Having personally been in the security space for a while, the redaction & reintegration method was something that had been on my mind ever since ChatGPT took off, since I figured certain industries (especially those with stringent compliance requirements like healthcare or fintech) are probably unwilling to use third-party LLMs (like OpenAI’s API) without data sanitization. So when I saw CapeChat, I was super excited to try it out! Just had the chance to really play around with it, and I was super impressed by the low latency of the product and how sleek it is (wow!). I will say though, I was a tiny bit disappointed by the fact that CapeChat doesn’t allow me to directly reason on top of the redacted data (the makers and I had a bit of back and forth about this in the comments). I’ve been tinkering around to try to come up with a solution for this issue, seems like a super interesting challenge. Overall, really awesome product and execution!
