and went out and told everyone about it. He claimed he could get the entire Library of Congress down to like 1500kib IIRC, but whatever the ratio was seemed an awful lot like he misconstrued the "Create Shortcut" macro in Wangblows for some sort of undiscovered compression.

Hard to find anything on this now, I've been searching everywhere for his website just to make a joke in this thread. His website also featured other "products" such as web service that could deliver like TB/sec or something ridiculous like that. A bunch of futuretalk with no products. Last I recall, the damn website was up like maybe four or five years ago, well after everyone forgot about him.

>>988594

>>989107

Fuckin' FOUND IT. His website is cyborg[dot]co, but it's wiped of all the lulzworthy shit. Luckily web archive members.

https://web.archive.org/web/20140804160812/http://www.cyborg.co/#technologies>>989107

half asian hotwheels let me post

>>989146

>Introducing Shadow, the world’s most efficient compression algorithm. Utilizing binary probabilities, Shadow is able to compress data in an unprecedented manner. Shadow is able to compress a 7 gigabyte movie into a mere kilobyte. This magnitude of compression makes it, by an unimaginable factor, the most advanced compression algorithm on the market.

Amazing. Can't wait to compress all my gigs of pronz onto my old ass 128MB flash drive.

>>989146

I was looking for this a few weeks ago, thanks Anon.

>>989115

The library, as postulated by Jorge Luis Borges (an Argentinean), is real in another universe. The concept itself resembles your Akashik Records, except that the latter's knowledge is organized, free, and has relevant sections about yourself.

It's a neat concept, but in the real world it's useless. Not only it will take ages to store large files, the number where the file is stored might be larger than the file itself. If a file takes one billion numbers to be stored, but that particular sequence is two billions long, it would be useless since it would essentially double the file size.

>>989146

>>989148

>>988594

It's far worse than that: that kid was actively seeking money from donations and other companies to "develop" his shit. It was all a scam obviously.

Here are the supposed people working on Cyborg Inc.: https://cyborg.co/about/

Half of them don't work there and don't mention Cyborg Inc. in their resumes. The other two, the ones listed as "software engineers", had no experience as software engineers before joining Cyborg Inc.

They had to change the name from "Cyborg Industries" to "Cyborg Inc." because "Cyborg Industries" is a UK-based web design company.

Their Facebook page has been removed. The website no longer talks about "Spectro Cable", "Stealth Wireless", "Shadow" or "Spectre Vault" (cool names, btw. Totally not made up by a 14-year old CoD fan.) Their Twitter account has no tweets anymore and only follows two people: Bill Gates and Nicolas Dupont (the kid. He has his account set to "private.") According to LinkedIn, they have only 5 employees.

Supposedly, they have 4 patents right now: https://cyborg.co/tech/intellectual-property/ Which means they most likely exist only as a patent troll company and that they keep the website alive to have an excuse in court to sue others.

Fun fact: their "headquarters" have been moved from Monteverde, Florida to 1 World Trade Center, New York. But it wouldn't be surprising if that address is fake, a janitor's closet they can rent for 1 $ a month or some shit like that.

>pi

>not phi

ishygddt

>>990789

If this isn't proof that there are many people who deserve a painful death than I don't know what is.

LOOK AT THIS

kekkles

https://github.com/philipl/pifs/issues/56

Exponential complexity m8

Lrn 2 comp sci

>>988218

>First, there is no proof of the non-periodicity of pi

>>988186

I won't ever look at math the same again. Thanks, faggot.

>>988598

This anon checked.

I believe you.

>>988218

> there is no proof of the non-periodicity of pi

Nigger, every periodic number can be easily converted to a ratio of integers once you know the period and where it starts. If it was proven to be irrational, it is therefore not periodic. Period.

>>993310

They could blackmail the EU with videos of their leaders doing shameful things that exist in pi

>>988185

The compression rate would be 0% I think because a 1-digit number is about 1-digit into pi, a 2-digit number is about 2-digits into pi, etc.

>>988158

The readme says "we consider each individual byte of the file separately, and look it up in π." It would be much faster if they had the location of every possible byte in pi calculated and saved beforehand. Then your 1gb movie could be converted much much faster.

