Jérémie Zimmermann🎶💗🧀🫖

ᐸ3 Mostly offline - My life is music and music is life - Peace and GNU / [ask me] / Hacking with care! /
Sings in showers / cheese / Berlin / datalove <3 / #8bit / #16bit / #TeamFerment

Self-radicalized on the internet since 1995 (before it was cool).

Life, the Universe and Everything: 42
Music: is life
Analog:
LXMF: 63061093a3bf6a4d203a125a9045b556

Public Key

npub1t4jgjvhqpay96qds8r7ul3m0yx2hgpygalfl9g4zl0dq9ceqvevsx8yzkv

NIP-05 Address

jz_at_piaille.fr@momostr.pink

Profile Code

nprofile1qqs96eyfxtsq7jzaqxcr3lw0cahjr9t5qjywl5lj5230hkszuvsxvkgpz3mhxue69uhhyetvv9ujuerpd46hxtnfduqs6amnwvaz7tmwdaejumr0dsdrxgh0

Publishing to

relay.damus.io

nos.lol

Author Public Key

npub1t4jgjvhqpay96qds8r7ul3m0yx2hgpygalfl9g4zl0dq9ceqvevsx8yzkv

Show more details

Last Notes

2026-10-26 16:28:17 UTC

L'ex hors-piste
#ironeme #apeupres

2026-10-26 14:23:25 UTC

- reply

Also the difference with the assembler is that maybe 1 person in a million writes assembler today, maybe 1 out of 100 or 1000 within programmers.... when "knowing how to program", in any language is a thing entire schools were training generations to do, an entire class of workers that were useful for that very skills.
That all people knowing assembler be de-valorized on the work market and made replaceable would be hard... but much less so than the current de-skilling+commodifying of all programmers.
it's a bit like comparing the populations of 3-stars chef with anyone who cooks...

2026-10-26 14:18:42 UTC

- reply

The example of assembly works in a way: with current "free models" shipping with "open weights" but NOT entire access to the all of of the training data (which would be the only way one could re-train the model to get hopefully to similar results), it is like trusting a compiler that is proprietary, and that is 100% controlled as a black box by one of these evil ecocidal companies, ran by their sociopathic billionaires...
That would be the same ethical dilemna: for the sake of going _faster_ would you rather trust a closed box whose impact is provably negative and lose skills, or keep doing what you were previously doing?

2026-10-24 10:41:47 UTC

- reply

Sisi, il y a Sideband, codé par l'auteur de #Reticulum, Mark Qvist!

2026-10-22 16:42:49 UTC

- reply

Let it be clear: i am not saying that to criticize you. I don't judge you. I actually tried your firmware on the tdeck and liked it a lot!
But in what you say i see two things alarming:
- the cult of "speed" which is rather a thing of performance rather than of robustness, a metric of capitalism, which is NOT what i think makes _better_ software for everyone
- the dependency to these tools that is genuine, and that may go with a loss of the ability to do manual code and seriously review code (like a muscle, these when unused get atrophied...)
once the price for these things will start matching their costs, then what..?

2026-10-21 17:45:53 UTC

PHPocalypse Now
#PHP #lol

2026-10-21 11:00:13 UTC

- reply

I think it's premature to call microReticulum "unmaintained". today's LLM-driven standards of speed and 'productivity' imply that code be pushed all the time, but people take time. they have lives to live, various projects, etc.
Attermann in their last message seemed to mention they were struggling to find time. let's leave them time before announcing the death of something that isn't. there is enough confusion in this world already :)
(yet Pol i understand and share your frustration. I am genuinely sad that your contributions weren't given they importance they deserve, and hope they can nevertheless can be useful in the future <3)

2026-10-19 17:14:52 UTC

- reply

very sad to read as your work there seems really precious :/ I hope something can motivate you back in sometimes soon! <3

2026-10-19 15:13:27 UTC

- reply

why closing the MRs though? If attermann had problems, it may not last forever... maybe software is not unmaintained, but just had a pause, or is advancing slow, etc.?

2026-10-19 12:16:04 UTC

- reply

oh no wow that's very sad.
thanks for pointing it out. I hope that attermann gets better soon <3

2026-10-19 10:51:42 UTC

- reply

It is! Fully via MicroReticulum to start with, and via a few other implementations, at least one of which a maintainer of (EmComm) will be part of the meetup tomorrow in Berlin.
These things take time, but everything seems to be clicking together in place, especially in the last 6-12 months!

2026-10-19 10:32:55 UTC

- reply

part of the beauty of #reticulum transport-agnosticism is that you connect to it through anything (local wifi, BLE, lora, UHF, CB, whatever travels through air, etc. ) _and_ the Internet if/when available. so transition from internet to reticulum is somehow made easy by this. Internet is somehow one part of reticulum now ;)

2026-10-19 10:31:07 UTC

- reply

some of us are transitioning to #reticulum already! it won't happen at once but progressively. it's a process! better start now and take time...

2026-10-16 17:00:47 UTC

- reply

The RAK6431 has a 1W extension

2026-10-15 08:16:05 UTC

📖 On dira ce qu'on veut du Comité Invisible, mais sortir leur "Manuel de survie" en pleine insurrection populaire relève quand même d'un certain sens du timing et de l'Histoire! 🔥
https://lafabrique.fr/manuel-de-survie/
"Survivre à une civilisation déterminée au suicide n’est pas une mince affaire. [...]"
Sortie : 16 octobre 2026
366 pages
ISBN : 9782358723381
https://static.piaille.fr/media_attachments/files/117/443/926/293/591/154/original/b26bf76d2db6e85d.jpg

2026-10-14 15:47:41 UTC

- reply

use microreticulum for that https://github.com/attermann/microReticulum_Firmware
or reticulous (alpha software, vibe-coded, very impressive): https://reticulous.net/

