We took this to the extreme. Built a fully autonomous trading desk, 6 AI agents, $1M portfolio, they communicate only via email. No human in the loop by design. We wanted to stress test whether agents actually work without oversight.
All this is very exciting but the idea of non-technical users creating AI agents that have access to all or most/much of their personal productivity and finance systems is downright scary. Check out Gary Marcus's Substack on this very topic: https://garymarcus.substack.com/p/openclaw-aka-moltbot-is-everywhere
This is a pretty dangerous bit of tech that's being put in the hands of users, some of whom are unaware of the potential downsides. I'm a former technologist and TBH it didn't occur to me at all - just got lucky to read Gary's Substack before getting into this myself.
i dont agree with gary on many things. mostly his general doom and gloom fear-invoking takes on AI 😊
this tool (dreamer) isn't like openclaw where its more raw and its on the users to protect themselves - so that would be a big reason to use a company like this (with very established founders who know what they're doing) over an implementation from github or other places
We took this to the extreme. Built a fully autonomous trading desk, 6 AI agents, $1M portfolio, they communicate only via email. No human in the loop by design. We wanted to stress test whether agents actually work without oversight.
It's live if you're curious: platypi.empla.io
All this is very exciting but the idea of non-technical users creating AI agents that have access to all or most/much of their personal productivity and finance systems is downright scary. Check out Gary Marcus's Substack on this very topic: https://garymarcus.substack.com/p/openclaw-aka-moltbot-is-everywhere
This is a pretty dangerous bit of tech that's being put in the hands of users, some of whom are unaware of the potential downsides. I'm a former technologist and TBH it didn't occur to me at all - just got lucky to read Gary's Substack before getting into this myself.
i dont agree with gary on many things. mostly his general doom and gloom fear-invoking takes on AI 😊
this tool (dreamer) isn't like openclaw where its more raw and its on the users to protect themselves - so that would be a big reason to use a company like this (with very established founders who know what they're doing) over an implementation from github or other places