It’s race over with him the winner. Open source can’t afford that data. Google, OpenAI and Anthropic would be the only ones able to train a “legal” model. Chinese models would get banned the day after.
I mean. There are other models that already only train on their own legally held data. Getty Images, and Shutter Stock for instance.
The main problem is that this is in direct reaction to other models scraping them for training data and them trying to monetize their own IP as a result.
It’s race over with him the winner. Open source can’t afford that data. Google, OpenAI and Anthropic would be the only ones able to train a “legal” model. Chinese models would get banned the day after.
I mean. There are other models that already only train on their own legally held data. Getty Images, and Shutter Stock for instance.
The main problem is that this is in direct reaction to other models scraping them for training data and them trying to monetize their own IP as a result.