links

MiroFish — looks really interesting. It builds agents based on text you feed it (anything really!) and simulates them. The genius idea here really is, you don't even have to think about agents, you need it, it'll figure it all out itself.

You feed it a document, and it builds a world of AI agents that argue about it on simulated Twitter and Reddit. Each agent gets an LLM-generated persona — age, MBTI, profession, stance on the issue — and they interact across both platforms simultaneously.

Writing code is cheap now - Agentic Engineering Patterns. Some good points by Simon:

Coding agent tools can help with most of this, but there is still a substantial burden on the developer driving those tools to ensure that the produced code is good code for the subset of good that's needed for the current project.

The challenge is to develop new personal and organizational habits that respond to the affordances and opportunities of agentic engineering.

These best practices are still being figured out across our industry. I'm still figuring them out myself.

Karpathy has a video explainer on LLMs for a general audience. I can't think of a better and more accessible introduction to LLMs. Its helpful to know what are embeddings beforehand to better understand LLMs.