Thanks! It does help, my misgivings about AI are strictly related to quality control and the potential for claude to fabricate information during benchmarks, as opposed to some sort of reactionary distaste for it. So I was genuinely just curious about your process and how pitfalls are avoided. I do think that there's currently a flood of plugins that are not particularly useful due to AI allowing people to rapidly ship (and charge money for) products that the developers themselves do not really understand, which can be a problem.
But you have put a disclaimer, the plugins are free, and it is clear there is a vision behind these that differentiates them from the frankly low-effort compressors and mojo machines that are being put out lately. Most importantly though it actually sounds good and is clearly standing up to A/B testing on my end.