Hacked Copilot into CI. 10 min = device OAuth → secret → curl → GitHub Action. Zero new services. It WILL break (undocumented), but it's the quickest way to prototype PR summarizers / changelog bots. Guide + code: kamil.chm.ski/github-copil... #GitHubActions #AIEngineering
💡 What if you could get React's useReducer mental model but entirely server-side?
That's SSR+ (Server-Side Reducers) - HTML partials with embedded state that self-assemble in the browser.
No frameworks. No hydration. Just fast, reliable UIs.
cimatic.io/blog/html-pa...
13 failed releases reminded me of a timeless lesson: stop predicting, start observing.
I let AI pile on “smart” fixes until my release script collapsed. Let It Crash - shipped on the next try. 🤖✅
My story➡️ kamil.chm.ski/let-it-crash...
#DevOps #RustLang #CICD #LetItCrash
July has been a truly incredible month for model releases from China - Moonshot (Kimi K2), Z ai (GLM-4.5) and 5 new releases from Qwen
I think it's undeniable that the best available open weight models now come from the Chinese AI labs simonwillison.net/2025/Jul/30/...
My favorite commits start with a minus sign.
Lighter code → clearer roadmap → faster value.
Junior or senior, AI or not: question every line, delete what doesn’t earn its keep.
Protect the product—not the syntax.
GitHub released Spark yesterday, their extremely well crafted prompt-to-app platform for creating and iterating on React apps with user auth and persistent storage
I like it a lot! I reverse engineered it with Spark itself, the details are fascinating simonwillison.net/2025/Jul/24/...
Building @cimatic_io MVP led me down some interesting technical rabbit holes 🛠️
One small win: Rust ↔ TypeScript type drift with two open-source crates that generate Zod schemas + TS types from Rust structs.
🧵 Full post: kamil.chm.ski/bridging-rus...
#RustLang #TypeScript #OpenSource #IndieHacker
I scraped the schedule for Open Sauce 2025 this morning and built an alternative schedule interface with the option to add everything to your calendar (via ICS)... working entirely on my iPhone, using OpenAI Codex and Claude Artifacts
Prompts + transcripts: simonwillison.net/2025/Jul/17/...
Vibe coding thousands of lines is easy, but how do you know it's what users *actually* want? My new article shares how I ensure AI-generated code aligns with user needs. Read it: kamil.chm.ski/vibe-coding-... #AICoding #UserExperience #ProductDevelopment #IndieHacker
Still in development, but here's where I'm headed: cimatic.io - I'm curious what jumps out to you as the biggest opportunity for improvement.
📚 **Master GitHub Actions Optimization**
Want the complete guide to workflow anatomy and CI Vitals optimization?
Read the full breakdown: cimatic.io/blog/github-...
#GitHubActions #CICD #DevOps #DeveloperProductivity #Automation
🚀 **5 Quick GitHub Actions Optimizations**
1. Move linting/quick tests to job start
2. Add caching for node_modules, pip, etc.
3. Use artifacts to share builds between jobs
4. Remove unnecessary dependencies
5. Add retries for external calls
Small changes = big savings.
🐌 **Common GitHub Actions Performance Killers**
❌ Sequential jobs that could run parallel
❌ No caching (downloading deps every run)
❌ Expensive operations before quick checks
❌ Flaky external dependencies without retries
❌ Rebuilding same artifacts in multiple jobs
🎯 **Optimize with CI Vitals**
**🏎️ Workflow Execution Time**
- Remove unnecessary `needs` dependencies
- Parallelize independent jobs
- Cache dependencies and build outputs
**🎯 Noise-to-Fix Ratio**
- Retry logic for network steps
- `continue-on-error` for non-critical steps
⚡ **How GitHub Actions Actually Runs**
1. Event triggers workflow
2. Jobs start in parallel (unless `needs`)
3. Steps run sequentially
4. Artifacts/outputs pass data between jobs
5. Workflow completes when all jobs finish
Parallel = faster. Dependencies = bottlenecks.
