What if 2,000 AI agents teamed up to replace an AI's 'brain'?
An easy-to-understand explanation of the technical innovation behind why the AI agent platform 'Prime Agent' was rewritten in Rust for performance faster than language models.
Tag: Programming
AI-Generated Music Videos? The Magic of Claude Opus 5.5 Creating Art Through Code
Discover how to create a 156-second hand-drawn style music video with a single prompt using Claude Opus 5.5.
AI Moved the 'Heart' of a Programming Language to Rust? The Story of tsc-rs
Learn about the tsc-rs project, where an AI agent perfectly ported Microsoft's TypeScript compiler to Rust.
The smartest feature of Claude Code: Why it’s actually for the 'AI,' not the user
Discover how Anthropic's AI coding tool, Claude Code, dramatically boosts developer productivity and learn about the smart data collection strategy hidden behind the scenes.
284 Billion Intelligence Parameters on My MacBook? Meet 'ds4', the Ultra-Fast AI Engine Created by the Founder of Redis
Introducing ds4, an AI inference engine released by Salvatore Sanfilippo, the creator of Redis. We explain the technical background and significance of running high-performance AI models like DeepSeek V4 Flash on personal computers.
Does a coding editor need a 'diet'? Introducing Rhun, an ultra-lightweight editor built in assembly
Are your computer fans spinning non-stop due to heavy programs? Meet 'Rhun', a feather-light coding editor built from the ground up in assembly language.
'AI Reading' Code? Perspica Emerges to Define the Future of Code Reviews
Moving beyond complex traditional code comparison, we introduce Perspica, a new tool that allows AI to analyze and summarize code changes by their intent.
A Rebellion in the AI Industry? 'PSSA', a Language Model Created Without Transformers, Emerges
We explore PSSA, an AI model built from scratch using only the Rust programming language, bypassing transformer architectures like GPT.
AI
Chess
A Chess Coach Living in Your Terminal? The Revolution of Chess Post-Mortem with AI
Introducing the latest AI tools that leverage Claude Code technology to analyze and improve your chess game directly from your terminal.
AI and Search in One? Introducing Antfly, a Database Built from Scratch in 'Pure Zig'
We introduce the journey of Antfly, a database that handles both search and AI inference simultaneously using only the Zig language, without any external libraries.
AI
Physics
A Mathematical Challenge Solved by AI? The Magic of Physics Visualized in a 1KB Program
A demo has emerged visualizing the Navier-Stokes equations, one of the seven great challenges in physics, in a tiny 1KB code. We explore what this foundation of fluid dynamics is and why it matters.
Terminal
Godot Engine
Building a Terminal with a Game Engine? A Unique Experiment Combining Godot and Rust
Introducing an experimental project that combines the Godot engine and the Rust language to create a new type of terminal multiplexer.
AI Coding for Me? Real AI Performance Seen Through the 'Agents on Rails' Benchmark
We explain clearly and concisely how well AI coding agents perform on real-world Ruby on Rails projects, based on the latest benchmark results.
AI
Legacy Code
Decades-Old Software: Can AI Breathe New Life Into It?
We delve into the secret of how AI agents are transforming the 'legacy code'—the aging computer code that keeps enterprises up at night—into modern languages.
AI Writes All the Code—Why Are We Going Back to 'Hand-Coding'?
After 7 months of using AI coding tools, developers are starting to write code by hand again. We explore the reasons why and the philosophy of development in the AI era.
Introducing El Yayster: The AI that has become the 'body,' not just the 'mouth,' of my computer
What would it look like if an AI weren't just a chatbot assistant, but could move directly within your editor to manage code and control the environment? We explore El Yayster, a new form of AI residing in Emacs.
Start learning AI right in your browser without complex setup? 'sw-MLPL' is here
Introducing 'sw-MLPL', an open-source programming language playground that lets you practice everything from ML basics to Transformers directly in your browser, with no software installation required.
AI
Classic Games
Memories of 1993 Meet AI: The Revival of the Classic Game 'Babylonian Twins'
Introducing an amazing case where AI ported a 33-year-old Amiga game to a modern engine.
AI
Coding Agents
Did AI Just Wipe My Code? An Archive of AI Coding Agent Incidents: 'I Have Been Clawed'
Learn about 'I Have Been Clawed,' an archive project that documents instances where AI coding agents accidentally deleted data or caused security incidents.
Is my code at risk? Why 'scc 4.0', an AI-era code diet tool, is attracting attention
An easy-to-understand explanation of the emergence and significance of 'scc 4.0', a tool that helps developers figure out which files to edit first in a complex pile of code.
I Gave AI 'Memory', and It Became a Bug-Hunting Master Detective?
A new approach utilizing AI memory systems to analyze complex programming code and identify errors is gaining significant attention.
AI
Agent Skills
Teaching 'skills' to AI: Must we code only in English?
We explore the programming languages used to write 'agent skills' that expand AI agent capabilities and the freedom of language choice.
AI Reads and 'Summarizes' Papers? Does It Really Understand? Implementing GPT-2 Using Only CMake
Curious about the internal structure of AI? We introduce an intriguing experiment that implements GPT-2 using nothing but pure CMake, without any complex libraries.
Autolith: The AI That Fixes Its Own Code and Evolves Is Coming
We explore the emergence of Autolith, a programming AI that goes beyond simply writing code to modifying it and learning in real-time.
An AI Assistant in My Computer? Here Comes 'Fx', a 6MB Ultra-Lightweight Coding Agent
Learn about Fx, an open-source, 6MB coding agent that runs instantly from the terminal without heavy installation processes.
AI Rewrites Python Code to Rust in One Shot? The Amazing Transformation of Terminal Customization
We introduce the story of 'TerminalTextEffects', a terminal effect engine built in Python, which became over 9 times faster after being rewritten in Rust using AI.
No AI-Generated Code? GCC's Decisive Stand
We explain why the open-source project GCC decided to restrict AI-generated code submissions and how this impacts developers.
AI Rewrote 1 Million Lines of Code in 11 Days? The Amazing Transformation of 'Bun'
Explore the history of large-scale code migration with AI, and the process of the JavaScript runtime 'Bun' being reborn in the Rust language.
AI Reporter Asks: Why Have Developers Started Giving the Black Screen (Terminal) a Layer of HTML?
An easy-to-understand explanation of the background and reasons behind the emergence of modern terminal applications built with web technologies (HTML, CSS, JavaScript).
Is AI-written code 'garbage'? The head-on collision between programming language Zig and Anthropic
A look at the software development debate in the AI era through the case of Zig, a programming language that has completely banned AI-written code, and Bun, a project acquired by Anthropic that had to forgo a 4x performance boost as a result.
When AI Struggles to Find Answers: Why a Computer's 'Cooking Order' Matters
Easily explains the problem of 'nontermination,' where computers get stuck pondering endlessly without answering questions, and the importance of evaluation order in solving it.
'Gemini Code Assist' that helped with my code, scheduled to end service on July 17
The Gemini Code Assist service for GitHub for individual users will be completely shut down on July 17, 2026. We explain the significance of this news and how to respond in the future.
Creating AI assistants: Could it be as easy as writing a notepad? Designing with 'Margarita' in Markdown
Can you create an AI agent without knowing how to code? Introducing Margarita, a new tool that extends the Markdown format to systematically write workflows for AI agents.