The world of AI-powered coding saw major breakthroughs this week, with tools becoming more powerful and easier to use than ever before.

Eliza Labs made waves with their auto.fun launch, a platform that turns complex Web3 operations into simple drag-and-drop tasks. Users worldwide can now create AI agents that automatically manage cryptocurrency wallets, trade digital art, or interact with blockchain games – all without writing a single line of code. Early users in Tokyo have created agents that monitor NFT prices 24/7, while a school in Berlin is testing agents to help students learn about smart contracts through visual templates.

Accessibility remained a key theme, with tools bridging the gap between professional developers and beginners. While not released this week, free services like Gemini Code Assist continue gaining popularity for their ability to explain coding errors in plain language. A 12-year-old in Mexico City recently used such tools to fix a website bug, showing how AI coding assistants are making technology creation more inclusive.

New research highlighted both progress and limitations. Studies show top AI models still fail on 74% of real-world software tasks, struggling most with understanding vague instructions. This reinforces the need for human-AI teamwork – where people handle creative problem-solving while AI speeds up repetitive tasks like testing code or finding files.

The rise of AI agents brings new challenges too. Security experts warn that poorly designed agents could make mistakes in financial transactions or data handling. This has led to growing demand for “AI agent supervisors” – human roles that monitor and guide automated systems, particularly in banking and healthcare applications across Europe and North America.

Looking ahead, educators are exploring how to teach AI agent design in schools. A pilot program in South Korea now includes basic AI workflow creation in their computer classes, preparing students for a future where managing smart helpers becomes as fundamental as using spreadsheets or word processors.