By skyforbes My theory is that these humanizing tools are essentially fine-tuned language models with carefully crafted prompts behind the scenes. If that's the case, couldn't you theoretically achieve similar results by just engineering the right prompts for standard LLMs like ChatGPT or Claude? What would those prompts even look like to get comparable humanization quality?
I'm curious if anyone has experimented with this approach and found prompt patterns that actually work for making AI text sound more natural and human-written. Or do these specialized tools have some additional processing that goes beyond what prompting alone can achieve? It would be great to avoid the account limitations and just handle everything through direct API calls with the right prompting strategy.