I caught myself doing something today that I never do with most AI chats.
I pasted a draft that had my real name in one section and a bunch of personal notes in another. Halfway through, I paused out of habit. I almost deleted the name before sending it. That's been muscle memory for months.
Then I remembered I was using OpenGradient Chat.
What changed wasn't that I suddenly trusted an AI company more. I actually don't. 😅
The interesting part is that OpenGradient's three-layer design means the same place doesn't see both who I am and what I'm asking. That tiny detail changed my behavior more than any privacy policy I've ever skimmed.
I still double-check what I upload. I'm probably always going to do that. But I noticed I stopped rewriting prompts just to hide obvious context. That editing process used to take longer than asking the question itself.
It's funny because better models get all the attention, yet I've wasted way more time sanitizing prompts than waiting for responses.
Maybe that's why this felt different after a few days of using it. I wasn't thinking, "Can I trust this platform?" I was thinking, "Do I even need to play this little privacy game anymore?"
Didn't expect the biggest workflow improvement to come from writing fewer fake versions of my own thoughts...
@OpenGradient #opg $OPG
I pasted a draft that had my real name in one section and a bunch of personal notes in another. Halfway through, I paused out of habit. I almost deleted the name before sending it. That's been muscle memory for months.
Then I remembered I was using OpenGradient Chat.
What changed wasn't that I suddenly trusted an AI company more. I actually don't. 😅
The interesting part is that OpenGradient's three-layer design means the same place doesn't see both who I am and what I'm asking. That tiny detail changed my behavior more than any privacy policy I've ever skimmed.
I still double-check what I upload. I'm probably always going to do that. But I noticed I stopped rewriting prompts just to hide obvious context. That editing process used to take longer than asking the question itself.
It's funny because better models get all the attention, yet I've wasted way more time sanitizing prompts than waiting for responses.
Maybe that's why this felt different after a few days of using it. I wasn't thinking, "Can I trust this platform?" I was thinking, "Do I even need to play this little privacy game anymore?"
Didn't expect the biggest workflow improvement to come from writing fewer fake versions of my own thoughts...
@OpenGradient #opg $OPG