The article cites replication as important and then paradoxically argues for releasing the model in lieu of replicating the training and verifying the result.
This is the AI equivalent of a pharmaceutical company releasing all their research and patents and then taking heat for not providing reagents and doing the organic synthesis for everyone too.
The difference is a pharmaceutical company is for-profit, while OpenAI's (a non-profit) mission (https://openai.com/about/) is "to build safe AGI, and ensure AGI's benefits are as widely and evenly distributed as possible." (emphasis mine)
I don’t see how corporate structure changes anything. It’s a huge gift of tons of work, which is far and away the hard part. They’ve enabled literally anyone who wants to to train their own model on whatever source material they want.
The article cites replication as important and then paradoxically argues for releasing the model in lieu of replicating the training and verifying the result.
This is the AI equivalent of a pharmaceutical company releasing all their research and patents and then taking heat for not providing reagents and doing the organic synthesis for everyone too.