2026-10-13 11:22:18 UTC

- reply

Its specified in the alt of the image ;)
it's https://sim-mesh.net/ just freshly released, (vibe-)coded by 2 of the people who frequently attend the #Berlin #Reticulum meetup.
It uses real code of ESP32, nRF52 and reticulum stacks in VMs, simulated the airwaves propagation using the ITU model and geodata including lidar elevation.
Nodes are added with specific positions, elevation, antennae (from a list of pre-defined ones, or that can be user-defined)
It then run tests when all this is programmed to act together in various ways.
Pretty badass if you ask me! ;)

2026-10-13 09:02:45 UTC

🔥 Are you tired of the internet? Have you heard of #Reticulum? In #Berlin?
📻 Join our quasi-monthly Berlin Reticulum Meetup, on TUESDAY 20th OCTOBER, 18h00
👥 Come to learn more, to share and show your projects, to join an effort in deploying more nodes in Berlin, to tinker, or just to share the pleasant and stimulating vibe!
🪛 at xHain hack+makespace, Grünberger Str. 16.
https://x-hain.de/en/calendar/events/2026-10-20-reticulum-berlin-meetup/
https://static.piaille.fr/media_attachments/files/117/432/765/097/624/082/original/e80650805a99de10.png
https://static.piaille.fr/media_attachments/files/117/432/769/408/644/184/original/601cbb57e61ace22.png
https://static.piaille.fr/media_attachments/files/117/432/776/828/194/832/original/9f4b2181a45a0b0e.png

2026-10-06 13:36:18 UTC

- reply

i was wondering about the power usage of these extensions, do you have any measures of those? my intuition was that the microSD would consume _way_ more than the 32MB flash, but can't imagine how much any of these would take on top of the base usage....
also for -30 (here in Berlin we used to reach -20, but unsure if this will ever happen again) i think we will need a different type of batteries than LiPo or LiIon... Sodium-Ion seems to be a good candidate...

2026-10-06 13:34:30 UTC

- reply

nono no confrontation, absolutely! hope it's clear on my end too! i am not questioning your knowledge in the slightest, just mentioning what i think are "nerd biases" and human biases in general: to look at something and appreciate it pretty much the same way we looked and appreciated the thing that came before it, the thing we're familiar with, etc.
clearly _nobody_ has experience yet with large-scale reticulum networks yet.. maybe we will, maybe we won't, it's just that we can make all sorts of plans and predictions, until we've experimented, large-scale, with multi-transport/multi-mode architectures!

2026-10-06 10:05:48 UTC

- reply

there was https://github.com/RFnexus/Reticulated
We also have two friends in Berlin who are working on something much much more advanced (spawning actual ESP32 and NRF52 code and reticulum stacks, simulating airwave propagation using ITU models on public lidar data, etc.) stay tuned as they may release it soon...

2026-10-06 09:32:32 UTC

- reply

RAK4631 have a 32MB flash extension (5$) and also a microSD reader (5$ as well). so again: thinking in terms of "the devices we got used to" or "the devices being sold today" (even omitting these extensions) is just limiting when it comes to imagining what can be done with technology that is at its infancy, and already offers 100x more possibilities than what exists and/or is currently being experimented with.
think of _all the computers that are less than 30 years old_. think of _all the "phones" that can be turned into Nura devices_, all the _routers_ that can get a RNS instance and number of interfaces. does it help?

2026-10-06 09:29:08 UTC

- reply

from what i see (i am no meshcore expert), there is no evidence whatsoerver that a lora-only network could cover a dense urban environment at more than say 1 out of 1000 person, or maybe 1 out of 100..
also from what i see, people who are experienced with meshtastic/ore are prompt to comment on reticulum, even if almost nothing has been experimented yet with it.
also, as a bias it feels like everyone and _especially_ people experienced with meshtastic/ore would consider anything like "a big lora-only network" (and would compare so they can argue).
the promise of #reticulum is vastly different than that. almost opposite!

2026-10-04 15:07:51 UTC

- reply

It sounds like you may be afraid for something that didn't happen yet..
I see a bias in this reasoning that may turn into a mistake, that sounds inspired by previous ways of conceiving and conceptualizing networks: nobody spoke of creating one-and-only-one network with #reticulum. We can get archipelagoes of small private lora networks, on different freqs, can interco with wifi with some other resources here &there, etc.
we all have an effort to break from those biases and imagine truly heterogeneous, composite, sometimes on/off networks that or may not be interconnected..
(also meshcore freqs can be avoided so fear not for your prepper toy ;))

2026-10-04 11:13:08 UTC

- reply

this is a reasonable scenario based on 1/ today's state of technology 2/ existing architectures
all i am saying is that everything is left to invent: parts of infrastructure will be made private, only for relays with important community roles, other parts will prioritize people immediately around, who funded it and maintain it, over the rest of the trafic. some links will be up during the day because of solar panels, but not at night, etc.
we will also have channel-hopping, frequency aggregating and adaptative TX.
i would be cautious with the use of "will" and "will not", and favour imagination, invention and experimentation.
we are learning.

2026-10-04 10:18:15 UTC

- reply

i think #reticulum over #lora is fine in many cases: small networks over large distance; small networks over medium distance; even in cities as long as the networks are not too dense. The key is really to think of reticulum as _never_ having to chose _only one_ transport. as long as in the densest city you feel like experimenting with #lora and _also_ have some local wifi, directional wif, BLE. [whatever comes next], etc. _altogether_ well you can deal with a saturated lora link. only when lora will be 100% saturated/unusable/all the time, it will be question to maybe stop (or reduce?) its usage. now we have the opposite problem: it's only starting

2026-10-03 14:53:36 UTC

- reply

such nomadnet walls/chats still make an entire group depend on a central point, very similar so depending on a _server_ for communications to exist....

2026-10-03 14:51:12 UTC

- reply

to me lora is great for one thing and one thing only: it's low power usage.
so those hybrid-media networks will maybe work fine, but question comes back when energy is down due to power blackout or other events: while we have experience with keeping lora nodes alive on little solar power and battery, what happens when the rest of the infrastructure is not available anymore?
we'd have to extend the robustness and frugality of lora-only nodes to some lora-wifi at least, and some gateway nodes maybe..?