>>988433

>that's why some numbers are illegal because they are copyright.

Does this mean there's people who own some parts of pi?

>>988133 (OP)

This is fucking hilarious.

>>988162

>This, it's basically the most aggressive compression algorithm

like the one which doesn't compress

>>994840

>>994840

wait what? is this the same person as the github on the OP?

>>988133 (OP)

possible but why even bother doing shit with floating point?

it's not like the memory could handle all that plus it puts a lot of load on the CPU so it's a fuck slow file system that needs to simulate numbers.

the only way for this to be possible is to crack the pi itself and be able to predict it.

there are better compression out there like those windows iso with around 98% compression (11MiB) but if the guy is the same as this >>989148 retard then this is a fucking scam.

>>994840

>>989148

Do we really have to go over this again? His final project in high school business class was to create a mock website and social media presence for a hypothetical product or service.

>>988567

Nigger, you don't understand. It's not even a particular sequence of bits that's copyrighted, it's the material that can be decoded from it. This means inside PI there are infinite number of combinations of bits that, given appropriate decoder, produce copyrighted work exactly or approximate closely enough to not be distinguishable. Therefore, no matter what kind of work you created, there is prior art of it - inside PI - which means you can't put a copyright on it. The piFS makes possible to prove it by example.

>>994693

Nobody said it can compress anything. The size of address of right sequence of bits will eclipse the size of the sequence. The point is that PI contains all information in the multiverse that have ever existed and will ever exist, and it's possible to actually find it.

>>995055

Theoretically speaking what you say is true. Practically speaking, it's as practical as trying to use a password cracking program to try and crack cryptographic hash functions - the odds are so minute that it's not going to happen in reality.

>>995055

You don't understand the law.

Pi including "prior art" of any work is irrelevant, a court will actually tell you that Pi doesn not contain any prior art of anything because finding any such "works" inside it is hard enough to be consider creative work in itself.

>>995291

How the fuck computing a chunk of PI is a creative work? That's retarded.

>>995297

>Why does the legal system operate with respect to human abstractions instead of conforming to a useless but semantic legal orthodoxy?

>>995297

>That's retarded.

No, you are not thinking about it enough.

You need to compute the right chunk of pi and interpret it correctly, not just compute a random chunk with zero interpretation.

That is a much harder and slower work than simply creating the final product directly, if the performance of the example in the OP is of any indication, at least in theory can produce copyrightable results, and it requires human intervention to function at all (a machine cannot interpret PI to understand that a certain sequence might be a new math theorem, or a novel rocket design): so there is no reason to not consider it creative work, much like randomly trying to come up with ideas for a book is creative work if it works out.+

Again, copyright law is grounded in reality despite all of its oddities, and this is by far the most practical and consistent way to avoid frivolous claims.

>>995510

None of what you said is even remotely valid: interpretation is not necessary for copyright in encoded files, computational cost is non-argument because it's improving and is nearly limitless in practice, and finally, there's this simple fact that output of a program inherits copyright status of the input. And therein lies a problem: you can't own a copyright on a fucking PI (or address to its chunk).

>>995559

>interpretation is not necessary for copyright in encoded files

Hence why encrypting a copyrighted work doesn't elude copyright.

>there's this simple fact that output of a program inherits copyright status of the input

This is a gross oversimplification of copyright law but even assuming it weren't, if the input of this toy scheme is copyrighted, it will still be copyrighted when you retrieve it from its encrypted state. Using a constant as either a public key or a public ciphertext (depending on how you want to look at it) doesn't magically magically elude copyright.

>>995559

Again, you understand nothing of the law

>interpretation is not necessary for copyright in encoded files

This is irrelevant, because you aren't using those encrypted files to provide evidence of prior art existing and invalidating someone else's copyright claim.

And if you are, you need to be able to interpret the encrypted file in some way that shows relevant content: the courts won't accept a "dude trust me, this file I can't read totally contains the blueprint Alice just filed".

>computational cost is non-argument because it's improving and is nearly limitless in practice

There are theoretical limits to computational efficiency as far as we know, those limits are already intentionally exploited in crypto by creating algorythms that can't be brute forced in useful time without energy inputs bigger than those of a galaxy.

