Curated signals for developers. Updated hourly.
In Parts 1-3, we built a storage engine: WAL, memtable, SSTables, and compaction. Every operation is a single GET or PUT. But real applications rarely need just one write at a time. Problem: Multiple Writes Need to Be Atomic Consider a bank transfer:
A few years ago, I spent about a month cleaning up duplicate charges. Every service involved already had idempotency "handled": the client sends an Idempotency-Key header, we store the response under it, and if the key comes back we replay the stored
A while back I built a small tool called SDKProof. it checks how well an AI coding agent writes an SDK's current API - the stuff that changed in the last major, that the model tends to get wrong because it learned the old version. Claude Opus 5 came
I recently tried to share an open-source project I've been working on called Open Vectorizer . It's a raster-to-SVG vectorization engine written in Rust. It runs locally, compiles to WebAssembly, has a reproducible benchmark suite, and competes surpr
I've always been amazed by how Oracle Database handles updates to a unique column-performing set-based operations that don't violate the unique constraint, yet when executed row by row, it temporarily permits duplicates. SQL > create table franck ( v
I have a side project: a four-in-a-row board game where the opponent is a neural network I trained myself. It has been on Google Play for a while, and last week I shipped a browser version of it. The interesting constraint was this: the browser versi
Every AI provider publishes a price in dollars per million tokens. Every comparison table ranks by it. Every build-versus-buy spreadsheet runs on it. That number is misleading, and not in a small way. It can be wrong by 10x, and wrong in the directio
JWT vs Session Tokens : The One Sentence That Actually Matters If you've built more than one web app, you've had this argument with a teammate: "just use JWTs" vs "sessions are simpler and safer." Both people are usually rightjust about different sit
A YouTube playlist can look healthy today and silently lose useful context later. A video becomes private, an upload is deleted, a viewer restriction changes, or YouTube hides an entry behind a generic unavailable placeholder. By the time someone not
Summary of 2026 Updates (v2 - Python ADK 2.0 Migration) This updated edition incorporates several key architectural improvements, production hardening, and dependency upgrades over the initial release: Python ADK 2.0 Core Upgrade : Upgraded to google
At some point in the life of almost every growing product, the team has a version of the same conversation. The codebase is causing pain. Features are taking too long. Bugs are multiplying. The developers are frustrated and the founder is losing conf
I spent a week trying to answer one question about my own AI agents: when one of them does something stupid in production, how do I prove the fix worked? For normal software the answer is boring. You have monitoring, an incident, a regression test, a
Modern applications don't just need features-they need observability . When an API becomes slow, a database query takes too long, or an unexpected error occurs, developers need answers quickly. That's where SigNoz and OpenTelemetry come in. For the A
How you doin'? Let's talk about the Iconic thing we heard a lot: "server down, try later." Your daddy said it while trying to book a Tatkal ticket. Your cousin said it the day JEE results dropped and the portal turned into a spinning wheel of despair
Most agent frameworks help you build a workflow. The harder part starts after that: the workflow has to run as a long-lived process, fail clearly, restart carefully, and be inspectable after the fact. That's the gap I'm building AgentOS for - an open
There's a delightfully literal answer to "Kubernetes with Docker": kind - Kubernetes IN Docker. Each node is a Docker container running a full Kubernetes node image. On an Ubuntu workstation it gives you a real, throwaway, multi-node cluster in about
Modern AI agents rarely complete a task in a single model invocation. Instead, they execute multi-step workflows: Retrieve documents Call APIs Query databases Generate intermediate plans Invoke external tools Produce a final response Most AI safety m
I've been building a small sync tool - it moves files between a remote project and a local folder, git-style ( clone , pull , push , status ). The moment you write a tool that touches other people's files on both sides of a network, you inherit a cat
Copy-paste an effect from anywhere and you get a finished snippet. It works, or half-works, and the moment you want to change one thing you're reading forty lines trying to find which one does it. TinkerFX is 79 web effects built the other way round:
The biggest update to the Model Context Protocol (MCP) since its launch is almost here. Lead maintainers froze the release The post MCP’s biggest update removes the machinery many servers were built around appeared first on The New Stack .
A deep-dive into the mechanisms that make software progressively harder to evolve. Introduction An application written ten years ago still works. Users rely on it every day. The servers run. The tests pass. On the surface, nothing looks wrong. And ye
I run pigfox.com as a one-person shop. It's two things now: 27 tools, 16 of them free , covering OSINT, email security, TLS and document forensics - and three live smart-contract demos on Base Sepolia, one of which lets a language model decide who ge
Table Of Contents What is a Git worktree? Why use worktrees instead of several clones? A practical directory convention Everyday Git worktree commands Consolidating several independent clones Safety rules before starting Phase 1: Inventory every clon