Philipp Pohl-Krüger

Philipp Pohl-Krüger

@matheus23.com

Building iroh with the amazing folks at number 0 (n0.computer). Generally striving to increase user agency and excited about commons networks. Only works for Canadian CEOs, apparently. Rust, cryptography, CRDTs & more on my feed

665 followers373 following683 posts

I can't reply but... sendme? Lots of good packaged versions that are even wire compatible with the CLI nowadays.

iroh.computerirohless net work for networks
Rebane@rebane2001.bsky.social

okay heres the requirements to anyone replying: must not pass through a 3rd party must not require double-transfer must work cross-device must work outside of lan/proximity must work with other people's devices/setups must be fast for large files must not be insecure must be not annoying to use

The future is sooo weird. I'm in an airplane. I marvel at the fact that I can have 1h (wqow) of free wifi from here. I'm interested in how the wifi behaves when using QUIC (how it affects congestion control etc.) But I only have my phone. So I ask an agent to build me a test app.

Within this hour, I've gone and created this test app, ran some experiments from my phone (just connections to iroh's relay QAD endpoints) and captured the trace logs for that. And I have time to spare to talk to you about this. Wat.

(6) is not about QUIC, it's about HTTP3 header compression (QPACK). Just an FYI for anyone who's doing QUIC via iroh.

so what has broken the internet today? well, 1. php 2. node.js 3. gitlab 4. 4 5. musl 6. QUIC/HTTP/3 7. Signal 8. Apache Traffic Server 9. Pillow for Python, even! 10. bind 11. LiteLLM 12. XZ Utils 13. Bind (Java is reported, but supposedly it's just their version of the 4 issue)

Philipp Pohl-Krüger reposted

Not gonna RS the thread that made me think about this (you've likely already seen it, or a similar one) BUT Every person I know who has cut off or gone low contact with a parent took YEARS of agonizing and trying to make things work because damn that's their PARENTS.

Err wtf this is big?

Astra ⎔@astrra.space

@personhood.removal.surgery pointed out that they have just casually invented a good way to do suffix cache reuse and uh holy shit i just tried implementing this for my deepseek thingy and it just... works???

There are some interesting ideas for future jobs by modern monetary theory proponents: The Job Guarantee. The idea is to have the government fund community or care work at livable wages.

aly@aly.codes

This is getting a little fatigued and garbled, but my point here is that there is essentially no vision for what new jobs could look like here, and all our visions for new jobs are really just old jobs that don’t exist anymore. So I do think there’s a real possibility of mass unemployment!

Excited to talk about iroh and networking :) We'll try to record this! ✌️

Niklas Korz@niklaskorz.eu

Tried posting on atmo.rsvp, but that didn't work, so it goes here: Our next #NixOS and #Rust meetup takes place on Tuesday, October 6, at the DHBW Mannheim! As main talk, we'll have @matheus23.com from @iroh.computer, the QUIC-based peer to peer network. rheinneckar.events/events/67013...

I wonder if chain of thought was so successful because it makes token generation use more computing resources than prompt processing. Reading a source file and producing a source file are otherwise forced to use the same amount of compute per token.

Just looking up some literature. This is clearly discussed. Mostly as a way of allowing a model to spend more compute on harder problems, not on the read vs. write axis, which I guess is a more correct framing.

Philipp Pohl-Krüger reposted

it is sooooo cool that we have an actual explainable algorithm here on bsky

spacecowboy@spacecowboy17.bsky.social

PSA: you can find out which of your likes caused For You to show any given post at foryou.club/explain Just enter the post URL and hit "Explain":

Philipp Pohl-Krüger reposted

Official announcement will follow next week: we're meeting on Tuesday, October 7, for our next @NixRust meetup in #Mannheim. We still have time for a second talk, so if you have a topic or project related to #RustLang or #NixOS that you'd like to present, get in touch!

Gooooood thread feat. cryptography terminology, Mark Miller, Granovetter Diagrams and a little bit of iroh.

zooko2.bsky.social@zooko2.bsky.social

The encrypting/verifying key should never have been called a “public” key. It should not, in general, be public/broadcasted/posted/discoverable. Instead treat it as sensitive and ensure that it is known only to the agents that need it, and no-one else. ⤵️

Debuggers are overrated :) (The tool, not the informal job description lol)

bushr.at@bushr.at

I have been programming in Rust for half a decade, embedded, web, desktop, even contributed to the standard library! Not once have I ever successfully used a debugger to fix a logic bug in a Rust program.

Philipp Pohl-Krüger reposted

Dear American tech leaders: There is no way to pause or slow down offensive cyber research without hobbling (American) cyber defenders. But there *are* ways to accelerate cyber-defense without accelerating cyber-offense! ⤵️

Philipp Pohl-Krüger reposted

Excited to have a ✨ guest post ✨ from @djwyen.bsky.social in our @inkandswitch.com Keyhive lab notes! The local-first setting is new to a lot of cryptographers; here Derek shares some of his insights working on the BeeKEM paper, cross-fork security (CFS), and fundamental protocol design tradeoffs 🐝

inkandswitch.com06 · E2EE in the Local-First SettingExamining design considerations for developing cryptography for the local-first setting.

Philipp Pohl-Krüger reposted

Philipp Pohl-Krüger reposted

Replying to @kandake.africa

The three directors of Northsky Social are all women, @jay.bsky.team was CEO of Bluesky until she decided to move back to the work she loves. Your PR bits are stepping on toes.

Philipp Pohl-Krüger reposted

Been playing around with @iroh.computer a little bit with the rust docs (still a rust noob). Very impressed and like it a lot! I’ve abandoned so many project ideas over the years because accessing a machine on a home network inbound via IP was just not feasible. Wish I could remember them all.

Philipp Pohl-Krüger reposted

Routers are lame. We don't need routers. Just connect your devices with ethernet and use IPv6 link local addresses. Or Wifi Direct. Or Wifi Aware. All IPv6 link local addressing. (Every now and then I dream about IPv6 link local addr support in iroh)

A 118B model scoring 40% on DeepSWE (compared to the 700+B GLM 5.2 that scores 44%) is super impressive. Will want to download and check out.

Benjamin Muskalla@bmuskalla.dev

Today we're releasing Laguna S 2.1, our most capable model to date. It's a 118B total parameter MoE model with 8B activated per token, a context window of up to 1M tokens, and thinking and no-thinking modes. Small enough to run on a single @NVIDIAAI DGX Spark poolside.ai/blog/introdu...

Philipp Pohl-Krüger reposted

Had a fantastic time at @localfirstconf.com. Started off relatively stark with discussions acknowledging the general political instability of our times and how much data sovereignty matters. Ended with an absolutely wild demo of @inkandswitch.com's patchwork project. Lots of good bits in between.

Philipp Pohl-Krüger reposted

Imagine cities powered by renewable energy, buildings overflowing with greenery, and communities in harmony with nature. Solarpunk blends sci-fi, art, and activism to show hopeful, sustainable futures with DIY innovation, solar power, and social inclusiveness. Learn more ➡️ w.wiki/9BBv

Yeah.. after having worked on a bunch of bug reports with LLM assistance, this is the next obvious step beyond the chat+harness interface.

dax@thdxr.com

we have an issues channel where every good github issue goes and we can ask opencode to reproduce it and fix even started throwing posts from here in there