That makes sense. I've thought for a while that having an agent that takes initiative rather than reacting to inputs could be really useful, and I imagine it takes a lot of trial and error to make it take just the right amount of initiative.

So people are hyped because they don't know cron?

Yeah and people were hyped for Dropbox because they did not know rsync and ftp.

Dropbox wasn't given access to your bank account 2fa. There should maybe be slightly more gatekeeping around installing software that unironically advertises itself as RCE: https://docs.openclaw.ai/gateway/security#node-execution-sys...

There is a large amount of gatekeeping called installing and configuring this software. It is not a trivial task that normies can easily accomplish. You have to walk past so many red flags, that you would rightly be called in idiot if you lost anyting of value.

I'll be more concerned for the public when its a double click. Currently it's just a way for techies to fafo. And I do enjoy that there are many people out there messing around with it. It is closer to the 90s experimental net mindset and than I've seen lately. It is also fun that its not a big corpo release. It is not often quick and dirty small team software blows up this big and gets noticed by the world at large.

You forgot cron. rsync without the periodic poll is not a good Dropbox-replacement.

It is combination of:

- Cron-style heartbeat manager

- Easy customization with markdown only

- Good out-of-box memory management that just works

- Good set of tools out-of-box that just work

Like Jack Dorsey said about project success, limit number of details and make those details perfect

Four months ago, I was playing with basically the same framework to explore the idea of "consciousness," using Claude Agent SDK as the harness and Opus 4.5 as the LLM.

I was thinking: wake up every hour, look at some webcams and the weather forecast (senses, change), maybe look at my calendar, maybe read my personal emails for important things, proactively chat with me for work or just fun via email invites.

I played with it for a bit, then got back to "serious work."

I am such an idiot for not seeing the broader value. One thing is that I was sure some multi-billion dollar company was already doing this, and I am super paranoid about the Lethal Trifecta.

Don't worry, you're not an idiot. This is not gonna pan out.

Just like when I discounted cryptocurrency in 2012, yet again I may be overthinking things.

That is the lesson that I would like to share on this platform: If you have an idea, any idea, just do it. Do it now. Build now.

It turns out that there are a lot of morons. Role the dice, you are likely not one of the bag holders, as one of the readers of this comment. I wish I had been younger when I had realized this circle of life.