We Want Our Agents and We Want them NOW!

Meta has released a personal agent that has skyrocketed to the top of the mobile phone stores!  Muse is here and it is kind of a big deal... and it's kind of cute too.

But what does this mean?  Here is my take on Muse and its implications for the wider agent ecosystem:

1. There is definite demand for AI agents.  We are clearly not afraid of these doomsday scenarios.  We WANT Agents!  And we WANT them NOW!
2. We may start out using occasionally but that will progress to wanting them always on.  So, more than just wanting agents, we want them helping us 24/7/365!  
3. In order to run agents 24/7 we really need to be able to TRUST them and this is where Muse may fall from grace.  As we know LLMs hallucinate and agents can go rogue.  There are ways to increase trust and the two current primary ways to do that are guardrails, sandboxes and harnesses.  All these amount to deterministic constraint where a program is checking the LLM (aka brains of the agent).  

Overall, I consider the Muse craze a positive sign because the more visibility into problems the greater the probability there weaknesses will be fixed.  So, you go get those agents!