Dec 24, 2025 · 3 min read · The Problem: Context is a Graveyard In software engineering, having Git means we don’t just store code — we preserve its history. We know which decisions are current and which ones are legacy. Now in LLMs we use transformers, a neural network archite...
Join discussion
Dec 9, 2025 · 6 min read · At the start of this whole journey, I naively thought the hard part would be “storing stuff.” Turns out the real difficulty is getting an LLM to use stored information consistently, without guessing, drifting, or confidently inventing memories that n...
Join discussionNov 14, 2025 · 3 min read · Welcome to Day 3 of the 5-Day Agentic Revolution! On Day 1, we introduced the agent's Brain (the LLM). On Day 2, we covered the Hands (Tools). Today, we tackle the most human-like capability: Memory. The core challenge of building advanced AI is simp...
Join discussion
