Activity for andyschwab.link

I can't see @mu.social without thinking of en.wikipedia.org/wiki/MU_puzzle

Interested to see where this goes... FWIW, I do think tools that help individuals really control how they interface with the internet (like Cloudflare's, but also like AT Proto) will be on the winning side of the AI transition

Robin Berjon@robin.berjon.com

Imagine things being so bad that your best solution is to cut yourself off from Google Search traffic. I don't think that people realise just how much of a bad actor Google has been, for year. No other company has damaged the web nearly as much. 🧵 www.adweek.com/media/publis...

Been thinking about plc.directory as a single-point-of-failure So as a learning exercise, I threw together DIDMARC - an example of an independent witness for the directory 🔍 This approach uses Starknet to be fully trustless, but a trusted mirror could perform similar checks to identify abuse

Just finished @laurenshof.online's powerful piece: "No Change" It is a critical read But I'd push back a bit on the history; I think the sovereignty of nations has, by its nature, always been in flux. And the desire for decentralized "sovereignty" without the state is part of the same story 1/n

FR#166 – No Change

Many people in Europe see the open social web as a path to digital sovereignty. This week demonstrated a problem: the assumption that European governments actually want to be sovereign.

June 12, 20269 min read

Connected Places

Read article

Endorsed a builder on at.fund

Testing out @mu.social 👀

So a surprising benefit of running LLMs locally is that it has given me a chance to introduce the tech to my kids on more-or-less my own terms

Was encouraged to see The European Social Stack announcement (https://european.social/). TLDR: this is good for more than just Europe.

Reflections on "Don't be evil"

European digital autonomy is a global good

June 10, 2026

T

Thoughts & Comments

Read article

My first computers were backroom cast-offs: a TRS-80, XT 8088 clone, etc. Their limits, and the workarounds they demanded to make them useful, were my intro to CompSci Having a bit of that experience again getting open models running on (woefully) underspec'd edge devices - and it's kinda fun 😀

Running Qwen locally on a system that is clearly not designed for it reminds me of telnet chat over 1200baud modems... ...watching... the... words... appear... It is, nonetheless, exciting that it works! But I also need to upgrade some things

I just finished @robin.berjon.com's excellent deep dive on the "search levy" problem - and how to fix it Strongly recommend digging into the arguments & recommendations here; the report is a remarkable distillation of our moment I have thoughts, but need to sort them out before posting further 😅

Valuable reflections on building trust.

David Gasquez@davidgasquez.com

Really enjoyed this talk from @bcantrill.bsky.social on why trust _is_ infrastructure. youtu.be/WF7J7qtZ8TA Recommended watching!

"Whatever word we use, I want it to preserve a clear division: humans on one side with responsibility, machines on the other as a boring tool."

Armin Ronacher@mitsuhiko.at

More musings after some people got upset about the word clanker. lucumr.pocoo.org/2026/5/26/cl...

I wish this weren't behind a paywall, but Ben Thompson's latest on Stratechery about AI Datacenters is brutal, and honest, and important: stratechery.com/2026/data-ce...

I... really appreciate this thread The PLC directory is a big, existential risk in a partitioned internet world (if, say, nations weaponized domain tld operator control) Addressing this risk has sticky trade-offs, but is important work long term

Edmund Edgar@goat.navy

I think we already know how to do decentralized identity, the issue is that is has trade-offs, the most relevant ones here being 1. need for an RPC endpoint to resolve 2. blockchain cooties I guess people won't accept those trade-offs as long as the centralized thing seems to be working for them

To measure things honestly means "propagating the uncertainty" Incidentally, that is exactly what Agency Graphs are meant to do: github.com/sodal-projec...

Andrew Nesbitt@andrewnez.mastodon.social.ap.brid.gy

The Mismeasure of Open Source: https://nesbitt.io/2026/05/09/the-mismeasure-of-open-source.html

I take a very long time composing stuff, and I can be, um, loquacious in my desire to be precise. In the past, I would double, triple, and quadruple check my content for typos and such. Now, I find I'm tempted to leave a few errors in, just so people know I didn't paste a prompt response 🙃