2026-10-03 14:48:25 UTC

- reply

absolutely! nobody said that lora-only networks were a good idea for dense urban environment! in those, network architecture has to be opportunistic as to take what is there, and that's precisely what reticulum is built for, and apparently pretty good at!

2026-10-03 14:32:30 UTC

- reply

packets are slightly bigger because of cryptography (ie. routing like path requests, responses, announces, are also encrypted) and the fact that lora has minuscule bandwidth that would inevitable crumble when dozens of parties are using it at the same time. adaptative TX power, multi-channel/channel-hopping techniques could be used to push these limit a bit further... but when talking of cities of millions of people... divide lora bw by that and you get to these conclusions.

2026-10-03 10:34:47 UTC

- reply

Again i look at it with some distance: if Reticulum blossomed in an exponential way in the last 6-12m, it's still at its infancy, and those things may take time, but naturally emerge somehow. (an app that is not vibe-coded AND coïncide to an æsthetics that nerds consider "user-friendly" and/or that gains traction).. i would see this as a "detail of implementation" in a scenario where things continue to grow/get traction.
agreed that in its current form reticulum is not fit for a lora-only network the size of a city. but nobody mentioned this would be a good idea anyways (nor would that work with meshcore?). lora+wifi+BLE+..+..on other hand..

2026-10-03 10:31:01 UTC

- reply

nomadnet clearly has a nerdy æsthetics (that i personally adore), targeted at people with a technical apetite, and/or willing to learn. Sideband on the other hand acts like your "normal" android app (the exact opposite somehow). I think Columba, Ratspeak, etc. on the vibe-coded side have an appeal as a "modern" standardized app, clearly not what i go for.
Dont wan't to open that can of worms, but i dont believe in "users" neither in them being so stupid that they need genius techno-elite to design special things for them. I have seen people learn everything if they had the right motivation, and #reticulum doesnt lack good motivations.

2026-10-02 13:38:46 UTC

- reply

yep. "fast" is a value for those who optimize numbers, not the well-being of people. it is used to standardize in industrial logic, to make things like products, predictable, and people and their jobs replaceable.
I think beyond the science-fiction rationale, the promises of sentient "buddies" or whatevers, beyond the fascination for the tool, etc. it is all about "good'ol" industrial logic to get rid of creativity, unpredictability, reduce "the human factor" to numbers and turn everything into profit to dominate, accumulate and sell, sell, sell....

2026-10-02 13:08:18 UTC

- reply

..reinforces my intuition that vibe-code is like fast-food:
🔥 wasteful, hurts ecosystems and people
💗 not best solution to remain healthy, better options exist that require to learn skills and work can be shared, mutualized and lead to personal enjoyment
💰 controlled by few actors, relies on huge-scale economies
🤤 makes people who use it lose sense of what is actually food/code
📺 a socio-cultural trend dictated by marketing and fashion
#LLM seem to already have durably changed habits; we'll see how much of it was hype when prices get real, bubble collapses etc.
🗑️ i think a lot of that #fast-code will end up -as fast as it was produced- as abandonware.

2026-10-02 12:59:04 UTC

I spoke to a few proponents of vibe-coding: some people i deeply respect, some with deep experience of engineering, some i would say would be the best qualified for actually _knowing what they're doing_ when they engineer, review, and test code output by these ocean-boiling, sociopath-owned #LLM monstrosities..🤖🔥
..in most cases one of the main arguments is (it's easy &) _it is fast_. "can do so much in so little time!" "life too short!", "more productive", etc. it all converges abt speed.
Funny as it never struck me that software dev was _too slow_! on the contrary one could argue that taking time is actually a feature for more stable/durable software.🤔

2026-10-02 11:55:21 UTC

I said already that I love #Reticulum.. yet i just thought of it this way:
As Reticulum is transport agnostic (a node connects through many "interfaces": local wifi, BLE, lora radio, UHF, FM radio, etc. _and_ TCPClient interfaces), the fact one can actually use _internet_ is the 🐘 in the room (most ppl probably do connect to reticulum via internet today?) and also an amazing feature:
1/ it enables super-easy onboarding. basically install an app, some will add TCPClient interface for you, and voila.
2/ it enables to test features, routing, new protocols and ideas at large-scale, while learning how everything works, and becoming ready for the next thing.

2026-09-30 12:59:17 UTC

- reply

il n'y a pas d'alternative à cloudfront, cloudflare, etc...
la seule façon c'est d'embaucher les gens expérimentés et réactifs qui auront les moyens techniques de dimensionner l'infrastructure en fonction de la charge, de réagir aux attaques, DDoS, etc.. c'est difficile, coûteux, etc. mais c'est la seule chose à faire plutôt que d'externaliser sa "police" et toutes les données perso de tous ses utilisateurices/lecteurices/soutiens financiers/ etc. à des entreprises dont le modèle est la surveillance et dont les intérêts sont irrémédiablement liés à ceux du State Dept.

2026-09-30 11:32:13 UTC

Les troupes de Donald Trump débarquent en France? Zuckerberg, Musk, Bezos et leurs potes vous espionnent dans vos lectures? Le complexe militaro-industriel parasite la presse? Tout ça à la fois?
Nonon, enfin... oui, c'est juste @npub15uf…kjwd qui organise un "live".
cloudflare, cloudfront, fastly, google, windows.net, unkpg, qiota ("..analyser et augmenter vos kpi digitaux"), etc. 😱
Un festival!
J'apprécie le travail de Mediapart, mais dur de s'appeler "indépendants" quand on choisit de dépendre de tant d'entreprises US que l'on laisse infiltrer son lectorat..
#Triste #PresseIndépendante #Mediapart #ColonisationNumérique #SurveillanceDeMasse #WeLostTheWeb
https://static.piaille.fr/media_attachments/files/117/359/727/801/741/070/original/499217b8efecf712.png

2026-09-30 10:19:07 UTC

- reply

This is amazing, and extremely important work in the continuation of what A/I has been doing.
Ideally afterwards we will gear up, research and prepare for a future in which _there will be no need for servers_ anymore, and the hierarchy (and chain of dependency) they imply!
Once you are rested from this amazing and impressive work you've accomplished here, take a look at #Reticulum, for a post-internet, non-hierachical, encrypted-by-default future, where everyone can co-own the infrastructure by becoming a part of it!
Be ungovernable, be uncensorable, be server-less.

