What's the deal with ChatGPT's "No humans were in the loop"?

I don't understand the distinction between ChatGPT and Claude, as apparently both are trained on large datasets of text and couldn't function without humans who wrote the code that made the models possible, wrote the datasets the models are trained on, and manually fine-tuned the models.<br><br>What does "no humans in the loop" even mean? What's the difference supposed to be?<br><br>Furthermore, isn't it hypocritical of OpenAI to boast about how they don't use humans to train ChatGPT when they are hiring people to train and fine-tune their models manually?