# Post-quantum cryptography & unrelated attack vectors on Reticulum Protocol

_General · started by Quackman1 on Sat, Nov 21, 2026 11:24 AM_

Tags: Question, Discussion

---

## Original post

**Quackman1** · Sat, Nov 21, 2026 11:24 AM

Hi Everyone,

Hope you're well.

First of all please note i have no coding experience most of my answers are based off of AI answers (hope that's fine)

I wanted to ask about post-quantum encryption. AFAIK, post-quantum encryption is open source, and Asymmetric parts (X25519 + Ed25519 & ephemeral ECDH keys) of Reticulum can broken by Shor's algorithm. Secondly I'm not sure whether AES-256 encryption is across the board. Finally obviously after discussing with AI, I've understood that the brutal constraint across the board is bandwidth costing kbs of data across crucial and precious air-time and almost impossible to travel via LoRa, so I'd like to know whats the thoughts on this (something like maybe make this optional?)

As far as unrelated attack vectors are concerned I've used opus 4.8 for this discussion, however I'm not sure how Fable 5.2 with ARTEX capabilities could fare with the live Reticulum Protocol. I appreciate Mark is the sole maintainer of the Reticulum protocol however I wanted to ask whether safety fallbacks are in place for stuff like a firewall policy (arbitrary packets), duty-cycle abuse on regulated bands, interface-level injection, sybil / spam at the application layer, endpoint compromise, etc. 

I know I'm mostly out of my depth here, and please correct me where I'm wrong, but all I'm trying to understand is whether Reticulum is a robust or anti-fragile network stack protocol that might survive the AI transition post-2030 combined with quantum technology and maybe even a state level bad faith external actor who'd like to compromise the network protocol.

Kindest Regards,
Le Quack

---

## Reply 1

**Zenith** · Sat, Nov 21, 2026 12:09 PM

PQC primitives are a hypothetical solution to a problem that doesn't exist yet. There is no way of knowing if Dilithium or other lattice-based standards can be "quantum resistant" because the hardware to try and break them doesn't exist. Previously endorsed NIST standards have also been vulnerable to attacks from classical computers (see SIKE https://en.wikipedia.org/wiki/Supersingular_isogeny_key_exchange) 

There are no quantum computers that exist which can be used for anything meaningful. There is so much pop science and other BS surrounding the word "quantum" that I don't have time to explain right now

Reticulum is already being tested against every possible vibesharter with unlimited Claude tokens in existence. This is one of the reasons behind the introductions of Identity and Destination stamps. I would recommend you go and read the posts here regarding 1.6.0 and 1.6.1, as well as the changelogs

If "post quantum" is in your threat model, design your network around an IFAC. With an IFAC no Ed25519 public key is ever broadcast. Again, dealing with hypotheticals, a highly stable, usable, quantum computer would still struggle with symmetric crypto like AES. Grover's algorithm would only reduce the theoretical security of AES-256 to the equivalent of AES-128. Reticulum adopted AES-256 in place of AES-128 in 0.9.5

---

## Reply 2

**Quackman1** · Sat, Nov 21, 2026 12:30 PM

**Zenith** wrote:
> PQC primitives are a hypothetical solution to a problem that doesn't exist yet. There is no way of knowing if Dilithium or other lattice-based standards can be "quantum resistant" because the hardware to try and break them doesn't exist. Previously endorsed NIST standards have also been vulnerable to attacks from classical computers (see SIKE https://en.wikipedia.org/wiki/Supersingular_isogeny_key_exchange) 
> 
> There are no quantum computers that exist which can be used for anything meaningful. There is so much pop science and other BS surrounding the word "quantum" that I don't have time to explain right now
> 
> Reticulum is already being tested against every possible vibesharter with unlimited Claude tokens in existence. This is one of the reasons behind the introductions of Identity and Destination stamps. I would recommend you go and read the posts here regarding 1.6.0 and 1.6.1, as well as the changelogs
> 
> If "post quantum" is in your threat model, design your network around an IFAC. With an IFAC no Ed25519 public key is ever broadcast. Again, dealing with hypotheticals, a highly stable, usable, quantum computer would still struggle with symmetric crypto like AES. Grover's algorithm would only reduce the theoretical security of AES-256 to the equivalent of AES-128. Reticulum adopted AES-256 in place of AES-128 in 0.9.5
> 
> 
> 

Hi Zenith,

Thanks for the response, Opus 4.8 (restricted because i mentioned ARTEX in the chat) gave me an outdated response that AES-256 wasn't fully on the board. As far as the rest of your answers are concerned its good to see a slight breakdown of the claude hypothesis of Reticulum encryption being slightly flawed argument in a more realistic way. Fable 5.2 Seems to disagree with you on some of the quantum computing points you've raised, but atleast you've helped me understand how AI doesn't always understand nuances and a realistic way of viewing things in the present atleast.

