The Memo - Special Edition - GPT-4 release - 14/Mar/2023
OpenAI's GPT-4 model is a visual language model (VLM) like DeepMind Flamingo
FOR IMMEDIATE RELEASE: 14/Mar/2023
Welcome back to The Memo.
Get ready for the avalanche of 1 billion+ tech articles about this slightly-underwhelming model!
The BIG Stuff
OpenAI releases GPT-4
I’ve spent a few hours summarizing the paper and playing with the model directly, and I reserve the right to update this summary as I continue reading and testing:
Note that OpenAI have chosen to hide nearly all details about the model. In the GPT-4 paper, they note:
Given both the competitive landscape and the safety implications of large-scale models like GPT-4, this report contains no further details about the architecture (including model size), hardware, training compute, dataset construction, training method, or similar.
Try it (must be a ChatGPT Plus subscriber at $20/m): https://chat.openai.com/
Read more: https://lifearchitect.ai/gpt-4/
My ‘Journey to GPT-4’ livestream will begin in just over 48 hours from this email.
EDIT: 17/Mar/2023: Here is the livestream replay:
All my very best,
Archives | Unsubscribe new account | Unsubscribe old account (before Aug/2022)