The six-month rookie error

The frame is compound interest: a good habit reinvests into itself and curves upward, a bad one is quicksand, and the difference at five years is enormous ▸ 1:55. The operative corollary he keeps from Vercel’s Guillermo Rauch: get to your own feedback loop as early as possible ▸ 5:01, because an unexamined habit compounds unexamined.

Then the confession the title promises. Their first venture, pre-Tencargo, was Innovation Station, a kids’ channel about the process of inventing, and they launched it as seven or eight channels at once, Spanish, English, Portuguese, Italian, Russian, Arabic, Chinese, “porque podíamos” ▸ 9:55, on the premise that more channels meant more chances something goes viral ▸ 10:20. The production floor of that premise: eight-hour days of generating narration by playing Google Translate’s voice at ChatGPT from a phone and re-recording the repetition, per topic, per language, then video, then shorts ▸ 11:02. Verdict: “nos quemamos solitos” ▸ 11:52. Cost: six months, not the four he first remembers ▸ 10:40.

la premisa que no examinaste vuelve disfrazada de idea nueva →

The expensive part is that the lesson didn’t transfer. The same premise reappeared later wearing different clothes, “hagamos la mayor cantidad de programas posibles, alguno va a pegar” ▸ 7:55, and unlearning it the second time took four more months ▸ 8:33. The do-over prescription is the inverse of the spray: one channel, one language, and small cheap tests, this topic, this voice, this format, “todo es como hacer una receta”, iterate until the recipe works ▸ 12:01. And the honest postmortem includes the assumption they never tested at all: that kids want content about imagining new things, when YouTube’s gravity says people overwhelmingly want to learn what already exists ▸ 12:57. Six months bought them the feedback loop they could have had in six weeks, if they’d stopped to close it…