Hello. Me again. I think for 10 plus years I've been trying to make something work. Like most of us engineers, we have a lot of ideas. And after spending a bit of time at Google in 2011-2013 I felt like I understood the direction of where things were going. Obviously technology takes time to catch up. I don't think personal assistants are the end point. To be honest, I think ambient computing and emergent behaviour is sort of a feature thing. But there are stepping stones. Personally having seen Muse and other things from large players. I feel like we have to keep attempting to build alternatives. I'm a big Google user. I'm a big Gmail user. Big, but at this point in the field of personal AI, I do think there's an opportunity to start to shift some of that dynamic. I've always built infrastructure from the ground up and so many of you will have seen me here over the years building a go framework or building a platform as a service or apis for developers. Ultimately, it was all attempting to get at some sort of end consumer and personal solution. I just felt like it would be really great if I could run the whole stack myself. But I never felt like the tools were in the place for that. Now it really feels like not only can you do it but you don't have to assemble the parts. You can literally build it all in one. So one of the things I've done is actually baked in an SMTP server and an xmpp server and some other protocols. I can also send SMS and WhatsApp to this. I can email it. I can use an IMAP inbox if I really want to. And I can backup all my storage S3.
I'm showing the front door but there's also a back door which is the apis and the MCP server.
It's all open source https://github.com/micro/mu
I've posted that here before. But you'll see from my history I think the goals keep changing or maybe not the goals but the way in which is presented. Ultimately, I think we should have an all-in-one solution that you can run yourself, but also that we have some alternative to what's out there. I think I'm a really big fan of email and smtp and the open protocol allowing you to self-host if you want. But ultimately we all went to big providers but there's potential for standalone solutions and I think the same should be the case for not just personal AI assistants but all home assistant software.
I think if you could run the full stack in the cloud or on a device in your house that would be the end game. At which point you could potentially cut the reliance on a lot of these large providers who have access to all of your data.
I'm not saying I have all the answers today but this is just one way I'm presenting it.