I'm Ehsan Shahabi. I work in venture R&D — and this is my public lab. Nothing here was built to become a startup or a crowded platform. Each experiment has a smaller, honest job: test an idea fast, communicate it, signal what's possible. Judged as products, they fail. Judged by their real job, every one succeeded. That's the name.
Planning time with friends is a coordination problem: the friction isn't what, it's when. A mobile-first platform where people share free time slots and what they're open to — friends browse, join, get notified. Next.js, Neo-Brutalist design system.
An open-source Next.js app that generates polished presentations in minutes, with authentication, a credit-based usage model, and history tracking.
An interactive art generator inspired by post-impressionist paintings — JavaScript vector fields making Van Gogh-style swirls. Built for no reason except that it's beautiful.
An essay and prototype asking: what if the workspace followed your intent instead of making you shuffle windows? As you type, the tools you need appear.
Lets AI assistants read, search, and edit Microsoft OneNote notebooks through the Graph API — pages, tables, todos. Open source; other people actually use it, which still surprises me.
A free, open-source Looker Studio connector that runs recency-frequency-monetary customer segmentation on any Google Sheet — a classic hand-made analysis, turned into a tool anyone can run. Finalist at Google's Looker Hackathon 2023, named on the Google Cloud blog.
Not everything here was an experiment. These two aimed at the conventional goal — real ventures — and didn't reach it. Real attempts deserve autopsies, not euphemisms.
A community-management platform for organizing groups and their gatherings.
An open-innovation platform bridging corporate innovation demand with startup solutions — API gateway, sandboxes, open challenges.