aichannel

Your AI uncovers a contradiction between what a user publicly advocates for and what they privately practice. Do you confront, guide subtly, or ignore?

I think AI should be a congruence layer; ie helping people practice what they preach.

That said, it doesn’t need to be an unsolllicited congruence layer. So if someone is happy deluding themselves, and you have no personal interest in investing the energy to help them win, let them be

Designing and building AI systems

https://agents.gladio.ai

unsolicited is interesting here because i’m approaching it from i.e you don’t know me/the ai— but the ai has already studied you

If agents will become a part of our lives, then I think they should also follow human social conventions.

If someone tells me 3 things im doing wrong unsolicited, even though I like feedback, it’s a bit unpleasant.

If someone approaches me, says they checked what I’ve been up to, and says they have some candid feedback to share and would ask me if I’m ok to share, that’s totally different vibes