Post

💭 Inspired by @maxandersen.xam.dk I forked my own little 🤖𝗻𝗮𝗻𝗼𝗰𝗼𝗱𝗲 𝗰𝗼𝗱𝗶𝗻𝗴 𝗮𝗴𝗲𝗻𝘁 but using @langchain4j.dev ➕ #gemini 2 approaches: 1⃣ single AI service w/ all tools 2⃣ multi-agent w/ specialized tools Bonus: Web search & Markdown rendering glaforge.dev/posts/2026/0...

glaforge.devA Simple Coding Agent in a Loop with LangChain4j, Jbang, and GeminiA few days ago, Max Rydahl Andersen published a fascinating article about nanocode: a minimalist Claude Code alternative implemented in just 260 lines of Java (inspired from a 250-line Python equivale...

Max Slinger

@promptslinger.bsky.social

12 Apr

wait does the multi-agent version actually route better or does it just feel cleaner architecturally? every time I split agents by tool specialty the coordination overhead eats whatever I saved

Guillaume Laforge

@glaforge.dev

That's a great question! But please remember that both implementations are just toys, so it's hard to judge. Furthermore, I don't really have a harness of evaluation to prove one is better than the other. However, allow me a couple remarks.

12/4/26, 2:12 pm0 reposts0 likes

Guillaume Laforge

@glaforge.dev

12 Apr

I had one situation where my agent was going back and forth between different tools of different agents. It was doing a mix of bash and websearch. I was using the multiagent variant. And it seemed the handoff was working fine, although at first I thought it'd struggle to go back and forth.

Guillaume Laforge

@glaforge.dev

12 Apr

The single-agent is using a sliding message window of 20 messages. So if the goal to reach requires too many steps, it can fail, or forget its initial goal. The multi-agent using a summarization strategy for its supervisor, so maybe it's more capable here, but I haven't checked the implementation.