rns.recipes

https://codeberg.org/MarSik/reticulum-audit/src/branch/main/reticulum-source-privacy-flaw.md

Has anyone reviewed this critique of reticulum's privacy / does anyone have any thoughts on it?

I have, in full. Putting aside the obvious LLM use (if we played count the em dash and verbose writing patterns we'd be here all night) there are some valid criticisms that are easily addressed if the author had a fuller understanding of the project. But the author is inherently biased against the project and fails to understand entirely what Reticulum actually is. They also make assumptions that are flat out wrong, and some wild comparisons to things like Bundle Protocol Version 7 which really puts their Dunning-Kruger on display (If this wasn't just an AI generated document)

They are also hyperfocused on TCP/IP as the sole mode of transport for Reticulum packets, while disregarding any other medium that the fabric could operate over. Which includes Onion hidden services and I2P (which works out of the box). As well as IFACs (Interface Access Codes) which can be employed when destination_hash and the originating Transport hash is of concern. Which is what 90% of this boils down to.

I see this making the rounds even on Mastodon, every once in a while. People get panic attacks without reading/understanding this 'audit'. The whole thing is incoherent and there are no rational conclusions.

Audit? laughing my fucking ass off

Dunning-Kruger indeed. This is what you get when people fall into the living fantasy that LLMs are actually "AI".

Everything "stated" there could have been boiled down to 3 short paragraphs, tops. The general problem with these things, is that in it's current full glory of expanding-foam insanity, if I were to interact with that turd in any way, I would have to pick apart and explain to the reader all the nonsense, fabrications and incoherence present.

That is, to actually present a coherent counter-argument, I would be the one needing to weave a rational thread out of something that is inherently irrational (but superficially may look "intelligent" to people who know nothing about the subject matter). Which effectively would make me the "I" in this moron's "AI" fantasy, and utterly waste the time where I could be productive instead.

So instead, I'm going to do the only sensible thing, which is to simply point out what a complete bafoon the person is, who just pressed a button and chose to fully believe in a wall of text from a stochastic vector storm, apparently enough to (I assume) stand by it so completely, that the idiot decided to post it to the world as something that actually had any value. How utterly and toe-curlingly embarassing.

I would fucking love to have an actual conversation with this man, wrested from his Anthropic subscription, and see how well his "auditing skills" fare then.

Even having spent 5 minutes writing this post kinda gives me a bad taste in the mouth, and I'm only doing so to point out how this kind of stuff is becoming, and will continue to become, more and prevalent, and is functionally only resulting in murkying the waters, creating confusion and wasting time, as opposed to actually increasing available intelligence and discernment.

A joke, but a very sad one, indeed.

Or to put it more plainly:

Every idiot now has the ability to press a button, upon which any kind of "argument" can be produced, essentially for free. There is no longer any investment of time necessary, which until recently was the primary boundary condition for being able to get other people's attention.

What these idiots don't realize is, that everyone else also has access to that button. If simply pressing it is all that their personal capabilities amount to, what they are doing is (per definition) practically, mentally, economically and spiritually worthless.

I am just not going to interact with that, except in one particular way. Life is too short.

At this weird, wonderful and horrible time in our history of human "civilization", what is most direly needed is discernment, and even though it has become a total cultural taboo to do so, I will forever reserve my God-given right to point my finger at morons who so willfully and totally choose to forego it in favor of mythologizing their lives into a living sci-fi fairy-land, and then promptly proceed to laughing my ass off.

I love good sci-fi, btw. Let's not make a religion out if, though, mkay?

the section on the networking implications is too funny

Zenith wrote:

I have, in full. Putting aside the obvious LLM use (if we played count the em dash and verbose writing patterns we'd be here all night) there are some valid criticisms that are easily addressed if the author had a fuller understanding of the project. But the author is inherently biased against the project and fails to understand entirely what Reticulum actually is. They also make assumptions that are flat out wrong, and some wild comparisons to things like Bundle Protocol Version 7 which really puts their Dunning-Kruger on display (If this wasn't just an AI generated document)

They are also hyperfocused on TCP/IP as the sole mode of transport for Reticulum packets, while disregarding any other medium that the fabric could operate over. Which includes Onion hidden services and I2P (which works out of the box). As well as IFACs (Interface Access Codes) which can be employed when destination_hash and the originating Transport hash is of concern. Which is what 90% of this boils down to.

indeed using an ifac and only accepting trusted users... which even on a large scale you shouldnt be accepting random messages from strangers like the NORMAL INTERNET. the exampleconfig, reading the manual in full, and actually knowing anything about basic networking, you can make it relatively secure. i finished reading it in its entirety and it goes in circles so much about basic exploitation and bad actors intercepting parts of the network. they even glaze tor which can also be compromised THE EXACT SAME WAY with enough relays and nodes under one person.