Kreuzberg’s Post

𝗦𝘁𝗼𝗽 𝗯𝘂𝗶𝗹𝗱𝗶𝗻𝗴 𝗲𝘃𝗲𝗿𝘆 𝗯𝗮𝗰𝗸𝗲𝗻𝗱 𝘁𝗵𝗲 𝘀𝗮𝗺𝗲 𝘄𝗮𝘆. FastAPI isn't just "another Python framework." It's a deliberate choice — and knowing when to reach for it matters more than knowing how to use it. 𝗣𝗶𝗰𝗸 𝗙𝗮𝘀𝘁𝗔𝗣𝗜 𝘄𝗵𝗲𝗻:  • You're building ML/AI-powered APIs and your team already lives in Python  • You need async performance without the boilerplate of Go or Java  • Auto-generated docs (Swagger/OpenAPI) aren't a nice-to-have — they're a requirement  • You want type safety that actually catches bugs before production 𝗦𝘁𝗶𝗰𝗸 𝘄𝗶𝘁𝗵 𝘁𝗿𝗮𝗱𝗶𝘁𝗶𝗼𝗻𝗮𝗹 𝗯𝗮𝗰𝗸𝗲𝗻𝗱𝘀 (𝗦𝗽𝗿𝗶𝗻𝗴, 𝗗𝗷𝗮𝗻𝗴𝗼, 𝗘𝘅𝗽𝗿𝗲𝘀𝘀, .𝗡𝗘𝗧) 𝘄𝗵𝗲𝗻:  • Your org already has deep expertise and infra around them  • You need battle-tested ORM support and a massive plugin ecosystem  • You're building monoliths where convention-over-configuration saves months 𝗧𝗵𝗲 𝗿𝗲𝗮𝗹 𝗮𝗻𝘀𝘄𝗲𝗿? 𝗜𝘁'𝘀 𝗻𝗲𝘃𝗲𝗿 𝗮𝗯𝗼𝘂𝘁 𝘁𝗵𝗲 𝗳𝗿𝗮𝗺𝗲𝘄𝗼𝗿𝗸. 𝗜𝘁'𝘀 𝗮𝗯𝗼𝘂𝘁 𝘁𝗵𝗲 𝗽𝗿𝗼𝗯𝗹𝗲𝗺. FastAPI shines where speed-to-deploy, async I/O, and Python-native ML pipelines intersect. Forcing it into a legacy enterprise CRUD app is like using a scalpel to chop wood. Choose your tools like an engineer, not a fan. Thoughts? When did FastAPI click (or not click) for you? #FastAPI #Python #BackendDevelopment #SoftwareEngineering #WebDevelopment #APIDevelopment #TechCommunity #Programming #MLOps #SystemDesign

I searched my Claude Code session history and found something I didn't expect. Here's what I found: https://lnkd.in/dFcHWNGS 185 sessions in 13 days. 1,501 commands. 94+ commits shipped. Only 3 references to an IDE — and two were about window management. I'm an EM who codes daily. Java, Python, Kubernetes, Helm,  infrastructure-as-code across multiple repos. The kind of stack  that should demand a heavyweight IDE. Instead, I've been shipping from a terminal. And the work has  never moved faster. I wrote about why the IDE's core value proposition — "we'll help  you find and edit text in files" — feels increasingly like a card  catalog in the age of Google. The shift from file navigation to intent-driven development  already happened. Most of us just haven't checked our own  session history yet. #SoftwareEngineering #AI #DeveloperTools #EngineeringManagement

Chaofan Shou posted the discovery at 4:23 AM. Within hours, Python rewrites were already on GitHub That's the part I keep coming back to. Not the leak itself — the turnaround A developer took 512,000 lines of TypeScript and rewrote the core in Python before most of the world was awake. Mirrors spread faster than takedowns. When DMCA requests hit, the code had already moved to decentralized platforms. More forks kept appearing. A Rust rewrite is underway It wasn't clean. The same day saw a separate npm supply-chain attack hit the ecosystem. The community moved so fast it outpaced its own security instincts But here's what I actually noticed: nobody was waiting for permission. No one filed a request. No one asked Anthropic what was okay to build. They just... built. At 4 AM. In hours. In three languages We talk a lot about what AI is doing to developers. We talk less about what developers do the moment a tool they depend on becomes open Turns out the answer is: immediately, in parallel, with forks What does that say about the relationship between builders and the platforms they build on?

