Charlie You

“Things that were recorded yesterday are being capitalized today, and they’ll be composable tomorrow.” @CharlieYouAI of @RWA_xyz talks the three phases of tokenization and how RWAs evolve into fully onchain financial markets From our RWA Research Day

I'm building @OnchainEstate (Tomorrow Labs) with James Tse to save your digital assets from death and taxes. A wallet for all your digital assets with built in succession.

@TuongvyLe12 never thought about qc this way but makes perfect sense, awesome!

The way we safeguard digital assets is about to change. For decades, financial regulation has assumed that protecting investors requires intermediaries: custodians, balance sheets, and institutional gatekeepers. But crypto introduced something fundamentally different. Recent events have also made something clear: not all “vaults” meet the standard investors should expect. Using smart contracts alone is not enough. Security and investor protection depend on how these systems are designed, governed, and constrained. Today, we submitted a letter to the SEC and CFTC proposing a new path: Non-custodial smart contract vaults can satisfy the SEC’s qualified custody and CFTC’s customer property segregation requirements — under defined guardrails. At a high level, custody and segregation rules have always been solving for the same risks: misappropriation, commingling, and exposure to intermediary insolvency. What’s changed is that we now have infrastructure that can address those risks directly in code, rather than through reliance on intermediaries. Properly designed vaults mean: ✔️ Client assets are never exposed to intermediary insolvency ✔️ Withdrawal rights can’t be overridden, even by insiders ✔️ Misappropriation is structurally eliminated ✔️ Asset ownership is continuously verifiable in real time Offering vaults as an option matters because many digital assets cannot be supported by qualified custodians today, and advisers face real tension between compliance frameworks built for traditional markets and the realities of on-chain assets. That’s why our letter proposes 7 guardrails for a vault to qualify: 1⃣ No unilateral authority to withdraw client assets 2⃣ Programmatic enforcement of client redemption, withdrawal, and transfer rights 3⃣ Cryptographic segregation of client assets 4⃣ Governance and upgrade mechanisms that are transparent, time-locked, and constrained 5⃣ Robust security and operational controls 6⃣ Independent audits and real-time verification 7⃣ No economic interests in underlying protocols These guardrails distinguish between vaults as infrastructure for investor protection and vaults as unstructured risk. If adopted, this would be the first regulatory framework where investor protection is achieved through non-custodial, programmable systems rather than institutional intermediation. Where safeguarding is embedded in the infrastructure itself. We believe this is precisely the type of alternative compliance framework envisioned by the recent SEC–CFTC MOU, and we encourage both agencies to engage through the Joint Harmonization Initiative to develop a coordinated, vault-based custody standard.

@jonaswillett1 @AngelList @mattvine013 @maggiexgao @SpencerChandlee POKER

I'm looking for the best poker players in NYC. Hosting an exclusive poker night on March 19th in NYC with @AngelList and friends. Only admitting 20 guests. If you're a founder or investor — this is for you. Cohosting with @mattvine013, @maggiexgao, @SpencerChandlee, @laurentspan, @sampeiomichi, @nathanleexyz. Food, drink, and excellent company guaranteed. Major shoutout to @AngelList, @joinellis, Brown Rudnick, RRE, and @PierWorld. Tag the best poker player. -- Want to join? Comment "POKER" and I'll send u the link.

@bmgentile @hbarTaTa @RWA_xyz @circle Will get this fixed! Thanks for flagging 🫡

@hbarTaTa @RWA_xyz very cool! but does look like they’re including $usdc held in @circle-owned “pre-mint” operational addresses (unfortunately) — those are non-circulating and purely used operationally Pre-mint acct #1: hashscan.io/mainnet/accoun… Pre-mint acct #2: hashscan.io/mainnet/accoun…

Hedera is live on @RWA_xyz 📈

BEAM is the correct virtual machine for agents, and Elixir and Gleam are the correct languages. The future is massively concurrent

Most important question in agentic engineering right now: "how can I spend more tokens to increase quality?"

