About

One AI news issue a day, written so you can check every claim.

LodeHQ covers new AI models, tools, research and policy for people who build with AI or decide what to adopt. Each afternoon we publish one issue. It ends, and then you're caught up.

Every story has three parts

1. The claim, in the headline. "Mirendil's first model, Fern 0.1, can read up to 10M tokens at once." Read only the headlines and you still know the news.
2. How it works, in plain words. Three to five sentences, with jargon explained where it appears and a number from the source: "the context window is how much text a model can read in one go; here that is about 23 MB of code."
3. The original source. The announcement, paper or repo, never a comment thread, so you can check our reading.

Why "lode"

lode /lohd/, noun: a vein of valuable ore running through ordinary rock.

Most of what gets posted about AI each day is rock: reposts, takes, and the same launch told ten times. A few things actually changed. We follow that seam and leave the rest.