2023-05-29
-
OpenAI’s plans according to Sam Altman
- Deleted as of 2023-06-02 WebArchive.
-
Japanese translator
-
GPU resource crunch
- The reason multimodality in the early demos of GPT-4 is not available to the public is because we don’t have enough GPUs.
- Longer context width can be done, but Transformar still can’t provide it since it is O(N^2)
- Related: The process of resource exhaustion.
-
The scaling rule is still in place (because it’s eating up computing resources), but it’s only increasing at a rate of two to three times per year.
This page is auto-translated from [/nishio/OpenAI’s plans according to Sam Altman](https://scrapbox.io/nishio/OpenAI’s plans according to Sam Altman) using DeepL. If you looks something interesting but the auto-translated English is not good enough to understand it, feel free to let me know at @nishio_en. I’m very happy to spread my thought to non-Japanese readers.