How Latin America Built Its Own AI Model for $550K | Rodrigo Duran Rojas, LatamGPT

We do have this network of neutral institutions in different countries that at the very end are legitimate for the AI community.

Working with the academia legitimates you not only in front of your peers but also in front of the other governments that may change.

We have today an amazing guest, Rodrigo Duran Rojas.

He's the executive director of Chile's National Center for Artificial Intelligence, which is CENIA, and one of the principal architects of LATAM GPT, the first open large language model for from and for Latin America and the Caribbean, a 70 billion parameter model trained by more than 60 institutions across 15 countries and on a budget of roughly $550,000.

Just sit with that number because it tells you what innovation can happen when there is some constraints.

So we're going to talk to him and talk a lot about different, different things.

Rodrigo has been kind enough to speak to us.

He's commuting on a train, but some of these best conversations sometimes happen on the train.

So Rodrigo, welcome to the CIO Connect podcast.

Thanks,

Steve.

Sanjay, I'm very happy to join you today and so sorry for the circumstances around me, but I really wanted to have this conversation today.