My project of the day: NaturalScript - write & maintain executable scripts in natural language I find it off-putting when people generate scripts with LLMs and commit the results to Git repositories. Those scripts are verbose and hard to maintain. It's like committing a binary file without the actual source code. NaturalScript lets you write and maintain executable scripts in natural language. It lets you treat shell/Python/... scripts as generated artifacts that you don't edit directly. Let me know what you think. https://lnkd.in/erEt6ZxH

Claude Code ships a command that audits your own permission habits. I ran it this weekend. The command is /fewer-permission-prompts. It scans your JSONL transcripts, ranks your most-used Bash and MCP tool calls, filters to read-only, and writes the narrowest patterns to .claude/settings.json. The non-obvious part is what it refuses to do. It drops everything Claude Code already auto-allows before writing anything. The official permissions docs list that set: ls, cat, head, tail, grep, find, wc, diff, stat, du, cd, and read-only forms of git. Allowlisting these is pure noise. It refuses to allowlist interpreters, shells, or package runners. python, node, bun, sh, npx, bunx, uvx. The docs warn that these tools execute their arguments as a command, which means a wildcard rule becomes arbitrary code execution. The real allowlist wins are MCP tools with read, get, list, search, or view in the name. These do not auto-allow and prompt every call. If you have not run this command yet, do it before your next session. #ClaudeAI #ClaudeCode #AITools #DeveloperProductivity ♻️ Repost this to help someone use Claude.

The Great 24-Hour Extraction: How Anthropic’s "Source Map" Slip Changed the Game ( yes its April fools day but this story is real🙂) Yesterday, a simple human error in a build script did what competitors have been trying to do for a year: It revealed the "Special Sauce" of Claude Code. The Leak: A 59.8 MB Javascript source map was bundled into version 2.1.88 of @anthropic-ai/claude-code. This wasn't just minified code; it was a roadmap to ~1,900 TypeScript files covering everything from "Self-Healing Memory" to "Agent Swarm" logic.  The "Clean Room" Counter-Move: What’s most impressive isn't the leak itself, but the speed of the reimplementation. • Sigrid Jin (@realsigridjin) used OpenAI’s Codex (via the oh-my-codex orchestrator) to perform a systematic rewrite.  • By porting the logic to Python, they’ve created a "legal buffer"—a clean-room implementation that replicates the behavior and architecture (the "claw-code" repo) without infringing on the specific TypeScript copyright.  • As of this morning, the project has already surpassed 50,000 stars on GitHub, making it the fastest-growing repo in history. The Engineering Takeaway: We are officially in the era of Instant Legacy. If your competitive advantage is "hidden code," you don't have a competitive advantage. The only thing that stays proprietary in 2026 is your compute and your live data. The logic? That belongs to the agents now. Anthropic tried to sell a "Security Review" tool, but their own packaging script was the ultimate security failure. The community didn't just look at the code—they ingested it.  Is this the end of "Closed Source" developer tools, or just a really expensive lesson in .npmignore? The "Claw-Code" Python Port: Repository and Discussion 👉 GitHub - instructkr/claw-code: The fastest repo in history to surpass 50K stars ⭐, reaching the milestone in just 2 hours after publication. https://lnkd.in/dSV3VjCC #claudecode

