Artificial Intelligence1 min read

What Picasso taught me about building AI Agents

By · Published by Everything Blog

In short

The modern approach organizes agents within basic directory structures containing instructions, skills, and context saved as plain Markdown files. Agents execute tasks using only minimal necessary tooling without relying on heavy external memory systems. These self-contained units can be triggered autonomously via cron jobs or webhooks, using headless commands like Claude's `-p` flag to generate raw results or structured JSON.

Key points

  • Advancements in core AI models eliminate the need for overcomplicated software abstractio…: Advancements in core AI models eliminate the need for overcomplicated software abstractions.
  • Self-contained project folders serve as clean, versionable units triggered by webhooks or…: Self-contained project folders serve as clean, versionable units triggered by webhooks or cron jobs.
  • Session troubleshooting is simplified by reviewing locally saved JSONL transcripts of too…: Session troubleshooting is simplified by reviewing locally saved JSONL transcripts of tool calls and outputs.

Sources checked by our analysis pipeline: 1

The great master Pablo Picasso is famously attributed with saying:

"It took me four years to paint like Raphael, but a lifetime to paint like a child."

His work, Le Taureau (The Bull), captures this beautifully. Across 11 iterations, Picasso progressively stripped away the details of a bull until only a few lines remained. Yet the bull was unmistakable.

In my journey of learning to build AI agents, I can relate. It has come full circle.

Like many engineers, I started with abstractions and sophisticated software systems. Frameworks, orchestration engines, multi-agent architectures, state machines, vector databases, memory systems, and elaborate planning mechanisms.

The more I experimented, the more I found myself stripping things away.

The models were getting better. More things worked with a single prompt. But that wasn't the whole story. I had been overcomplicating things.

This is now my workflow.

I open my terminal.

mkdir my_awesome_agent

cd my_awesome_agent

claude

and start a conversation.

"Claude, I want to build an agent that does XYZ. Help me initialize it. Create an AGENTS.md and let's figure out what else it needs."

Artificial Intelligence