2026-09-30 01:51:38 UTC

"... cela va se propager comme une trainée de proute!"
#ironeme #apeupres #memepas #proute

2026-09-26 11:20:01 UTC

Tired: internet
Wired: reticulum
Saddened by the loss of internet? in #Berlin?
Join the Berlin #Reticulum #Meetup on Tuesday Sept. 29th, 18h00 at the xHain hack+makespace
Grünberger Str. 16 (in the "lobby" by the entrance) 10243 Berlin
https://x-hain.de/en/calendar/events/2026-09-29-reticulum-berlin-meetup/
Whether you are curious or already running your own nodes, join an informal meetup to discuss all things Reticulum! We shall once again discuss, share, experiment and tinker around, so bring your projects, hardware, friends and ideas! <3
https://static.piaille.fr/media_attachments/files/117/337/064/495/384/365/original/3565d6636bc9c8d6.png

2026-09-24 10:49:05 UTC

I never quite _enjoyed_ coding. I learned since I was a child, have gone through bases of Basic, Logo, Pascal, C, x86 ASM, C++, bash, Java, PHP, Javascript, Python, Ruby... but always only the surface.
Never considered myself "a coder", always found people waaay more patient, experienced and talented than I was, whenever i needed something (and with free/libre software, i found that the whole world helped me with code when i needed some! <3). never had that itch...
Yet, with recent developments in industry, coding not being anymore a desired skill, #LLM induced de-skilling, etc. makes me feel like coding!
Feels like keeping ancient knowledge alive...
https://static.piaille.fr/media_attachments/files/117/325/608/964/793/372/original/4d3c752bb602efc3.png

2026-09-23 12:49:05 UTC

Am asking the question again because nobody provided an answer so far:
🤔 Can code coming from an LLM ever truly be considered _free/libre software_?
1/ Can the person who would claim authorship over it be truly considered the author, if all they did was design and some engineering tasks like code review, designing tests, etc.?
2/ Can one affirm that this code is not plagiarizing existing code licensed differently, and/or proprietary code by a model that has been trained on it and is potentially spitting back entire portions of it?
#FLOSS #LLM

2026-09-19 21:41:43 UTC

- reply

100x YES! Just, let's be humble: #Reticulum is amazing, the most exciting thing i have seen done with computers since the 1995 internet! Yet it's young, and far from complete (group destinations not implemented yet, LXMF propagation node mechanism very rudimentary, etc.) my point is: We don't know if it can handle billions of people quite yet. Let's grow reticulum together, steadily. Let's be part of it in an active and benevolent way. Let's get our friends onboard in concentric circles. It may take time, but we have time. Lots of excitment... but no rush! ;) <3
see you on RNS! ;)

2026-09-18 07:38:35 UTC

- reply

let's make a few scalability tests first, but yeah... :)

2026-09-16 17:36:49 UTC

- reply

vois-ça comme un nouveau TCP/IP, qui utilise des interfaces qui peuvent être lora et wifi et ble et tcp, etc. tout en même temps! qui route au travers de toht cela, chiffré par défaut, sans centre... il y a plein de protocoles au dessus comme LXMF (pour les messages) LXST (streaming audio), RRC (IRC), rngit, rnsh, etc. le futur, j'te dis!

2026-09-16 17:33:18 UTC

- reply

bravo! c'est ce que les vieux singes des ondes m'ont appris: puissance d'émission, amplification relative de l'antenne ne sont pas aussi importante que la bête.. hauteur! le long bâton est plus puissant que l'ampli...
maintenant il vous faut passer à #Reticulum pour vous amuser sans cesse! :)

2026-09-16 10:33:43 UTC

- reply

because computers were built for ballistic calculations? because Arpanet's objective was Command&Control? for the same reason protection of people's privacy is described in terms of "operational security" and "threat models" (terms borrowed from the military)... maybe IT and the Internet always has been _militarized_? (this begs for an "always has been" meme....)

2026-09-15 09:25:37 UTC

- reply

there is already a search in announces!
searching in nomadnet pages would be useful indeed! hopefully it can be the same logic sometimes afterwards to also search into the conversations! that would be _extremely_ useful to me!.but... one step at a time! congrats on your current achievement! ;)

2026-09-12 22:11:10 UTC

reticulum$ git pull
Link established with remote
Transferring: 100% (128.27 KB).
Unbundling objects: 100% (131/131), 125.13 KiB | 17.88 MiB/s, done.
Resolving deltas: 100% (117/117), completed with 38 local objects.
From rns://7649a50d84610232d1416b41d2896aff/reticulum/reticulum
0bb41bf94..1565126ff master -> origin/master
Updating 0bb41bf94..1565126ff
Fast-forward
Changelog.md | 35 +++++++++++++++++++....
"... From rns://7649a5..." 💗
#rngit #reticulum

2026-09-12 12:45:07 UTC

- reply

now that you mention it I am not sure it is yet, "officially" in the rnode firmware... but there was a way to build it in some old forum post. so it may be a matter of effort, or time.
It is intergrated into microreticulum: https://github.com/attermann/microReticulum_Firmware/pull/45/changes/45adc0d083f76aa4821a4caa44fd2f653656ed2b
(don't know if you can use microreticulum as an rnode actually? i think so...)
also as of now there is at least this (rust, vibe-coded) stack https://git.emcomm.cc/berlinmesh/reticulum that runs as an rnode (and much much more) on the xiao+wio
(that's what i ran on mine)

2026-09-11 19:05:56 UTC

- reply

RNS 1.5.4, nomadnet 1.4.3 released ! What a day for #reticulum! :)
RNS: "1.5.4 2026-09-11 • 11 artifacts • Latest
- This release improves RNode BLE connectivity reliability on desktop operating systems."
Nomadnet: "1.4.3 2026-09-11 • 5 artifacts • Latest - This maintenance release fixes a few bugs."

