Working in the open
Why we plan to publish the notebooks, the negative results, and the reasoning — not just the finished paper.
There's a version of research that only shows you the destination: a polished result, a clean figure, a confident conclusion. It's efficient to read and almost useless to learn from, because it hides the part where someone was confused and slowly became less confused.
We want to do the opposite, as much as we responsibly can.
The case for small
Small teams have one structural advantage: nothing sits between a question and the person trying to answer it. There's no roadmap to defend, no quarter to hit. When a result points somewhere unexpected, you can simply follow it. Most of the findings we care about came from someone being allowed to chase an anomaly for a week.
The case for open
Sharing work in progress costs something — it's tidier to wait until you're sure. But three things happen when you publish early and often:
- You get corrected sooner, by people who know things you don't.
- Your negative results save someone else the same month you just lost.
- You think more clearly, because writing for an audience forces you to.
The goal isn't to look finished. It's to be useful to the next person who picks up the thread.
What that looks like here
Concretely: when we run an experiment worth talking about, we'll write it up — including what we expected, what actually happened, and where we think we were wrong. Some posts will be short notes. Some will be closer to papers. A few will simply say "this didn't work, here's why we think so."
If that's the kind of research you like to read, subscribe — and if you disagree with any of it, we'd genuinely like to hear why: hello@anthelic.ai.
— The Anthelic team