Hi,
I recently found a product (datawizz.ai) that introduces(?) the concept of SLMs (specialized language models). As far as I understand the claim is that they are a proxy which captures inputs and outputs from OpenAI and other providers, and uses the data to train OSS models like Llama.
They look like a serious company, yet it looks like a violation of OpenAI’s TOS.
Am I missing something here? The idea looks totally promising, but I don’t want to risk my account.
Thanks!