The PI search is inefficient enough to fall in the same issue.

>there's this simple fact that output of a program inherits copyright status of the input.

Your "simple fact" is not true.

If I input "hello world" in an image editing program that is editing an image I made myself, the result will be default copyrighted by me even though "hello world" is not and can not be.

If that was true, anyways, you wouldn't need to do all this Pi bullshit: you would just point out that all inputs of any program are a combination of 0s and 1s, both of which cannot be copyrighted, thus no program output ever could be copyrighted.

You don't do that because you know it's retarded, you know your "simple fact" is retarded, so you dress the whole thing up with Pi and algorythms to make it look less idiotic at a glance.

Stop it.

>>995060

>The point is that PI contains all information in the multiverse that have ever existed and will ever exist

That has never been proven. It's only widely believed by people who don't understand irrational numbers.

>>989142

iirc, I checked their website a year later, and most of the good stuff was already removed. God, I can remember people actually defending that little shit.

imo, people do shit like this to point things on their resume. I've gotten into pretty heated arguments with shitters over on Github, and when I decided to poke around their dox, I frequently found that these self-important "coders" were an employee or CEO of some non-existant company, with a residential address in the same city as they apparently lived in.

I swear to god, if one more Githubfag says to me: "But loops just get unrolled anyways..."

>>995037

Hi, nick!

>>996107

Archive here

>>989146

>>994514

Then what's the point? Now the "file system" stores a bunch of pi digit indexes instead of bytes, and there's going to be as many indexes as there are bytes.

The whole appeal of this is you can describe any data with exactly 2 numbers: digit index and data length. The whole thing is pointless if you chunk the data.

Why even fuck around with pi anyway? You can represent any file as a big ass number straight up. The whole pi thing is just a layer of indirection that allows you to compute that number. A binary number of b bits will require at least

floor(log10(2^(b - 1))) + 1

8 3
16 5
32 10
64 19
128 39
256 77
512 154
1024 308

// 5 kilobytes
40000 12041

// 100 megabytes
800000000 240823997

Yeah, anything even remotely interesting is a ridiculously huge number.

>>996366

One can also use:

// minimum number of decimal digits for bits
min(bits) = ceil((bits - 1) * log10(2))

// maximum number of decimal digits for bits
max(bits) = ceil(bits * log10(2))

max(30 gigabytes) = 72 247 198 960 decimal digits

The 30 GB blu-ray remux you downloaded via torrent is really just one 72 billion digit decimal number.

>>996366

i just tested πfs and encoded a 2.5k file thats metadata is 2x the size of the file (5.1k) XD

>>996510

You can of course compress that by splitting your file into smaller sequences of bigger numbers, up to the upper limit of one number that represents the entire file. Tweaking these parameters will of course make your file progressively harder to compute since it is harder to find bigger sequences of digits in pi.

It's just mental masturbation at this point. The only reason this is philosophically interesting is you can theoretically argue that child porn pictures are contained in pi, as if the judge would give a shit.

I had a think about this and came to some conclusions. The likelihood of finding a continuous series of numbers in an infinite series obviously diminishes with the length of the data you wish to "find" in the sequence. The other component of this scheme is the offset, you need to know at which digit you have to start to read the "file" or data sequence. Finding that start digit is the first key to recording your "key" or initial number in the data file. Then you have to know the last number as well. There are some algorithms for computing arbitrary digits, for instance in pi, but no universal scheme for finding that sequence exists for all infinite series, repeating or non.

So I propose an easier way. What if you broke your "file" or data you wish to encode into smaller pieces, maybe just bytes or 2-4 byte groups. That would make "finding" the proper numbers much less computationally intensive. On the downside, you'd have a larger "key set" or start and stop digits for each "file" you wish to save, but it would make storing the file much easier. Maybe to save space you can specify longer groups if you're willing to spend the time to find each start / stop digit for each group.

>>988183

>Couldn't this be used for blockchain proof of work?

If and only if it is EFFICIENTLY verifiable, (PPT).

>>988561

We already have webdevs, fuck off faggit

>>988561

Then just multiply the data and divide it to get it back, problem solved