Prototyping

The first time an AI prototype I built got dismissed as a ‘bedroom project’, it felt personal. By the third time, I started to see a system. In this post I name the patterns AI builders hit inside large enterprises — provenance attacks, role-legitimacy challenges, capture mechanisms, balkanised AI efforts, the owner gauntlet, brittle delivery systems, induced in-group rivalry, and the shape-shifter’s double bind — show that each one is documented in organisational research, and lay out the tactics I have tested for getting past the dismissal hump.

I tried to recreate a Google AI Studio proof-of-concept in code. The first path failed on authentication; the second produced a working bidirectional streaming prototype; the third pushed it toward a realistic scenario.