Multimodal by Bakz T. Future

Share this post

User's avatar
Multimodal by Bakz T. Future
OpenAI may be text-only for a while

OpenAI may be text-only for a while

Bakz T. Future's avatar
Bakz T. Future
Oct 15, 2021
∙ Paid

Share this post

User's avatar
Multimodal by Bakz T. Future
OpenAI may be text-only for a while
Share

As much as I am excited for DALL-E and multimodal models in the future, I think OpenAI may be focused on GPT-3 and Codex for a while.

This is because of two reasons - the compute costs are really high to train and host multimodal models and also there's still a lot left to explore with text based models like GPT-3.

A faster, safer, more reliable version o…

This post is for paid subscribers

Already a paid subscriber? Sign in
© 2025 Bakz T. Future
Privacy ∙ Terms ∙ Collection notice
Start writingGet the app
Substack is the home for great culture

Share