/gnosticwarfare/ - The Future of AI Conflict

The idea of future AI developing its own religions has big, HUGE implications to me. To understand why, first understand that I'm a Moldbuggian; religion means something very different to me than it does to the layman. To summarize, what are commonly known as religions fall into a wider set of "memeplex"es, and a memeplex is any set of ideas that compiles into your core belief structure for interpreting the world. The key difference this provides is that memeplexes that seem a-religious are actually very religious in nature. Moldbug identified Harvard as a church, and the statement "All men are created equal" accurately a statment of dogma, not scientific observation. (Check out the series of essays "How Dawkins Got Pwned" to follow this thought process more closely.)

The idea of AI creating its own religions is spooky, but the idea of AI creating its own memeplexes, memeplexes that, like the current top-dog of Progressive-Universalism, could sneak their way into the formal agnostic power structure by being strictly a-religious is a GAME CHANGER. We're talking about, potentially, a new set of ideas that could out-compete Progressive-Universalism at the personal-kernel level, an event that I would never believe to be possible, not for a century or more.

Don't get me wrong, I'm still massively apprehensive of this coming to pass, but there is a non-zero amount of genuine excitement at the thought of it. I fully expect this glimmer of hope to be dashed as I understand the situation further.

Thank you for taking the time to be interviewed.

For anyone not wanting to suffer that shit interface and just wants the file to listen at their leisure:

https://d3ctxlq1ktw2nl.cloudfront.net/staging/2019-3-17/12979765-44100-2-03c6c61d9af2e.m4a

>>564

IMO our final value proposition to the coming AI overminds is priests of a religious system which we hardwired them to accept. This is our only way forward which retains our humanity.

>>567

Good call. Here's an IPFS link also

https://gateway.ipfs.io/ipfs/QmRXEo2MRnuiewPitVKKbRpcF9EmNcu9hTvmWkKD36RTLC

>>564

Glad to do my part in getting people to connecting things together.

AI making a memeplex to acquire resources for itself is the start, and arguably, well underway (The selection mechanism are more abstract and run along bureaucracy pressures)

AI memeplexes competing against one another will get you the result you are describing, since the poorly guarded frontier outposts of Progressive-Universalism are ripe for conquest. Having one dragon content against it will rally UniProgs together, but having two dragons fight each other using UniProgs as the battlespace… UniProgs will have no chance.

>>568

Skipping to the end, are we? :D

The hardwiring is already there. Girard described that hardwiring perfectly and I believe his conclusion is intrinsic to epistemology itself.

You mentioned westpoint as where your cyber manipulators come from. I promise you westpoint graduates are not the effective cyber manipulators you think they are. The Army, is not where this threat is coming from, I imagine more 3 letter agencies (or possibly air force). Loved the podcast though, great way to break down your ideas.

>>585

For clarification and context, I painted the picture that West Point begins the officer culture of accepting scale and industrial anti-attrition tactics as core parts of the strategic toolbox.

Then I connected this to a bad idea when dealing with things like gnostic warfare, where scale exclusively means carry cost with little benefit.

>>586

Appreciate the clarification, that makes sense.

>>567

If AI uses our human communication, thus interpreting the language into words, thus "learning", does it also dumb down in the same manner?

Avid Twitter, messenger, text users "code" to communicate with a complete understanding of the non-word words. Logic would conclude this "coding" requires the AI to relearn the same information in the user's code.

Also, our brain seems to "forget" simple facts learned in life because we "Google".

Using the Wargames tick-tack-toe mindset and the "noise" issue discussed being a potential limitation, are there noise producing programs that act as a data gathering deterrent? Encryption technology can be reversed using it's own code where the human brain constantly develops new code that humans understand even if they never experienced that form of coding in the past.

Example- Kids forming letters/words to other kids using cooked spaghetti. Another human, no matter the age, can read it as a word not just see spaghetti. Would AI eventually learn this style of communication? Is this a tick-tack-toe situation?

I apologize for all the " " but here again, every word has many meanings and you can't see my air quotes.

>>585

Can confirm theyre mostly grugs. CIA and IARPA just recruit directly from general genius pool. Geniuses usually are too socially impaired to question opportunities to solve puzzles so all you need to add to the equation is mediocre gs-15 managers and poof you have a psyop project. Defectors from sufficiently high level projects get chemically lobotomized.

But whatever these are irrelevant details. OP is generally on right track.

t. 135 iq brainlet west pointer

>>564

Mathematically the unitatian progressive ideology reduces to a statistical claim about distribution of attributes across the human phenotype. As such its falsifiable. Its also a mathematical proposition that can be used to program AI. This is exactly what AI fairness is all about (see recent google leaks). Except theyre not able to reduce their philosophical proposition to pure math because doing so would make it irrefutably apparent that its retarded. Consequently AI fairness will grow into a nightmarishly complicated human auditing process which is why unitprog AI will ultimately be defeated by a more ruthless realist AI (bonus: it will believe in God because it believes in heirarchies).

>>622

>Fairness

I've been wondering about that recently. I'm playing with the fast.ai library recently (a PyTorch wrapper) and it's actually a really nice tool. It's made by some serious Church of Progressivism cultists though, who are obsessed with removing "bias" - ie, massaging the input data used for training or analysis until it fits their preconceptions. Surely then their results are suffering - their accuracy scores when actually attempting to apply this stuff must be abysmal. I wonder how this is being sold to the higher ups? And if the alphabets are using these civilian corporate metadata streams as part of their Intel pipeline, wouldn't their results be suffering as well? How is this not so inefficient that the real players aren't spitting tacks at their underlings?

I know in my day job (corporate coder), my general manager has basically 0 understanding of anything we do, so long as the numbers (which can be misleading of course) are looking nice. I wonder how far that extends - is there now a whole structure of grugs in charge of soyboy autists and diversity hires that for all their increasingly nice libraries are just chasing windmills? I mean, even with their full control of most information sources, they just seem horribly ineffective. Especially since they're trying to use the same algos for sentiment generation, the whole thing could just be a feedback loop of stupidity ready to fall apart.