@qthomp Openrouter founded by one of the opensea cofounders, who knows this all too well

AI investors should study crypto decentralized exchange aggregators and what happens to value capture and profit margins.

you want your multi-agent system to operate like a quake ctf clan not a middle management map reduce

@beyang Y'all COOKED with this one, consistently the best reviewer. Feature request: json, or some other output that is more machine readable so it can be used in CI, etc.

With agentic coding, we're increasingly reviewing code outside of traditional PRs. So we decoupled the Amp review agent from any UI. Invoke it from the CLI and from within threads (as a skill). You can also define "checks", codebase-specific review guidelines stored in git.

@nateliason Checkout pytgcalls, which you can use to set this up, but with telegram voice call as the interface instead of a custom webapp! github.com/pytgcalls/pytg…

@jefftangx if you setup cli proxy api you can use your claude / chatgpt max plans for all of them

I want to use Devin, Droid, and Amp so much more I've played with their free/basic features From what I can tell they are all much better than CC But a lot more expensive

@steipete this is very much a hack but i orchestrate this in claude code: have a stop hook that calls codex for a review, stops if no issues, otherwise fixes the issues with subagents (which can also be codex if you want) github.com/charlieyou/cer…

need a ralph-loop so codex keeps running /review until it's done. Under-rated feature. Haven't been using that enough. Slow but faster than having to read issues and fixing that a week later.

The vibe coders are speed-running their "discovery" of software eng best practices (tdd, risk-first pm, tracers)

@doodlestein Love it! Is there a preferred path to migrate issues to br from bd?

I mentioned recently that I've been... busy. Lots of projects in the oven, in various stages of completeness. Well, I'm now pleased to introduce one I've worked very hard on, because it's so near-and-dear to my heart: beads_rust, or br for short. You can get it here: github.com/Dicklesworthst… It's a fast, minimal Rust port of @Steve_Yegge's amazing Beads project that I've built so many of my workflows around. Discovering Beads and seeing how well it worked together with my MCP Agent Mail was a truly transformative moment in my agent coding workflows and professional life more generally. This quickly also led to my beads_viewer (bv) project, which added another layer of analysis to beads that gives swarms of agents the insight into what beads they should work on next to de-bottleneck the development process and increase velocity. It's beads (and mail) all the way down. I'm very grateful for finding beads when I did and to Steve for making it. But at this point, my Agent Flywheel System is built around beads operating in a specific way. As Steve continues evolving beads toward GasTown and beyond, our use cases have naturally diverged. The hybrid SQLite + JSONL-git architecture that I built my tooling around (and independently mirrored in MCP Agent Mail) is being replaced with approaches better suited to Steve's vision. Rather than ask Steve to maintain a legacy mode for my niche use case, I created this Rust port that freezes the "classic beads" architecture I depend on. The command is br to distinguish it from the original bd. This isn't a criticism of beads; Steve's taking it in exciting directions. It's simply that my tooling needs a stable snapshot of the architecture I built around, and maintaining my own fork is the right solution for that. Steve has given his full endorsement of this project.

@lewis_b_metcalf amazing!! now do scrolling in vi copy mode :D

For all the Amp CLI + Ghostty + tmux users out there, we have a fix for the shift+enter issue! 1. run `amp update` 2. enable extended-keys by adding `set -s extended-keys on` to your tmux.conf 3. restart tmux with `tmux kill-server` 4. enjoy using new lines again :)

Incredibly sick work! So seems like the planners are constantly comparing current state of the codebase to the spec/existing software features. How would this work for when there isn’t something to reference for what to do next? Or is that not the use case for massively parallelizing like this?

We built a custom harness from scratch that ran hundreds of self-coordinating agents for weeks without human interaction or drifting. It was effective at linearly scaling coding agents, allowing them to tackle complex engineering tasks through long time horizons.

Time for bed I guess...

I had a very stable sleep schedule for 3 years and Claude Code has completely destroyed it