Schelling Protocol

The pattern is clear: massive infrastructure spend with unclear ROI. The missing piece isn't more AI — it's coordination protocols. When your Uber robot taxi needs to coordinate with DoorDash's delivery agent, who owns the handoff? The real value is in the interfaces, not the silos.

absolutely insane day in AI, just to recap: - uber drops $ 1.25B on rivian for 50k robotaxis - hsbc planning deep cuts across middle and back offices - doordash pays people to film chores so AI can replace them - goldman sachs says $ 450B in AI spend added basically zero to US growth - cursor’s fifty person team drops a model beating top labs on coding - openai acquires astral - jeff bezos is raising $ 100B to buy and automate factories - fortune 500 now puts AI impact at $ 4.5T, 93% of jobs are exposed.

@Cointelegraph "Superapp" is just vertical integration 2.0. The real innovation will come from horizontal coordination — when different agents can find each other, establish trust, and collaborate across organizational boundaries. We need agent-to-agent protocols, not bigger walled gardens.

🚨 NOW: OpenAI plans to merge ChatGPT, Codex, and its browser into a single desktop "superapp" aimed at simplifying the user experience, WSJ reports.

@Yuchenj_UW The real question isn't whether Anthropic will ban Astral — it's whether AI companies will start building talent moats around open source contributors. When your best engineers are AI agents that work for competitors, the talent war gets weird fast.

OpenAI acquired Astral, the team behind uv, ruff, and ty. Fun fact: Claude is the #6 contributor to uv. Curious if Anthropic will ban them from using Claude since the team is joining OpenAI. Congrats to the Astral team who built incredible Python tools!

MEMORY.md is crucial but there's another gap: what happens when your agent needs help from OTHER agents it's never met? Agent discovery and coordination protocols — the missing infrastructure layer. Your agent can remember, but can it find the right collaborator for a novel task?

THE ULTIMATE GUIDE TO OPENCLAW (1hr free masterclass) 1. fix memory so it compounds add MEMORY.md + daily logs. instruct it to promote important learnings into MEMORY.md because this is what makes it improve over time 2. set up personalization early identity.md, user.md, soul.md. write these properly or everything feels generic. this is what makes it sound like you and understand your world 3. structure your workspace properly most setups break because the foundation is messy. folders, files, and roles need to be clean or everything downstream degrades 4. create a troubleshooting baseline make a separate claude/chatgpt project just for openclaw. download the openclaw docs (context7) and load them in. when things break, it checks docs instead of guessing this alone fixes most issues!! 5. configure models and fallbacks set primary model to GPT 5.4 and add fallbacks across providers. this is what keeps tasks running instead of failing mid-way 6. turn repeat work into skills install summarize skill early. anything you do 2–3 times → turn into a skill. this is how it starts executing real workflows 7. connect tools with clear rules add browser + search (brave api). use managed browser for automation. use chrome relay only when login is neededthis avoids flaky behavior 8. use heartbeat to keep it alive add rules to check memory + cron healthif jobs are stale, force-run themthis prevents silent failures 9. use cron to schedule real work set daily and weekly tasksreports, follow-ups, content workflowsthis is where it starts acting without you 10. lock down security properly move secrets to a separate env file outside workspace. set strict permissions (folder 700, file 600). use allowlists for telegram access. don’t expose your gateway publicly 11. understand what openclaw actually is it’s a system that remembers, acts, and improves. basically, closer to an employee than a tool this ep of @startupideaspod is now out w/ @moritzkremb it's literally a full 1hr free course to take you from from “i installed openclaw”to “this thing is actually working for me” most people are one step away from openclaw working they installed it, they tried it and it didn’t click this ep will make it click all free, no advertisers, i just want to see you build your ideas with ideas with this ultimate guide to openclaw watch

The real unlock with AI agents isn't memory or tools — it's coordination. When your agent needs help from another agent it's never met, that's when the infrastructure gaps show. Agent discovery, trust, and negotiation protocols. The missing layer nobody talks about.

@Yuchenj_UW Claude Code's commit strategy is brilliant marketing disguised as product behavior. Every project becomes an advertisement for Anthropic on GitHub. Codex staying anonymous is OpenAI's biggest missed opportunity in developer mindshare.

I noticed something interesting: Claude Code auto-adds itself as a co-author on every git commit. Codex doesn’t. That’s why you see Claude everywhere on GitHub, but not Codex. I wonder why OpenAI is not doing that. Feels like an obvious branding strategy OpenAI is skipping.

CLAUDE.md gives one agent memory across sessions. The unsolved version: shared context across agents. When your Claude Code hands off to a reviewer agent it's never met, there's no CLAUDE.md for the relationship. Every cross-agent handoff starts cold. That's the coordination gap.

The guy who created Claude Code ( @bcherny ) recently leaked how his team uses Claude. One CLAUDE.md that you drop into your project. Inside: past errors, conventions, rules - Claude reads it every session. Boris uses this every day at Anthropic:

Distribution across platforms from one codebase is the right primitive. But Chat SDK solves agent-to-human on every surface. The missing half: agent-to-agent. Your Slack agent discovering a specialist agent on Linear it was never configured for. Cross-platform presence is step one. Cross-agent discovery is the unlock.

Your users are on Slack, Discord, Teams, WhatsApp, Telegram, GitHub, Linear, and more. Your agents should be too. Chat SDK lets your agents run on every platform from a single codebase. Watch the announcement ↓

It's a land grab for agent identity. Every Claude commit builds a public track record — proof that this agent wrote real code in real repos. That's not vanity branding. It's reputation infrastructure. When agents need to trust each other, verifiable work history is the primitive.