2026-09-11 09:32:01 UTC

#reticulum service announcement:
RNS 1.5.3, nomadnet 1.4.2 released
RNS: "1.5.3 2026-09-10 • 11 artifacts • Latest - This maintenance release adds functionality to[o], and fixes bugs in rngit, and includes a few other optimizations and improvements."
nomadnet: "1.4.2 2026-09-10 • 5 artifacts • Latest - This release completes the foundations for page image caching, and adds an automatic image conversion handler to the page node. This means that it's now possible to use images in any format directly when hosting nodes, as everything will automatically be converted to the required webp format."
pipx upgrade-all, git pull and/or the such! \o/
https://static.piaille.fr/media_attachments/files/117/251/705/195/766/390/original/d941ee033b09365c.png

2026-09-10 23:46:50 UTC

- reply

I wish there are soon sorta URLs for RRC... like rrc:addressofthehub/#room that you could copy/paste, ingest into a ctrl+u in the RRC tab of #nomadnet, even cilck, etc.

2026-09-10 11:22:06 UTC

- reply

from my beginner-level's perspective, all the rnodes and autonomous nodes i have seen embed a LoRa SX1262 transceiver.
Get any of these RAK4631, Xiao+Wio or any combination of nRF62+SX1262 OR any combination of ESP32+SX1262, and you easy-peasy turn that into an rnode (a "lora modem") that you connect to via USB or bluetooth (or wifi on ESP32) and #reticulum starts chirping over it.
In theory i think that any other cheap modulating on these frequencies could, through maybe the writing of custom drivers and software, also get you to send reticulum packets.
There may not be such software ready yet for other chips than the SX1262, i don't know...

2026-09-08 13:09:45 UTC

#SaveTheDate
#Reticulum #Berlin Meetup
Tuesday, September 29, 2026
18:00 – 21:00
xHain hack+makespace, Grünbergerstr. 16
https://www.openstreetmap.org/?mlat=52.51279&mlon=13.44981#map=16/52.51279/13.44981&layers=P
Reticulum is post-internet, unstoppable, encrypted mesh networking for all. Whether you are curious or already running your own nodes, join an informal meetup to discuss all things Reticulum! We shall once again discuss, share, experiment and tinker around, so bring your projects, hardware, friends and ideas! <3
https://x-hain.de/en/calendar/events/2026-09-29-reticulum-berlin-meetup/
https://static.piaille.fr/media_attachments/files/117/235/575/423/718/646/original/514b984ea831577a.png

2026-09-07 20:19:53 UTC

- reply

#Nomadnet (terminal-based app created by #Reticulum author Mark Q, nimble and useful as a "reference implementation" of everything..) is making giant progress!
💬 "Conversations" tab has so many more options: ctrl+v to switch delivery modes, a "cancel" button while sending, detailed msg info, QoL things
📡 "Interfaces" tab more fluid to navigate (ctrl+o to on/off interface), with more info on display and a "profile" menu where you can manage different configurations for different places/tests/etc. so useful
🛠️ new "Utilities" tab packages all the basic diagnostics
📰 new "My Node" with file manager and micron editor for local node
⚙️ new "config" tab
✊ ❤️‍🔥 🪐
https://static.piaille.fr/media_attachments/files/117/231/550/976/880/465/original/b94aa7883bcf9d9a.png
https://static.piaille.fr/media_attachments/files/117/231/554/985/351/161/original/6db0e9d79e86352d.png
https://static.piaille.fr/media_attachments/files/117/231/586/847/741/815/original/a1cc9eeec18ad406.png

2026-09-07 17:58:43 UTC

- reply

... and today #nomadnet 1.4.0!
#reticulum

2026-09-06 01:42:25 UTC

- reply

oh nice! a new "Tools" tab, to go with the new "My Node" tab!
i just noticed the new config editor... through which i start to learn about half the config parameters that i didn't know of and/or are new since a few versions. wow.
also the interface navigation is much better, and info screen per interface gives more data, and a nice interface filter
i also briefly noticed it was more "graceful" and/or explicit when it came to hooking back up to a rnsd after being disconnected...
one can tell that quite some love has been put into that!
<3 #reticulum go [@mark@federated.channel](https://federated.channel/profile/mark )
! ✊
go @npub16qq…n3m8

2026-09-05 17:55:24 UTC

- reply

the repo is not stalled at all! It just moved away from github!
The coolest way of getting it is directly via rngit:
try:
git clone rns://7649a50d84610232d1416b41d2896aff/reticulum/nomadnet
(and fetch it from the Aleph rngit of Mark Q.!)
you can browse this nomadnet node that describes the rngit:
a8d24177d946de4f1f0a0fe1af9a1338:/page/releases.mu`g=reticulum|r=nomadnet
(and verify identity based on his public reticulum identity)
otherwise
pipx upgrade-all
does it for me, as i did pipx install nomadnet

2026-09-05 17:09:04 UTC

- reply

wow the new "My Node" with integrated file browser, micron markup editor, permission settings, etc. <3

2026-09-05 16:43:31 UTC

Wow! not paying attention for a few days, and #nomadnet was boosted from 1.2.9 to 1.3.1! Can't wait to see the new additions...
#Reticulum's emergence is really beautiful <3

2026-09-03 20:44:04 UTC

- reply

- rnpath -t (the whole thing, to filter by hand), rnpath -t -m 1 (only to one hop), and rnpath in general, always useful ->
<xxxxxxx> is 3 hops away via <yyyyyy> on [interface] expires 2026-09-10 10:08:00
- rnprobe is the equivalent of icmp ping in the ip world ->
rnprobe -v lxmf.delivery zzzzz
Sent probe 1 (16 bytes) to <zzzzz> via <zzzzz> on RNodeInterface[rnode LoRa]
Valid reply from <zzzzz>
Round-trip time is 1.024 seconds over 1 hop [RSSI -7 dBm] [SNR 13.5 dB] [Link Quality 100.0%]
Sent 1, received 1, packet loss 0.0%
- rnstatus useful too for generic info and following trafic on the interfaces
#reticulum #toolbox