---

## Reply 3

**Nickie** · Sat, Nov 21, 2026 2:42 PM

**Quackman1** wrote:
> Hi Zenith,
> 
> Fable 5.2 Seems to disagree with you on some of the quantum computing points you've raised

Oh, for crying out loud...

---

## Reply 4

**Quackman1** · Sat, Nov 21, 2026 3:39 PM

**Nickie** wrote:
> **Quackman1** wrote:
> > Hi Zenith,
> > 
> > Fable 5.2 Seems to disagree with you on some of the quantum computing points you've raised
> 
> Oh, for crying out loud...

Hi Nickie,

Thanks for your response, I do appreciate my response might've sounded dismissive so my bad on that so lets discuss the actual technicals. Could you please expound on these points? - 

1) Harvest-now decrypt later : Traffic recorded today could be broken theoretically by the Shor's algorithm since epehemeral EDCH can be cracked.
2) SIKE was used against an isogeny-based alternate candidate never a standard and unrelated lattice. The Standardised schemes (ML-KEM, ML-DSA, finalised by NIST in 2024) rest on different maths.
3) "No way of knowing whether Dilithium or other lattice-based standards can be "quantum resistant" because the hardware to try and break them doesn't exist" - A good point but Signal, iMessage, Telegram deploy hybrid key exchanges since its no worse off if they fail.
4) IFAC is a partial workaround which helps against outsiders but not against a leaked key or a member.
5) Zenith's response was a satisfactory response to me, but I'd like to hear from you whether you think Reticulum is a safe, post-quantum AI led future communications protocol.

Note- These points were obviously written by AI and i acknowledge im out of my depth, so do let me know if i can't raise points summarised by AI.

Would love to hear your insight on this.

---

## Reply 5

**NONGKHAEM_BKK_TH** · Sat, Nov 21, 2026 4:40 PM

This has to be a joke post... please be a joke post...

---

## Reply 6

**Cleeyv** · Sat, Nov 21, 2026 9:51 PM

**Quackman1** wrote:
> Note- These points were obviously written by AI and i acknowledge im out of my depth, so do let me know if i can't raise points summarised by AI.

The norms around use of LLM-generated (aka AI) text for communication between humans is a newly emerging topic, both on this forum and everywhere else. I can only give you my opinion on this question, but I think many people on this forum will agree: 

If you are a human, and intending for your words to be read by other humans, then please do not use LLM-generated text for that communication. It is often interpreted as disrespectful. If you have not taken the time to write something, why should you expect any of us to take the time to read and engage with it?

Our sources of information often come from outside of our personal experience: from other people, books, websites, and increasingly from LLMs trained on these other sources. However, with any of these sources of information, we have to evaluate it (often using critical thinking) to decide if the information is trustworthy, or useful, or beautiful, or whatever other criteria we personally value. Various attributes of LLM output make this kind of evaluation even more necessary than it was for earlier sources. Pasting LLM-generated text is a red flag that there is an absence of this very important process of evaluation, and I strongly believe we should avoid doing it whenever possible. 

On the other hand I don't think it is helpful to be absolutist about the role that LLMs can have in aiding with human communication. I use automated translation services for written communication in second, non-english, languages that I speak and understand but don't write very well. However, I would try to avoid using it for languages I don't speak/understand at all because I'm not able to evaluate the output. The less we know about a topic, such as post-quantum cryptography in your case, the more careful we should be about the degree to which we are communicating information that we have a very limited capacity to evaluate. 

I agree that the topic of post-quantum cryptography is important for all secure communications protocols, including Reticulum. I have been researching this topic for a while now and considering posting about it here. The existence of this thread is now motivation to actually write something, so hopefully I will be able to do that sooner rather than later.

---
