Picky is not Menu AI

An email arrives from EWOR, a B2B-focused fellowship: the initial application passed review, complete the next form, record a sub-90-second video ▸ 18:41. Healthy skepticism is logged, maybe it’s automated, maybe the LinkedIn posting was just a referral funnel ▸ 21:42, but the prize is real enough to play: ~$500k in funding and mentorship in exactly his declared weak spots, hiring, marketing, finance ▸ 24:04.

The fellowship wants B2B, and the portfolio audit takes one sentence: “el único business-to-business que tenemos es Picky” ▸ 20:09. So the restaurant-menu app comes off the shelf, and with it the entry’s subtler teaching: the features Juan once ordered cut, “Julia, no, quitemos todo esto, vamos a hacer simplemente un MVP”, get reinstated, because the audience changed: “ahorita como ya vamos a tener que dar la idea final, o sea, convencer a los inversores, pues hay que dar precisamente todo eso” ▸ 20:48. MVP scope is for users; vision scope is for investors. Same product, two truthful sizes, chosen by who’s asking.

el MVP se recorta para usuarios; la visión se cuenta a inversores →

And then the moment that names the entry. Gemini, ever helpful, suggests the product sound more professional, something like “Menu AI”. The refusal is instant and total: “y yo: Picky, Picky. Ese nombre ya está, ya está el branding, no nos vamos a poner a molestar con eso” ▸ 24:39.

It’s the mirror image of entry 60-1’s architecture pushback. The AI optimizes for the generic case, and “Menu AI” is the generic case: descriptive, forgettable, indistinguishable from a thousand SEO plays. Picky has a ring menu, a personality, a bot named Pickyto, sunk brand equity the model can’t see. Knowing when to take the AI’s advice built this year; knowing when to refuse it is what keeps the year theirs…