@jabster28 yeah but headscale is quite limited (os tailscale server ). And I dont like the thought of a central control node.
This setup would be static / my nix config can know all nodes at all times so perfekt for wireguard in my op.

Actually in my #openchat project i've integrated #libp2p and it can do something similar to tail schale without central control ( bit that ainr far enough / stable enough yet ). Also cool thing to check in that space is edgevpn also libp2p based vpn.

Also #OpenChat made some developments.
A major re-structure of the core apis - based on past iterations and tests - is planned.
Clean formalization of the #libp2p based federation api's is also planned ( de-central p2p vpn spec )

#openchat should seamlessly connect any '#ai' components, web-servers, or hardware.
Still hope #llm to become a commodity; Can def say I'm positively surprised my capabilities of #Opensource models today!

Checkout my Test on 'strix-halo': https://blog.t1m.me/blog/local-llms-on-strix-halo-128gb-shared-ram

Local LLMs on Strix Halo 128GB Shared Ram: My Tests

Strix Halo 128GB RAM, 100% local LLM agents, my tests

@Shwop and I are at the #39c3.

Find us at the @fiff_de stand or send me a message if you want to meet-up to chat about @littleworld or #OpenChat.

I think #kubernetes move towards the GatewayApi is a nice thing, full control over TCP traffic sounda like the hook for #OpenChat to be used as #federated cluster ingress.

https://www.kubernetes.dev/blog/2025/11/12/ingress-nginx-retirement/

Ingress NGINX Retirement: What You Need to Know

To prioritize the safety and security of the ecosystem, Kubernetes SIG Network and the Security Response Committee are announcing the upcoming retirement of Ingress NGINX. Best-effort maintenance will continue until March 2026. Afterward, there will be no further releases, no bugfixes, and no updates to resolve any security vulnerabilities that may be discovered. Existing deployments of Ingress NGINX will continue to function and installation artifacts will remain available. We recommend migrating to one of the many alternatives.

Kubernetes Contributors

Mehr zu meinen Projekten könnt ihr in meinem Blog lesen.
Viele Projekt bei denen ich mitwirke sind öffentlich verfügbar, z.B. #OpenChat – und ab Ende des Jahres auch die #LittleWorld Matching-Plattform.

Ihr hört von mir. Cheers!

P.S.: Endlich richtig im #Fediverse angekommen zu sein fühlt sich gut an, danke @annette und @Sascha für eure Unterstützung! 🙌

(3/3)

#deepseek ist so ziemlich der erste Chatbot, den ich uneingeschränkt nützlich finde. Ist natürlich vielleicht ein wenig ungerecht, es gibt ja eine gewisse Lernkurve wie man die Zufallspapageien vernünftig nutzt und das letzte mal, das ich mit #openchat und #chatgpt experimental gespielt habe, ist eine Weile her.
PS: Vielleicht macht auch das vorselektieren, bloß nichts was als politische Meinung verstanden werden kann, etwas aus.

Blender, not for everyone? | Open Chat with Dalai Felinto. Ep. #5

https://peertube.kaleidos.net/videos/watch/5486cdf5-6127-455d-aee9-9c4816b9f7c8

Blender, not for everyone? | Open Chat with Dalai Felinto. Ep. #5

PeerTube

Keeping teams aligned | Open Chat with Dalai Felinto. Ep. #5

https://peertube.kaleidos.net/videos/watch/f86d1835-0a5c-4553-9e33-673fa0752d76

Keeping teams aligned | Open Chat with Dalai Felinto. Ep. #5

PeerTube

Fast execution: the Penpot approach | Open Chat with Dalai Felinto. Ep. #5

https://peertube.kaleidos.net/videos/watch/d098df77-4101-49ff-8399-4a53db13c1b7

Fast execution: the Penpot approach | Open Chat with Dalai Felinto. Ep. #5

PeerTube

Contributions at Blender: devs vs designers | Open Chat with Dalai Felinto. Ep. #5

https://peertube.kaleidos.net/videos/watch/33e08052-0362-46ed-9912-8825f501a845

Contributions at Blender: devs vs designers | Open Chat with Dalai Felinto. Ep. #5

PeerTube