Not impossible but there are risks in that plan. We could be in a similar problem than chatGPT is having, where part of the combinations would end being wrong a

Not impossible but there are risks in that plan. We could be in a similar problem than chatGPT is having, where part of the combinations would end being wrong and generating newer info, that would reinforce the error.