Latest story• Jan 12, 2026
Why 90% Accuracy in Text-to-SQL is 100% Useless
Text-to-SQL needs 100% accuracy in enterprise settings, but LLMs achieving 90% on academic benchmarks drop to 10-20% on real enterprise databases.
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
Yolobox – Run AI coding agents with full sudo without nuking home dir
YoloBox sandboxes AI coding agents in containers, giving them full permissions while protecting your home directory from accidental deletion.
Even Linus Torvalds is trying his hand at vibe coding (but just a little)
Linus Torvalds used Google's AI coding tool Antigravity for his AudioNoise guitar effects project, calling the approach 'vibe coding.'
Fall asleep by watching JavaScript load
Web app uses progressively slower JavaScript loaders and bedtime stories to help developers fall asleep.
How I used ChatGPT's $20 Plus plan to fix a nightmare bug - fast
ChatGPT Plus successfully debugged a WordPress plugin issue, identified server-level caching problems, and drafted hosting provider support emails.
Markdown Is a Disaster: Why and What to Do Instead
Developer argues Markdown is a flawed lightweight markup language and advocates for better alternatives like Orgdown.
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
Why I built a Kubernetes deployment tool
New CLI tool Displace translates Docker Compose setups into production Kubernetes, eliminating weeks of YAML configuration struggles for development teams.
Type-In Rescue: The C64 Autoboot Generator
A deep dive into refactoring a 1984 Commodore 64 auto-boot loader generator, reducing complexity and typing requirements while preserving functionality.
Coding Agents As An Interface To The Codebase
AI coding agents are emerging as conversational interfaces that let developers query and modify codebases using natural language instead of traditional IDEs.
How to practice Feature Testing with PHP
Feature tests verify entire system workflows rather than isolated components, catching integration bugs that unit tests miss in real user scenarios.
Devs doubt AI-written code, but don’t always check it
Despite believing AI code is buggy, only half of developers consistently review it before committing, creating verification bottlenecks.
How To Start A Mastodon Development Environment
Complete guide to setting up Mastodon's Rails-based development environment locally using Docker or manual dependency installation.
I Replaced Redis with PostgreSQL (And It's Faster)
Developer replaced Redis entirely with PostgreSQL features, trading slight performance for reduced costs and operational complexity.
The Next Two Years of Software Engineering
Five key questions will define software engineering's future: junior hiring, AI skill impact, role evolution, specialization strategy, and education relevance.
Hongdown: An opinionated Markdown formatter in Rust
Hongdown is a Rust Markdown formatter enforcing Hong Minhee's style conventions with distinctive formatting rules.
annote: Writing java using only annotations
Annote turns Java annotations into a programming language, executing @Var, @If, and @Loop statements through reflection instead of traditional code.
Free Dockhand Tool Simplifies Docker Container Management
Dockhand offers lightweight Docker container management for home labs, providing monitoring, troubleshooting, and deployment features through a simple web interface.
Is the Cult of Constant 'Trying Things Out' Killing Your Engineering Efficiency?
Constant tech experimentation without strategy hurts productivity through technical debt and context switching; data-driven approaches can balance innovation with efficiency.
Moving beyond RxJS: A guide to TanStack Pacer
Build a Pinterest-style React gallery using TanStack Pacer to handle debouncing, throttling, batching, and rate limiting without RxJS complexity.
6 fast (native) alternatives for VSCode
Native code editors built with Rust and C++ are emerging as faster, lighter alternatives to VSCode's resource-heavy Electron architecture.
Beyond Python: Why LLMs Need More Stable, Open Source Code
LLMs heavily favor Python in code generation due to training bias, but will likely shift toward more stable languages as maintainability becomes priority.
whenwords: A relative time formatting library, with no code
whenwords is a code-free library that uses AI prompts to generate relative time formatting functions across multiple programming languages.
How to Build an In-Memory Rate Limiter in Next.js
Tutorial for building an in-memory rate limiter in Next.js with fixed window algorithm, including Artillery load testing for validation.
Building a Game Engine in Node.js
Developer builds impressive 2D game engine in Node.js, battling coordinate system translations and physics engine limitations.