I work on a CETI project (communicating with E.T.) and wonder how OpenAI could be canned and sent to the moon and turned back in 3 billion years when ET visits Earth. I’ve read that “It’s powered by an astonishing 285,000 CPU cores and 10,000 GPUs.”. But would it be easier just to save the 45TB text behind it and let ET figure out how it can respond? Or the AI is a truely human invention and the text output would never chat with ET without it? So perhaps if you save the 45TB text and X code, these two with a single sample CPU/GPU would be enough to save it and let ET chat with “humans” and explore our way of thinking?
Related Topics
| Topic | Replies | Views | Activity | |
|---|---|---|---|---|
| Build yourself a portable AI core! | 3 | 677 | November 23, 2023 | |
| OpenAI’s Foundry will let customers buy dedicated compute to run its AI models | 5 | 7193 | March 28, 2023 | |
| OpenAI & Ensuring that AI technology isn't paywalled from the less fortunate | 2 | 464 | December 28, 2023 | |
| Offloading to local GPU to help expand OpenAI hardware access? | 0 | 302 | April 8, 2023 | |
| Letter from Galileo Galilei | 0 | 268 | October 12, 2023 |