Latest story• Aug 15, 2026
SpaceX has officially acquired AI coding startup Cursor
A $60 billion acquisition is now finalized, signaling a potentially major shift in the tech or business landscape.
SMRTR Programming
Programming stories
Get hand-picked daily summaries of the best, most informative programming articles from around the web. Browse recent summaries, then jump to the original source when you want the full article.
Latest stories
Stop Parsing URLs Manually in PHP
Stop writing regex to parse URLs — PHP's parse_url() and parse_str() combo has handled this cleanly since PHP 4.
The AI Situation in Software Development
AI doesn't eliminate implementation time — it just moves it from writing code to writing detailed specs and system designs.
What Are the Best Practices for API Security? A Complete Guide to Securing Modern APIs
A practical, code-backed guide to securing APIs using authentication, authorization, encryption, input validation, rate limiting, and monitoring in JavaScript.
Paperditor – a free, open-source WYSIWYG Markdown editor
Paperditor is a free, open-source WYSIWYG Markdown desktop editor that renders live as you type, with export and multi-window support.
Exploring Compose HTML for Server Side Rendering
JetBrains explores bringing Compose HTML to the JVM for type-safe server-side rendering, potentially replacing Thymeleaf and JSP with real Kotlin components.
SMRTR Programming
Get the next batch of curated stories in your inbox.
This archive is built from SMRTR newsletter stories. Subscribe for hand-picked stories without the extra noise.
More recent stories
What You Gain by Building Your Own Game Engine
Custom engines power just 13% of Steam games but 43% of units sold — here's why building your own still makes sense.
AI is removing the middle class of software engineering
AI-generated code is accelerating bad decisions, not eliminating the need for understanding. Engineers who can critically evaluate AI output are increasingly valuable; those who can't are a fast-growing liability.
Gemini 3.7 Flash
Google's Gemini 3.7 Flash brings multimodal reasoning with a 1M token context window, tiered thinking modes, and broad API capability support.
Caged Code – Vanilla Claude Code Binary in a Webpage
Claude Code now runs in a browser via a shared demo gateway, with optional OpenRouter API key support for any model.
How to Build a Production-Ready AI Agent for $0/Month Using PHP, cPanel, and Gemini Flash
Build a complete AI agent with PHP, MySQL, and Gemini Flash — including tool calling, conversation memory, and cPanel deployment.
The Hidden Cost of Software Rewrites
Greenfield rewrites keep failing because mature codebases hold institutional knowledge no spec captures — and reversibility is the real solution.
The simple way to add Bitcoin payments to your app
Second SDK brings Bitcoin Lightning and Ark payments to any app through one self-custodial integration, eliminating channel management headaches entirely.
Python for Mechanical Engineering
FreeCodeCamp's new 7-hour Python course targets engineers and covers NumPy, Pandas, Matplotlib, and AI-assisted workflows for real-world engineering problems.
Building an AI-assisted Receipt Reader that Doesn't Guess
AI receipt extraction for vehicle records taught that confident, well-formed AI output can still be factually wrong — and a seven-line prompt fix matters more than the model choice.
MCP-stama – An ultra-fast Rust MCP server with no dependencies
A single Rust binary MCP server that replaces heavy Node/Python tools with sub-millisecond response times and under 10MB RAM usage.
Coding was never the hard part
A developer argues that dismissing code as 'easy' misunderstands software engineering, which is about problem-solving, not just typing syntax.
I built a multi-step API testing tool that's designed to be self-hosted
TestFleet is a self-hosted, distributed API testing tool with multi-step checks, SSO auth, and Kubernetes-ready Helm charts.
Why To Use A Multi-Stage Dockerfile
Multi-stage Dockerfiles cut image size, reduce attack surface, and speed up deployments by keeping build tools out of production images.
Agents on Rails: the first benchmark report
Eight AI models benchmarked on real Rails tasks show Luna wins on cost, Opus on accuracy, but Rails API knowledge separates them all.
75% of developers I surveyed prefer Claude Code - here's why they choose it over Codex
138 developers weigh in: Claude Code leads on context and reasoning, Codex wins on cost and limits, and 20% use both.
How To Reduce Token Costs of AI Coding Agents
Every wrong AI coding guess compounds across a session, burning tokens repeatedly. The fix is surgical: match models to tasks, tighten context, and ground agents in live infrastructure data to eliminate costly correction loops.
Decant – Understand how you spend tokens
Decant gives developers a local analytics dashboard for Claude Code and Codex sessions — tracking costs, tokens, and agent activity privately.
How AI Agents Are Changing Software Development
AI agents pair LLMs with tools, RAG, and vector databases to tackle complex tasks — but security and reliability require serious engineering attention.