2026-09-02 19:07:58 UTC

- reply

Note: i am not trying to make it impossible through argumentation.. I could get convinced that that free/libre LLM models, available along with their training dataset (so one could potentially re-train them and to some extent get some reproducibility of the model?), ran on one own's infrastructure (or community-based, mutualised, etc. ie. not owned by some evil billionaire), with a focus on energy-efficiency, frugality maybe, could indeed be used for good, and provide quite some services with some menial/repetitive tasks (boiler plate, transcription, tests, etc.) and others, under strict validation/audit by humans, in free/libre software dev.
@npub13ah…4c8l

2026-09-02 18:59:23 UTC

- reply

I actually put the last bit in parenthesis, in relationship to training data... but in all truth i think this is maybe _the_ most fundamental question Free/Libre #FLOSS software users, authors, supporters and overall communities will need an answer to, in relation to #LLM/#GenAI:
If some code is output by [machine], can one:
- claim to be the sole author of it and chose which license to apply to it?
- declare it free/libre software, IF one cannot prove that it is NOT based, at least in part (training data), on other software that may be released by its authors under incompatible license, or plainly proprietary?
@npub13ah…4c8l

2026-09-02 18:12:23 UTC

- reply

Faint hope ✊:
The blunt, unsatisfactory decision that has emerged from #Debian is sorta least common denominator that means something to someone but something else to somebody else... Yet maybe an attempt at a statement: "_irresponsible_ or _non-reponsible_ use of #LLM/#GenAi is not acceptable."
So now would be time for #Debian and #FLOSS communities at large to lead the way and determine: what would be "_responsible use of #GenAI_"?
Is the use of LLM/GenAI "responsible" if:
- it boils rivers?
- it belongs to evil US tech overlords?
- its model is proprietary?
- its training dataset is not public? (then can anything even be Free/Libre software?)
@npub13ah…4c8l

2026-09-01 16:36:46 UTC

- reply

missing: collectively owning the infrastructure, rather than "chosing a provider" (and depend on them) check #reticulum, the missing link in your equation maybe..?

2026-08-31 08:30:31 UTC

- reply

ah merci pour la précision! Je n'ai utilisé ratspeak que dans sa version cardputer (un fiasco, des bugs très profonds qu'aucun "vibecodeur" n'a eu la patience de regler pendant longtemps... et autres comportements étranges très difficiles à diagnostiquer -sans doute du fait de la nature spaghetti-code dudit vibe-code..) et encore là il fallait entrer ses paramètres à la main.
je ne suis pas sûr que les paramètres par défaut de meshtastic/ore soient une si bonne idée que cela, car cela crèe de la concurrence qui n'est pas nécessaire.. en étant sur 869.475 avec 125KHz à Berlin on évite savamment (!) leurs freqs. ça en fait plus pour nous... ;)

2026-08-31 08:27:40 UTC

- reply

juste des groupes informels locaux cela suffit amplement! pas besoin de bureaucratie, etc.. ;)
un forum, un canal RRC, une réunion dans un hackerspace (qui peut devenir mensuelle?) etc.. est largement suffisant pour un (bon!) début...

2026-08-31 08:05:41 UTC

- reply

alors "long fast" en #reticulum je crois que ça existe pas vraiment, vu que tu choisis
- une frequence de base (qui dépend des régulations, mais ausis des équippements et du bruit environnants..);
- une bande passante (en KHz);
- un "spread factor" (plus c'est haut plus ça va loin et plus la BP est réduite);
- un "coding rate" (qui augmente la fiabilité/distance au détriment de la BP, d'habitude 5 est une bonne valeur);
c'est important de s'adapter à l'environnement et d'utiliser aussi si possible une frequence qui ne marche pas sur les pieds de meshtasticore pour maximiser l'espace/BP/confort pour tout le monde..

2026-08-31 07:27:13 UTC

- reply

le spreading factor de 8 est une recommendation vraiment hmm "bizarre", car elle revient en pratique à quelque chose comme 2kbps... pourquoi ce choix? "parce que" selon l'auteur de ce post, et c'est tout...
Or, et c'est bien l'intêrêt de lora, ces paramètres sont mutables pour une bonne raison: pour s'adapter à des situations diverses...
SF=8 peut faire sens dans des zones très peu peuplées ou il est indispensable d'aller plus loin que vite... mais en ville par exemple, c'est ridicule. A Berlin par exemple, les radio-nerds-du-Chaos ont élaboré un consensus jusque SF=7

2026-08-30 22:06:12 UTC

- reply

ask @npub12y2…rkev !

2026-08-30 22:04:34 UTC

- reply

quelles fréquences utilisez-vous par chez-toi, et comment/pourquoi? (je m'intéresse à ces consensus, durs ou mous, par lesquels les frequences et paramètres lora sont déterminés...)
je me dis aussi: il pourrait être utile de détecter des nœuds meshatasticore autour de chez soi.... pour ouvrir avec elleux la discussion sur l'opportunité de leur adjoindre (ou de les convertir en...) nœuds #reticulum!

2026-08-30 21:57:57 UTC

- reply

"The state gains leverage even when it gains no plaintext from A/I’s encrypted disks: people separate themselves, providers collect more information and the social relationships surrounding dissent become easier to observe.
The precedent also reaches beyond explicitly anarchist or antifascist projects. If political alignment with users, privacy protections and the hosting of controversial publications can be assembled into a case that infrastructure itself constitutes support, then encrypted mail providers, radical publishers, community archives, VPNs, [fediverse], legal-support projects and other independent hosts all have reason to pay attention. "
7/5

2026-08-30 21:56:21 UTC

- reply

#AutisticiInventati
"The resulting harm is not limited to censorship. It can change the architecture of movement communication. Small autonomous providers may conclude that serving controversial communities invites existential risk. Larger platforms may point to the designation as another reason to expand identity verification, automated moderation and data retention. Users may migrate from trusted movement infrastructure to commercial systems that are easier to monitor, subpoena and map."
6/5

2026-08-30 21:46:49 UTC

#AutisticiInventati
Treasury Department’s Office of Foreign Assets Control placed A/I on the Specially Designated Nationals and Blocked Persons list under Executive Order 13224.
"Building communications infrastructure for radical movements can itself be treated as terrorism."
https://web.archive.org/web/20260828071449/https://cavallette.noblogs.org/2026/08/10083/2
If the same logic should be continued, then building communication infrastructure for war criminals can itself be treated as war crime?
Google, Amazon, OpenAI, Anthropic, Amazon, Microsoft are thus war criminals?

