A page about me, and a roadmap that stopped hedging.
The rest of the site is deliberately terse — that's correct for claims a skeptical reader needs to check. It's the wrong voice for the one page whose entire job is trust through personality. And a roadmap pass that moved things both up and out, for once, not just up.
Two unrelated things today, both about honesty in a different direction — one about who's actually behind this, one about what's actually coming.
An About page
Someone read the site cold and told me it felt terse and a little emotionless — accurate, and mostly on purpose: a comparison table, verified proof numbers, an architecture page that states its own limits. That precision is correct where a claim needs checking. But About isn't a claims page. Its whole job is warmth, and there wasn't anywhere on the site that had room for it.
So: a real page, a real photo, and the actual story — I built Neeos for myself first, because it was the AI product I wanted and nobody was building it. I told family and friends about it, and the reaction told me there might be a market for it beyond just me. I run restaurants for a living and build the software those restaurants run on; Neeos has been a passion project alongside that. One person, not a company with a team. Not yet, anyway.
The one boundary I kept: the terseness stays everywhere else. It's load-bearing for the technical audience this product is actually for — a homepage or an architecture page that got chatty would read as less credible, not more approachable. About is the exception, not the start of a site-wide tone shift.
A roadmap that removes things
Every roadmap I've seen only grows — items get added, and the ones that quietly died just sit there forever, unfulfilled, because removing a promise feels worse than never having stated it precisely. Ours got a pass in the other direction too, and I think that's the more honest habit to build.
Cut outright: a web app. It had been sitting in the "planned" tier since the architecture page shipped, hedged as an intention. It isn't one — there's no web client, there never has been, and there isn't going to be. Leaving it on the roadmap was a softer way of saying "maybe," and "maybe" wasn't true. Google Drive sync went the same way for a different reason: it's tabled, not planned, and a roadmap item that isn't being worked on doesn't belong on a roadmap.
Promoted, because it's real: local LLM processing — running the language model itself on-device — moved from "planned" to active development, alongside two new active items: Apple Pencil support across Notes, PDFs, and documents, and offline processing using local models for the parts of adding something to Neeos that don't need a live connection to start. All three are real code, not intentions.
The roadmap's whole point was always "labeled honestly, not the same as done" — and a list that only ever grows fails that test just as much as one that overclaims. Taking things off is the other half of the job.