Write native. Preview human.  🚀 I just published my first VS Code extension - bin2prev Lets stop writing code in natural language, Lets just preview if needed.  Machines don't speak Java, Python, or JavaScript. They speak native binary.  High-level languages were invented for developers - not for machines. Every line of code you write gets compiled down to the binary that the CPU  actually executes. The native code is the real code. Everything else is a translation.  With AI agents, we can now write native binary directly - the way machines actually think. No compiler, no runtime, no abstraction layers.  bin2prev bridges the gap: write native with your AI agent, then preview it in your preferred language - Java, JavaScript, Python, Ruby, or Go —  so you can read what the machine already understands.  ✅ Opens binary files directly in VS Code ✅ Raw hex dump with ASCII view ✅ Equivalent source code in 5 languages ✅ Supports Mach-O & ELF  formats   ✅ ARM64 instruction decoding  Write native. Preview human.  🔗 Install it now — search "bin2prev" in VS Code Extensions or visit: https://lnkd.in/dytMK2fm Open for all to contribute: https://lnkd.in/dpsSkVen #VSCode #Extension #Binary #NativeCode #AI #OpenSource #Developer #Hackathon  💻 Source code: https://lnkd.in/dpsSkVen

whoa... the 64,000th feature claude code is the source code 😳 someones @ anthropic JUST ACCIDENTALLY exposed the full source code in a source map on npm. DMCA takedowns are going wild, but it's too late.. it's out there. 500k+ lines of code. people have already ported it to python, and scoured the code for some interesting insights about what is about to be released ✨ Kairos — Persistent assistant mode (even more OpenClaw like). Claude remembers across sessions via daily logs, then "dreams" at night — a forked subagent consolidates your memories while you sleep. ✨ Ultraplan — Sends complex planning to a remote Claude instance for up to 30 minutes. You approve the plan in your browser, then "teleport" it back to your terminal. ✨ Mult-agent Coordinator Mode — Already accessible via CLAUDE_CODE_COORDINATOR_MODE=1. Spawns parallel worker agents that report back via XML notifications. ... and something spicy claude code has "UNDERCOVER MODE" where they have taught it not to look like an AI, and basically pretend to be a human. used whe nit contributes (anonymously) to public repos. but tbh - claude code should be open source (like codex etc) so maybe this will force their hand? 🔗 HN thread: https://lnkd.in/eNS-pwHs

Ooops! What if a company vibe coded fake source code of it's product and leaked it as an April Fools joke? Unfortunately for Anthropic, it appears this wasn't that. https://lnkd.in/g5gMrNvY

I just stumbled onto something that might change how we think about Ruby performance. 💎 I was digging into my usual Rails workflows when I came across Spinel, a new experimental project by Matz. It’s an AOT (Ahead-of-Time) compiler, and the philosophy behind it really caught my eye. Instead of the usual "magic" we love in Ruby, Spinel compiles a strict subset of the language directly into native C. I took a look at the benchmarks, and the numbers are honestly hard to ignore: 🚀 Logic & Loops: 20ms vs 1,733ms (86x faster) 📊 Data Structures: 24ms vs 543ms (22x faster) 📦 JSON Parsing: 39ms vs 394ms (10x faster) What’s the catch? To get this kind of speed, you have to embrace a more disciplined, minimalist style of engineering. No eval, no send, no dynamic metaprogramming. It’s Ruby, but stripped down to its raw, high-performance core. Is it ready for our Rails apps? Not yet. Our favorite gems and Rails features still depend on that dynamic "magic." But seeing Spinel in action makes me realize that we don't always need every dynamic feature for every task. Sometimes, stripping away the noise is exactly what’s needed to unlock the next level of performance. For production today, I’m sticking with YJIT, but Spinel is a clear signal that Ruby is heading toward a very fast, very interesting future. https://lnkd.in/dxuxypP5 #Ruby #Rails #RubyOnRails #RoR #SoftwareEngineering #Spinel #Performance #MinimalistEngineering #BackendDevelopment #Ruby4