Non-Technical Member of Technical Staff™ @CosmicStack
I’ve known how to code for a long time. That sentence sounds more impressive than it should. I could read code. I understood what was happening. I had worked on projects. Give me something that
Every week, another AI agent launches. A new benchmark. A new model. A new reasoning technique. A larger context window. Another autonomous workflow. The progress is real. Today’s AI agents can
Your agent should not need a skill for everything. If your agent needs a skill for everything, it is not ready. I have spent enough time building agents to believe this strongly: It should not need a
A technical look at why the LLM Wiki pattern resonated, where it starts to fail at machine scale, and what serious agent memory likely looks like next. When Andrej Karpathy shared his LLM Wiki
The problem with AI agents right now Personal AI agents have a security problem, a cost problem, and an identity problem and most of them ship all three at once. On security: agents run shell
