The blog
Thinking, in the open.
Research notes, practical questions, and ideas we are still working through.
Editorial preview
Notes marked as drafts are newly written for the site, not previously published research articles.
A decision is not a sentence.Why a probability over a few well-defined options can be more useful than another paragraph.Editorial draftOne EEG model is not the whole interface.Reconstruction, representation, and intent decoding are different jobs. The distinction matters.Editorial draftThe model is only half the release.On input contracts, numerical parity, and the work that makes a research artifact usable.Editorial draft
Follow the work
Good questions. Measured answers.
New models, their evidence, and the occasional correction. Model cards come first; this site follows.