Hey Josh, good idea. Implementing that myself now. One question though - why do you answer today, but have the agent update it's docs tomorrow? Why not just update today?
It's mainly a byproduct of how OpenClaw works. Because each daily question is actually more of a brief conversation, OpenClaw doesn't necessarily know "oh, we're done chatting about this" so it doesn't explicitly know "Okay, now is a good time to save this."
So this functionally lets you return the question throughout the day if you wanted and then doesn't force the save until the next question the next day.
Technically it's always in OpenClaw's session history so it's always saved (you don't risk it completely forgetting). But this just means it won't save it to "permanent" memory until the next or until you say "Go ahead and save this to memory".
Is this agent useful for all tasks (work, projects, etc) ? Like, if you are working on a repo, is this connected to that too or do you keep it for more personal use / research / project ideas?
OTOH I'm thinking of having an "agent" based on my obsidian vault and this workflow could be useful I think, to cover what's not in the notes. Although I'm not using 24/7 agent autonomous (like openclaw) but maybe some chat via opencode for eg. Not sure something more complex is worth at the moment.
btw. Great content so far. I'm happy I'm following this
Franco! So I do very little coding through OpenClaw but I do a TON of both personal queries/tasks and business (especially around brainstorming ideas). For those, hyper-personalization is perfect.
I don't do much around personalization for Claude Code (which is what I use for all heavy dev work). I've added a few paragraphs in my root CLAUDE.md file that helps guide tone a bit, but it matters less to me there so I don't really overthink it.
Hey Josh, good idea. Implementing that myself now. One question though - why do you answer today, but have the agent update it's docs tomorrow? Why not just update today?
It's mainly a byproduct of how OpenClaw works. Because each daily question is actually more of a brief conversation, OpenClaw doesn't necessarily know "oh, we're done chatting about this" so it doesn't explicitly know "Okay, now is a good time to save this."
So this functionally lets you return the question throughout the day if you wanted and then doesn't force the save until the next question the next day.
Technically it's always in OpenClaw's session history so it's always saved (you don't risk it completely forgetting). But this just means it won't save it to "permanent" memory until the next or until you say "Go ahead and save this to memory".
Ah, OK, thanks
Is this agent useful for all tasks (work, projects, etc) ? Like, if you are working on a repo, is this connected to that too or do you keep it for more personal use / research / project ideas?
OTOH I'm thinking of having an "agent" based on my obsidian vault and this workflow could be useful I think, to cover what's not in the notes. Although I'm not using 24/7 agent autonomous (like openclaw) but maybe some chat via opencode for eg. Not sure something more complex is worth at the moment.
btw. Great content so far. I'm happy I'm following this
Franco! So I do very little coding through OpenClaw but I do a TON of both personal queries/tasks and business (especially around brainstorming ideas). For those, hyper-personalization is perfect.
I don't do much around personalization for Claude Code (which is what I use for all heavy dev work). I've added a few paragraphs in my root CLAUDE.md file that helps guide tone a bit, but it matters less to me there so I don't really overthink it.
yes makes sense, thanks!. I'll probably play around with something similar to what you've done in my obsidian vault and start from there