Started this experiment curious, ending it with some actual opinions
Month 1-3: Using AI to generate text and paste it in. Word count went up, quality went down, nothing sounded like me.
Month 3-5: Realised generation was the wrong use case. started using it to interrogate my own writing instead and results smh got more interesting.
Month 5-8: Figured out that output quality depends almost entirely on how much context the AI has. Same prompt, different context, completely different result.
Month 8-12: Found a setup where the AI reads my actual manuscript rather than a chat window. Everything before this feels like a different tool.
The learning curve is real and most people quit somewhere in months 1-3 when the generated text disappoints them. The actual value is somewhere else entirely.
[link] [comments]