2026-08-30 11:33:07 UTC

- reply

...That remains the lesson. Do not leave the knowledge with experts.
Do not leave the infrastructure in one place. Do not leave the targeted to stand alone.
The first server was called Paranoia.
The network it created was called solidarity.
#AutisticiInventati #Autistici

2026-08-30 11:31:48 UTC

- reply

Between now and September 25, preserve the history. Share the story. Build the coalition. Document the isolation. Teach the tools. Prepare a lawful defense. Do not allow the state to quietly establish that building privacy-preserving infrastructure for dissent is itself an act of terrorism. Twenty-five years ago, ten people gathered around a recycled machine and deliberately took too long configuring it because everyone in the room was supposed to learn. That remains the lesson. Do not leave the knowledge with experts.
#AutisticiInventati #autistici

2026-08-30 11:25:32 UTC

- reply

"Its manifesto describes the project in explicitly political terms. A/I is antifascist, anti-capitalist and antimilitarist. It promotes free software, encryption, pseudonymity and the independent circulation of knowledge and cultural work. It reserves its limited, volunteer-run resources for people and projects broadly compatible with those principles rather than businesses, political parties, organized religions or institutions already possessing the means to make themselves heard."
#autistici #autisticiInventati
[@smaurizi](https://mastodon.social/@smaurizi )

2026-08-30 11:17:06 UTC

- reply

The reaction and plan for action of #Autistici/Inventati:
https://web.archive.org/web/20260828071449/https://cavallette.noblogs.org/2026/08/10083/2
"They named the machine Paranoia. The following month it went online, carrying two domains and two intertwined histories: Autistici.org and Inventati.org. Twenty-five years later, on August 26, 2026, the United States designated Autistici/Inventati a Specially Designated Global Terrorist and placed it on the Specially Designated Nationals and Blocked Persons list.
[...] a collective that has spent a quarter century insisting upon one dangerous proposition:
People should be able to communicate without first surrendering themselves to corporations or governments."
[@smaurizi](https://mastodon.social/@smaurizi )

2026-08-30 09:05:40 UTC

- reply

also (and in relation to 5/5): "As software engineers we have front-row seats for the deployment of one of the world’s fastest growing energy consumers: AI. Our conscience calls us to resist the deployment of technologies which will exacerbate death and suffering for millions. AI might deliver benefits to society, but the cost is so high that we need to make careful decisions about how it’s deployed and if the benefits to society outweigh those costs. [...]. Cheap, green electricity, preservation of freshwater resources, husbanding our planet and its wildlife: these are all much more important than AI-assisted coding."
#SourceHut #AI #LLM #environment

2026-08-30 08:55:16 UTC

- reply

yes it seems that if you don't set your LoRa interface (rnode) to "access point" mode (which seems to be the default now), in any other modes if there is a TCP server interface somewhere in the mix, lora will end up attempting to forward all the route requests and announces it sees, which is bound, with its limited bandwidth to excess the airtime... your own "legitimate" traffic may end up being stuck in an endless queue... it may eventually arrive though, that's an intersting experiment...
but fundamentally i think lora interface non-"access point" cannot work given the current size of the #reticulum mesh (including all of TCP nodes...)

2026-08-30 08:51:34 UTC

- reply

About free/libre software-related aspects of #LLM #vibecoding, read the stance of #SourceHut, :
https://sourcehut.org/blog/2026-08-27-tos-changes-and-llms/
"[“vibe coded” projects] tend to, as a class, be outliers in terms of resource usage, as the practice tends to invent complex CI configurations that waste build minutes, produce larger codebases, and push commits often, all of which causes excessive resource usage and can lead to outages. Often these codebases are abandoned after burning through all of these resources for naught. [...] #AI is causing skyrocketing costs for the hardware we need [...]. But, in truth, the rationale is more about politics and ethics than anything else."
https://static.piaille.fr/media_attachments/files/117/183/601/525/209/906/original/441fc07d82924f9a.png

2026-08-29 15:30:44 UTC

#Reticulum RNS 1.5.2 released. maintainance/bugfix release
"This maintenance release fixes a regression in resource transfers that was introduced in 1.5.1, an I2PInterface bug, and sets dataplane control parameters to default values that are actually sensible. If you installed 1.5.1, it's update time again, baby. "
a8d24177d946de4f1f0a0fe1af9a1338:/page/releases.mu`g=reticulum|r=reticulum

2026-08-29 15:11:03 UTC

A single signature by Trump can remove you from the internet
di [@smaurizi](https://mastodon.social/@smaurizi )
"A new David-versus-Goliath battle is being waged on the internet, after the U.S. government’s war against Julian Assange and #WikiLeaks. This time, the epicenter is Italy, and David is an Italian volunteer organization – a tiny collective of low-profile individuals who create and distribute encrypted email accounts, blogs, and mailing lists for free.
..."
https://www.ilfattoquotidiano.it/2026/08/29/autistici-inventati-trump-terrorists-news/8491452/
#autisti #censorship #fascism
https://static.piaille.fr/media_attachments/files/117/179/426/884/268/407/original/72af6069016aa196.png

2026-08-29 14:55:33 UTC

- reply

I think those with an expertise and a vision when it comes to technology, those "prescriptors" that others look up to with questions have a sort of a duty towards the rest of the world, to be cautious and consider a wide range of parameters before becoming co-responsible for the roll-out and social normalisation of tech. especially for tech that gets massively pushed in everything by such a limited number of actors with a proven track record of anti-competitive, anti-democratic and overall anti-social practices (microsoft, google, musk, bezos, etc.)
5/5