🏗️ **GitHub Actions Building Blocks**
**Workflow** → Collection of jobs
**Job** → Set of steps (runs on one runner)
**Step** → Individual task or action
**Action** → Reusable code component
**Runner** → VM that executes everything
Understanding this hierarchy = better optimization decisions.
🧵 You write GitHub Actions workflows—but have you ever wondered about the detailed execution flow behind the scenes?
Here's a deep dive into GitHub Actions anatomy, from triggers to completion:
We ditched CGI in the late 1990s because of the overhead of starting, executing and stopping a process for every incoming request... turns out modern servers (plus languages like Go or Rust with a fast startup time) mean CGI isn't such a bad idea any more! simonwillison.net/2025/Jul/5/c...
This thread covered 3 of the 7 Pipeline Pillars. The complete framework includes more:
🔴 Advanced Testing Strategies
🟡 Pipeline Control & Orchestration
🔵 Multi-Platform & Multi-Cloud Support
🟤 Access Control & Security Architecture
📖 Read about all 7 pillars: cimatic.io/blog/cicd-pi...
9/9
Pillar 3: Metrics & Observability 🟢
Your pipeline generates wealth of data.
Track what matters:
• DORA metrics (lead time, deployment frequency)
• Developer experience metrics
• Pipeline performance trends
• Build/test bottlenecks
You can't improve what you don't measure.
8/9
Pillar 2: Feature Flags 🟠
The game-changer: Decouple deployment from release.
• Deploy safely with features off
• Gradual rollouts to minimize blast radius
• Instant rollback without code deployment
• A/B testing with real user data
Recovery time: Hours → Seconds
7/9
Pillar 1: Multiple Environments 🟣
Beyond dev-staging-prod:
• Team-specific sandboxes
• Ephemeral preview environments
• PR-based testing environments
• Compliance-ready promotion gates
No more "who broke the shared dev environment?"
6/9
Key insight: Unlike the sequential Golden Path, pillars can be implemented in ANY order.
Need better testing? Start with Pillar 4.
Security concerns? Jump to Pillar 7.
Scaling issues? Pillar 1 or 6.
The framework adapts to YOUR reality.
5/9
The 7 Pipeline Pillars (choose your adventure):
🟣 Multiple Environments & Promotion
🟠 Feature Flags & Gradual Rollouts
🟢 Metrics & Observability
🔴 Advanced Testing Strategies
🟡 Pipeline Control & Orchestration
🔵 Multi-Platform & Multi-Cloud
🟤 Access Control & Security
4/9
The Golden Path (your foundation):
1️⃣ Code Commit
2️⃣ Automated Build
3️⃣ Automated Testing
4️⃣ Staging Deployment
5️⃣ Production Deployment
6️⃣ Monitoring & Feedback
Master this first. Everything else builds on top.
3/9
Introducing the CI/CD Pipeline Architecture Framework:
🏗️ Golden Path: 6-step core workflow foundation
🏛️ Pipeline Pillars: 7 foundational capabilities
Think foundation + pillars that elevate your platform from basic deployment to enterprise-grade ecosystem.
2/9
🧵 After 20+ years building CI/CD pipelines, I've developed a systematic approach to pipeline architecture.
I've seen teams reinvent the wheel repeatedly, building similar solutions without a shared framework.
Here's the CI/CD Pipeline Architecture Framework I use: 👇
1/9
A year ago, I was deep in platform engineering interviews before I went all-in on indie hacking. I originally wrote down my CI/CD architecture knowledge for interview prep. Two weeks of polishing and turned into a comprehensive blog post. Hope it's a valuable resource for fellow engineers!
Just wrapped the biggest content piece for @cimatic.io yet 🚀
CI/CD Pipeline Architecture: Complete Guide
• 4,500 words
• Golden Path + Pipeline Pillars framework
• 20+ years of engineering insights
Full launch Monday morning.
#buildinpublic #CICD #PipelineEngineering