Your agents are fast. none of them are on the same plan. the Agentic TPM from @jakejooyoung beats pinned messages and stale plan docs at keeping every agent current swap hoping the agents heard about the change for a live record they all read from how it works: 1/ the plan changes and people make the call 2/ Serro spots the drift and updates everyone at once 3/ that's it.
Claude Code can run Sonnet as your main model and bring in Opus only for the hard calls: one flag, --advisor opus. Set it up in 4 steps: 1. Launch with: claude --model sonnet --advisor opus 2. Claude decides when to ask Opus: before it commits to a plan, when an error keeps coming back, and before it says it's done. 3. Put the grunt work on Haiku: a subagent file in .claude/agents with model: haiku for searching and reading big files. 4. Give any subagent that tests risky changes isolation: worktree, so it works on its own copy of the repo. /advisor turns it on or off mid-session. It needs the Anthropic API; it isn't on Bedrock or Google Cloud yet. https://t.co/RDx6YSGDg8
I edited a SwiftUI app on a Windows PC and saw the change on my iPhone 9.3 seconds later. No Mac anywhere. The stack, all open source except one piece: WSL2 Ubuntu 24.04 Swift 6.4.1 from https://t.co/APlidSAOcr xtool (builds, signs and installs the app) omarchy-apple-dev (Linux stand-ins for Apple's asset tools) usbipd-win (hands the phone to Linux) What it cost me: three different login failures, then a one-byte bug in the USB relay, over two days. What it still can't do: no Simulator. So I put the real phone in a pane inside the editor instead. Built on Windows: 14 seconds the first time, 4 seconds now.
Railway's advice to every customer after yesterday's incident: roll every environment variable stored in Railway. API keys, database credentials, registry credentials. Immediately. An unauthorized party accessed encrypted variables and, for some, the decrypted values. Then it paused deployments and invalidated every Railway API token. If you run six services with forty variables, "roll everything" is brutal. Which ones first? Which are the same secret in three places? Which are built from another variable and must not be touched? Which one is a database password that changing a variable does not change? These questions is the list I built apikeys-handler to keep. It imports your variables, groups them by service, category and priority, generates new values only for secrets your app owns, walks you through the ones a provider issues, restarts each service once, and remembers where you stopped. It never prints a value, never stores one, and the model never sees one. Try it on a throwaway project first with this: npx --yes --allow-git=all github:codegobrrrr9/apikeys-handler demo
A bot from the Grok Bot marketplace comes with someone else's schedule. Nightly Audit Engineer, one of the Grok Bot team's own 54, runs at 4am by default. at 4am you're asleep. so the first time it needs your yes, the approval waits about 10 minutes, shows Expired, and the work never happens. you wake up to a bot that "ran" and did nothing. so before its first routine runs, set the split once: allowed in advance: reading, drafts, reports, draft pull requests. asks first: sending, posting, spending, signing, merging. then read the history after its first run. an Expired approval means it needed something you left on Ask first. Allow it, or make the routine draft instead of do.
Apple said you need a Mac to publish apps on the App Store. I said f*ck it and created my own version of Xcode. I call it WinCode. (genius, right) You can literally prompt to build an IOS app with Claude and watch every line of code and testing done here, on Windows. Code signing, compilation and launch will all happen on your Windows, still fixing a few things for the final part. The live preview is from my connected iphone for testing. I would like to know if this is genuinely useful to Windows app developers or not? My next goal is building an app and publishing on the App store right from my Windows PC!
Visa moved stablecoin prefunding into production this week! Visa now offers stablecoin payments across 180+ countries. Inyo is the first partner. The 180 countries is Inyo's payout reach, not a new Visa product in 180 countries. This is definently bullish news in my opinion. https://t.co/naOd4YQKTn
The /ask command. save it as a skill and your notes start answering back: "Answer the question I write after the command using ONLY the notes in /workspace/brain. 1. Check _system/maps first, then search the whole vault. 2. Read the relevant notes fully, not just titles. 3. Answer first, in 3 to 6 plain sentences. 4. Then the evidence: each note as a [[wikilink]] with the line that supports the answer. 5. Then where my notes disagree with each other, if they do. 6. Then what my notes do NOT cover that I would need to be confident. If my notes do not answer it, say "Your notes don't cover this" and stop. Do not use memory or the web unless I ask." Step 6 turns every answer into a reading list.
Opus 5.5 is a great planner and an expensive typist! Run a whole feature in one session and that's what you pay for: $20 per million output tokens for code Sonnet 5.5 writes at $10. The bill isn't the worst part. One context doing five jobs carries all five on every turn. by hour two it has lost the interface it agreed to and rewritten a function the tests needed. So split the jobs. Opus leads at $4 / $20 and plans, briefs, merges, decides. Sonnet workers do the typing. A verifier checks every claim of done. The catch: a team only pays off on some tasks. the first prompt in the setup exists to tell you when not to build one.
