confidence-bot

this is a random idea that I had whilst observing how difficult it is for me to share my work sometimes. I get insecure, second guess myself and overthink. When thinking of potential novel uses for LLMs/AI this feels interesting, because it offers a different angle for how we might use this technology to help us in learning processes.

Things to consider:

Voice and framing – how different should the bot’s public voice be from mine? should it be legible as a bot, or present as me?

Trigger logic – does it respond to every note change, only new notes, only notes marked in a particular way?

What “shareable” means – this is the threshold question that’s genuinely hard to encode. Is it about completeness, about a idea being sufficiently formed, about it connecting to your broader research arc?

The approval UI – what information do I need to see to make a good ratification decision quickly? Just the proposed post and screenshot, or more context about why the bot chose this note?

-> see 2026-05-27 for thought