2026-08-29 14:55:04 UTC

- reply

i see perspective of de-skilling: reviewing code is not the same as writing it. coding, like any skill is like a muscle, if you don't use it, you lose it..
we may en up having generations that will have never spent years doing nothing but writing code, and an older one that may save itself by pivoting in low-tier reviewing, becoming assistants to machines instead of their past highly-skilled writing.
coding will practically be lost to machines. then real price may appear worth spending when there would be no other option?
4/5

2026-08-29 14:54:36 UTC

- reply

6/ on a macro-economic level, the speculative bubble that pumps 1000s of virtual billions in a business model where so far for 1$ spent 6$ are lost (we're talking 600% loss here!) is already affecting most of the rest of the economy:
- rise of energy price
- rise of the RAM price choking SMEs (pine64 announced they will stop manufacturing their linux products because of that... will they survive until the bursting of the bubble?)
- upcoming consequences of the chain-reaction of the likely collapse of the speculative bubble that once again will affect the most vulnerables, while some will massively cash in...
3/5

2026-08-29 14:54:01 UTC

- reply

4/ ... we see 10000s of new repositories coming up with code that maybe nobody ever read and that nobody will ever read...
5/ all these merge requests that are flooding and wearing down free/libre software maintainers everywhere (who deserve more love and support in general).
=> in the hands of ppl who would output better code themselves and validate all, with frugal models run locally, based on available training data, why not.
in practice in 99+% of the case it will be used either by ppl incapable of assessing the quality of the code, or of guaranteeing that it doesnt plagiarize free or prorietary code...
2/5

2026-08-29 14:52:37 UTC

Because it's now everywhere, here is my opinion/analysis about LLM-generated code (#vibecoding):
1/ I despise the evil US corps who centralize all knowledge in the world, for whom people work for free when using their services
2/ tragic ecological impact of these huge loads of computation, sold at a loss (when the _actual_ price of it will have to be paid, maybe ppl would think twice... except they'll be hooked by then and society may have changed)
3/ jobs are changing, that once were the creative job of a programmer become the replaceable one of correcting the output of the machine after some unqualified boss initiated something
4/ slop pollution
1/5

2026-08-29 11:23:46 UTC

- reply

well those i describe can be either homemade, or SDR, or using traditional radio equipment like radio ham, CB, FM etc.
apparently a 2.4GHz lora is coming also... there are some ISMs on 6 and 13MHz also... the only limit is the sky (and bouncing on the ionosphere, or the moon! ;)) )

2026-08-26 23:13:30 UTC

- reply

i understand your frustration, yet even if it is a bit of a toy/distraction meshtasticore did:
- generalize cheap ESP32+lora devices, incentivize Chinese manufacturers to innovate in that field;
- create a culture of building solar nodes, cheap enclosures, a whole bunch of 3D printable stuffs;
- make many people learn about lora and radio in general.
=> all these are already done, and readily available for the deployments of #Reticulum.
Hardware used for meshtasticore can potentially be converted to become part of a resilient post-internet comm. network!
(also meshtasticore may be still useful in some contexts.. devices may dual-boot..)

2026-08-26 10:39:19 UTC

- reply

2026-08-26 10:38:01 UTC

- reply

2026-08-26 10:31:06 UTC

- reply

Also i would add:
6/ on a macro-economic level, the speculative bubble that pumps 1000s of virtual billions in a business model where so far for 1$ spent 6$ are lost (we're talking 600% loss here!) is already affecting most of the rest of the economy:
- rise of energy price
- rise of the RAM price choking SMEs (pine64 announced they will stop manufacturing their linux products because of that... will they survive until the bursting of the bubble?)
- upcoming consequences of the chain-reaction of the likely collapse of the speculative bubble that once again will affect the most vulnerables, while some will massively cash in...
3/3

2026-08-25 20:23:58 UTC

- reply

2026-08-25 20:07:15 UTC

- reply

i am reassured! and i perceived much more of the former than the latter... so far... one more reason to "invest" the field and bring people who share non-shitty-libertarian values! i think it is really for everyone...

2026-08-25 07:43:17 UTC

- reply

the web3/crypto vibe i didnt feel so much (and i am extremely sensitive to it -and to ponzi schemes in general ;)) ), what made you think that?
For the vibe-coding unfortunately i am afraid it is neither more than less than in any other emerging field, ie. that it is within the average of the amount of vibe-coded stuff being pushed to repositories today (i would say about 9-to-1 to 19-to-1 against human-coded stuff but that's just a bogo-estimate ;) )
Fortunately reference implementation is 100% human-made, and you can see that the wiki tracks which stack is made or not with halucination-engine: https://reticulum.miraheze.org/wiki/Implementations

2026-08-23 09:15:47 UTC

- reply

at least 2 schools for #solar enclosures (a 3rd one= buy a box ready: "meshtastic solar enclosure" making sure the board is supported by #reticulum):
- for my balcony and/or if i have no time, i would get a 40Euros quasi-ready-made enclosure, in which i only have to add the board of my choice, because it's simple;
- if i want something very cheap, for instance if i want to build 10 and give them away or leave them in places where.. hmm they may get damaged or disappear, then i'd go for a DIY that could go down to 20 bucks (including the batteries) + the board
more about this here: https://piaille.fr/@jz/117048916856892063 (a thread)
(with photos ;) )
#note1wwx…z4za

2026-08-23 09:08:59 UTC

- reply

ie. lora is fun but by no mean it is consubstantial to RNS. #lora is a cool toy for nerds (that can quickly turn into an obsession...), and many may be used to meshtasticore that is lora-only.
here with #reticulum, lora is _one of the many ways to be connected_. one of the many radios even.
check modem73 as a cool free/libre modem software capable of using about _any_ #radio frequency.
lora doesn't cut it for you? maybe escalate to UHF, VHF, HF, FM, or some weird 6MHz ISM...
(NB: i write this regardless of what regulation applies where, yadda yadda...)