Ryan

I choose you Dreg

Honored to be the first test subject for the @claudeai Intravenus subscription. It's a bit uncomfortable to type now but I'm assured we won't need keyboards where we're going.

@GroundwireBTC lfg

It's time. Groundwire Alpha is released on Thursday. Hear all about it in our spaces at 8pm ET. twitter.com/i/spaces/1qxoN…

What the jestermaxxing discourse gets wrong is that you're not meant to perform for some intended reaction, you're meant to express what you really think and ideally tickle your own sense of humor because she'd rather play with you than judge you.

-@tynan

Bets are a tax on bullshit

@WhiteMarbleDAO @hanfel_dovned Oh hell yeah

WMS update: still migrating and checking stability on the new bigger machine; some promising results with the high-parameter models. Then we have some fun ideas to roll out for the classical scholarship and LLM intersection. (Pictured: @hanfel_dovned′s %monsters app)

@post_polar_ Lunarpunk

OG cypherpunks, cypherpunk-pragmatist hybrid, hardline cypherpunk, milady cypherpunk, CROPS moderate, defipunk, pragmatist, inverted pragmatist, institutional pragmatist, ct-vc podcast industrialist, lunarpunk, post-cypherpunk, sapphire punk

@sigilante Interested

OOMFs take note: I'm starting a systems engineering reading group in two weeks. “SEBoK Salon” will meet 25 March onwards. DM here for a link or watch a more secure channel than clearweb X.

We're going to have to create new ways to need each other.

@sigilante @urbit we're turning that arrow around

⍼ new logo for Azimuth dropped @urbit ionathan.ch/2026/02/16/ang…

@Voltropy mogged

ANNOUNCEMENT: We just mogged malware. Introducing Mog, a programming language for self-modifying AI agents. Mog solves the security problems of claws + the usability problems of sandboxing. 1/N 🧵

@samanthawillman @sonofalli Such a good idea

@sonofalli @master_malwyl this so u coded.

haven’t seen my husband in 2 days

Ownership and Trust (and AI) Soon, every Tlon Messenger account will come with its own @openclaw AI agent—hosted on our infrastructure with its own cryptographic identity. No setup required. At @tloncorporation we care about ownership and trust. So where does AI fit in?

@DennisonBertram @urbit hire this man right now

Everyone's talking about AI coding agents. What nobody's saying: the hard part isn't building anymore. It's knowing *what* to build. I used Claude to rebuild Urbit in Rust in 2 days. 150+ iterations. The bottleneck moved. Most people haven't noticed.

@cerebralbosons @VitalikButerin urbit.xyz/?zone=mito.sys…

@VitalikButerin Urbit was born for this

In 2026, I plan to be fully back to decentralized social. If we want a better society, we need better mass communication tools. We need mass communication tools that surface the best information and arguments and help people find points of agreement. We need mass communication tools that serve the user's long-term interest, not maximize short-term engagement. There is no simple trick that solves these problems. But there is one important place to start: more competition. Decentralization is the way to enable that: a shared data layer, with anyone being able to build their own client on top. In fact, since the start of the year I've been back to decentralized social already. Every post I've made this year, or read this year, I made or read with firefly.social, a multi-client that covers reading and posting to X, Lens, Farcaster and Bluesky (though bluesky has a 300 char limit, so they don't get to see my beautiful long rants). But crypto social projects has often gone the wrong way. Too often, we in crypto think that if you insert a speculative coin into something, that counts as "innovating", and moves the world forward. Mixing money and social is not inherently wrong: Substack shows that it's possible to create an economy that supports very high-quality content. But Substack is about _subscribing to creators_, not _creating price bubbles around them_. Over the past decade, we have seen many many attempts at incentivizing creators by creating price bubbles around them, and all fail by (i) rewarding not content quality, but pre-existing social capital, and (ii) the tokens all going to zero after one or two years anyway. Too many people make galaxy-brained arguments that creating new markets and new assets is automatically good because it "elicits information", when the rest of their product development actions clearly betray that they're not actually interested in maximizing people's ability to benefit from that information. That is not Hayekian info-utopia, that is corposlop. Hence, decentralized social should be run by people who deeply believe in the "social" part, and are motivated first and foremost by solving the problems of social. The Aave team has done a great job stewarding Lens up to this point. I'm excited about what will happen to Lens over the next year, because I think the new team coming in are people who actually are interested in the "social": even back when the decentralized social space barely existed, they were trying to figure out how to do encrypted tweets. I plan to post more there this year. I encourage everyone to spend more time in Lens, Farcaster and the broader decentralized social world this year. We need to move beyond everyone constantly tweeting inside a single global info warzone, and into a reopened frontier, where new and better forms of interaction become possible.

@tanya1866 Bravo

i built rime for the zec hackathon, a privacy hardened light client. zcash light clients leak metadata through memo-fetch patterns, block-range queries, timing quirks, and network identifiers, which allows wallet fingerprinting and reconstruction of which shielded outputs belong to a user, even though the underlying transactions are private. rime fixes this on the client side: PIR, dummy queries, full-memo fetch, bucketed scans, constant-cost trial decryption, tor routing, and stateless/ephemeral sync modes make normal wallet actions far harder to fingerprint. code + details: github.com/tanctl/rime