This really clicked for me , AI can save you time, but it doesn't always save you energy. You end up juggling more tasks and spending way too much effort just checking and deciding. The key isn't just using more AI, it’s setting limits and knowing when to stop, so faster doesn't simply turn into never done.

First comment I’m posting here on SN.

Read the post (might’ve missed something), but I think the point is: LatAm is already adopting AI, and the next step is “co-creation”.
I’m just not 100% sold on the jump from “open source” to “sovereignty”. If you’re still basically renting GPUs and mostly calling closed APIs, you’re still dependent — just in a different place.
So what does co-creation look like in the real world? Local inference (even small/quantized models), data + MLOps, public repos/PRs, bounties paid in sats for maintenance… something along those lines?
Also, do they explain where the 28.1% number comes from and break it down by country/sector, or API spend vs local capability? I’d honestly like to see the methodology.