One area where I think people in the GPT-3 community don’t discuss at all - is simply around GPT-3 access.
I think part of the reason is because no one wants to lose exclusive access to GPT-3 and are genuinely scared to speak up about it. I’m scared too - I love GPT-3 and like to think of myself as a super fan, losing access would hurt me deeply, but also I can emotionally relate to people who still haven’t got access.
Lately, I've been obsessed with multimodal AI content generation, but a topic which still touches a nerve for me is around the OpenAI beta application access process.
As far as I know, I was the first person to openly share how I got access to GPT-3 and create a comprehensive video about the topic, simply because I felt bad for people who wanted access but couldn’t get it. Many people over the past few months have thanked me for creating this video as it has helped many get access.
But the open questions still remain. Who gets access and who doesn’t? What is the selection criteria? How are we thinking about topics like diversity and building our futuristic AI developer community?
Why should I be the one to make a video on this topic? I’m not an expert, I don’t work for OpenAI. They haven’t paid me, either (I don’t want the money). Why haven’t they been the ones to make a video on it?
GPT-3 has almost been around for a year now. To my knowledge, there is no official statement from OpenAI about any of these topics and this creates a huge vacuum in our community. With a lack of official guidelines and virtually no transparency from them, many people I’ve spoken to who want access simply don’t know where to begin. At the same time, there are many people who have communication challenges, speak different languages, come from different backgrounds but still want access. What are we doing to help these people apply and get access to GPT-3?
Over the next few weeks, I will be reaching out and connecting more with the people who work at OpenAI. I’ve had initial conversations already, but to be honest, they were not productive and they did not take my suggestions seriously at all. I was personally pretty upset about it, because I feel I have also contributed a lot to this community and was seriously open to being constructive and collaborating. I’m not a journalist or something. I’m not out for blood or have some crazy vandetta. I’m a builder, creative technologist, and most importantly a GPT-3 developer.
I’ll do my best and I can’t guarantee where it will go. It may mean the end of my YouTube channel, which is entirely about GPT-3, but it’s a risk I’m willing to take to advocate for a topic no one else wants to talk about openly.
Update March 10, 2021 - Just following up for anyone reading this post - I had a chance to chat with OpenAI and we had a really productive discussion where I raised these concerns. Thanks everyone for their feedback!