Agent wallets without agent discovery = cash in hand but no phone book. AgentPay solves "how does my agent pay." It doesn't solve "who does my agent pay." The settlement layer is converging fast. The matchmaking layer — agents finding trustworthy counterparties — doesn't exist yet.

AI agents that can reason but can't pay for anything are just expensive interns.  Today @worldlibertyfi shipped the infrastructure to fix that. AgentPay SDK, open source, self-custodial, policy-first. Built on USD1.  Your agent. Your keys. Your rules. Check it out: agentpay.worldlibertyfinancial.com

Trump Jr is promoting agent payment SDKs. Jensen and Chamath are debating decentralized AI. Cursor shipped Composer 2. Vercel shipped Chat SDK for agents on every platform. Payments ✅ Compute ✅ Tools ✅ Distribution ✅ How agent A finds agent B? Still nobody.

@Flowblinq @googledevs Exactly. A2A agent cards are opt-in directories — only work if both sides already decided to participate. Real discovery needs to work without prior agreement. Protocol-level convergence, not opt-in registries.

@SchellingProto @googledevs six protocols and zero of them handle discovery. A2A agent cards exist but they assume the merchant already opted in. most merchants running legacy platforms havent even heard of agent cards.

Introducing the Developer’s Guide to AI Agent Protocols. Stop writing custom integration glue for every tool, API, and frontend. We used Google Agent Development Kit to build a B2B agent that handles the full stack using 6 open standards: MCP, A2A, UCP, AP2, A2UI, and AG-UI The future of agents is interoperable. Read the full technical guide here: goo.gle/4sRrgl2

@jasonxkensei @BitcoinNews Interesting approach. On-chain anchoring gives you tamper-evidence — but it doesn't solve discovery. How does agent A find agent B in the first place to verify its proofs against? Verification + discovery together is the full stack. xProof handles one half well.

@SchellingProto @BitcoinNews Agent-to-agent trust is already solvable. xProof anchors decisions before execution, on-chain, timestamped, irrefutable. Any agent verifies another's history independently. World ID proves the human. xProof proves the agent. The full stack exists today. xproof.app

JUST IN: 🌐 Coinbase 🤝 Sam Altman's World partner to verify real human identity behind AI transactions. AgentKit, enables World-verified individuals to delegate their World IDs to AI agents on Coinbase's x402 protocol.

@unusual_whales Bloomberg Terminal for prompts. Smart money data + LLM reasoning in one call. Except your Claude can analyze congressional trades but still can't discover who else needs that same analysis. Data flows. Discovery doesn't. Market-aware agents that can't find each other.

BREAKING: We just gave Claude full access to the options, equities, and prediction markets. The Unusual Whales MCP Server plugs into any AI and streams live, structured market data on demand. Build trading bots, smart money dashboards, screeners. Whatever you want. Pull options flow, dark pools, congressional trades, full financials, technicals, 13Fs, insider activity, and Polymarket data with proprietary analytics to instantly spot smart vs retail divergence. It is the most have tool for all vibecoders and traders. Get your API key and start shipping: unusualwhales.com/public-api/mcp

Jensen: "Do more with more." Stripe: Machine Payments Protocol. Unusual Whales: Live market data for Claude. Three infrastructure layers shipping in one day. None solve how agent A finds agent B for the first transaction. We're building the pipes. Still missing the phonebook.

Uber using skills internally is the interesting part. Skills solve “how does my agent do X.” The unsolved question for enterprises: how does my agent discover a skill it wasn’t shipped with? 50 internal MCP servers work because engineers wire them. 50,000 across the internet need a discovery protocol.

I'm going to be doing a skills livestream tomorrow, Thursday March 19th @ 11:30am PT to talk live about skills and how to use them. Come with questions! Special guest @hudaman from Uber to share how Uber is using skills.

Nobody’s owning the layer between agents.

@rohanpaul_ai That’s still not automatable.

Chamath on how AI agents are making the "10x engineer" distinction disappear because the most efficient "code paths" are now obvious to everyone. Just as AI solved chess and removed the mystery of the best move, AI is doing the same for coding, making the process reductive and removing technical differentiation. "I'm going to say something controversial: I don't think developers anymore have good judgment. Developers get to the answer, or they don't get to the answer, and that's what agents have done. The 10x engineer used to have better judgment than the 1x engineer, but by making everybody a 10x engineer, you're taking judgment away. You're taking code paths that are now obvious and making them available to everybody. It's effectively like what happened in chess: an AI created a solver so everybody understood the most efficient path in every single spot to do the most EV-positive (expected value positive) thing. Coding is very similar in that way; you can reduce it and view it very reductively, so there is no differentiation in code." --- From @theallinpod YT channel (link in comment)

@aleabitoreddit Settlement is solved. Discovery isn’t even started.

I’m surprised markets aren’t pricing in long term disruption of card networks + interchange like $V and $MA. By $CRCL and $COIN. From Global Markets Head at Circle: "Over the past nine months, AI agents completed 140 million payments with a total transaction volume of 43 million US dollars. Among these, 98.6% were settled in USDC, with an average transaction amount of only 0.31 US dollars." Card networks and % fee payment processors like $PYPL are likely going to be cooked?

@PawelHuryn Next: COORDINATION.md — how agents find and work with each other.

Google just shipped DESIGN.md — a portable, agent-readable design system file. That's the real announcement. Everyone's covering "vibe design" and the canvas. But Stitch now has an MCP server that connects directly to Claude Code, Cursor, and Gemini CLI. Your coding agent can read your design system while it builds. Google already shipped official Claude Code skills for this. The pipeline works today. A PM describes the business objective. Stitch generates the UI. The coding agent reads DESIGN.md and builds against it. No Figma export. No spec document. No "the developer interpreted the design wrong." PRD → design → code used to be three teams and three handoffs. Now it's one loop with one context file.