Trialling gadget a few years ago was one of the reasons I built https://t.co/9kIwwK1htf - I was working with a client on a high 6 figure @Shopify project and it **seemed** like gadget offered a lot of the things they needed but my goodness it was a nightmare. Sad to see any company shut down but I couldn’t figure out why it felt more difficult to use their platform than to build a backend and Shopify integration from scratch!
Introducing Claude Fable 5.2 and Claude Mythos 5.2, designed to tackle the most daunting technical challenges.
They’re capable collaborators on ambitious projects in many fields, and show improvements in areas like coding, reasoning, and vision.
My app got through Apple App Store release first time with no rejections and yours can to! Check out this repo for exactly how I did it. A codex or Claude can do all the hard work for you with these skills https://t.co/42leHhQhD4
After updating my multi sub agent work flows are no longer burning usage percentage like crazy 🤔 sounds like you guys fixed a bug rather than dropped a feature
Background subagents in Claude Code now use less of your usage limit.
A running subagent updates its status line every 30 seconds, and each update used to re-read its whole conversation. Now Haiku 5.5 writes it from short notes instead, so each update costs about 99% less.
Have you ever wanted to build an iOS app
But ran out of tokens before there was even anything on your screen? Or maybe you started with lovable and then realised it's web only? Point codex or Claude at this link and repo and let the skills talk you through the build out of your next app in hours rather than days and weeks! https://t.co/XkIO5PzIxM
I WANT you to steal this iOS app and make money from it!
https://t.co/cMXLO6in0e
I built this silly app with a full admin panel, iOS and web app to show how you can build with my platform Fóir - a full managed backend as a service. Everything is available for you, notifications, users, payments, data modelling and access - the whole shebang!
The repo has skills to take you through from scoping out your own app to getting it released on the App Store and you don't have to worry about any integration or backend development outside of whatever specific logic you need for your app in tiny worker functions or server routes.
Read about it here https://t.co/XkIO5PzIxM or just drop https://t.co/oxsHlYUxrt to your favourite coding agent and let it walk you through the process
@zebassembly@paper Call me crazy but charging for MCP requests as an extra just feels like a slimy way to make money in the short term while agentic workflows are new and hot
demesne is now written in Rust, and 1.1 is out.
It compiles Zanzibar-style ReBAC into Postgres row-level security, so your database enforces your authorisation. Generated checks in Rust, Go and TypeScript call the same predicates the policies use.
https://t.co/xmDf9KmtGd
Fixed on the way over from Go:
- a quote in a spec string could rewrite a policy
- a self or grant term under and/not compiled wider than it read
- a write the select policy hides changed nothing and reported success
- definers could come before the functions they call
New:
- one runtime, written in Rust: TypeScript runs it as WebAssembly in Node, Deno, Workers and browsers, and Go runs it on wazero with no cgo
- diff compares whole policies and functions with the live db
- check flags views that read past RLS
- specs can span several files
- migrations drop what a spec stops emitting
Was it worth it? The port surfaced two Go-era bugs that let a policy admit more callers than it read, both fixed. One runtime now serves Rust, TypeScript and Go: 537 cases replay identically in all three, and every push runs mutation tests.
Been thinking lots about how to close the feedback loop when building with agents. Back in the day we loved a feedback board with screenshots from clients. What if that flow could be driven by your users, your agent, their agent and also link directly to the data and backend running their app or website?