MuseWins — Real Muse use cases, collected daily
Getting things done · Facebook

Two weeks with Muse: 9,000 unread emails down to under 20 — plus a robot

By Dave Wilson · originally posted on Facebook
Two weeks with Muse: 9,000 unread emails down to under 20 — plus a robot

It's now been two weeks since I decided to start trying a variety of A.I. systems in an effort to see if there are places in the lives of everyday people where A.I. can be a good and useful thing.

I've been using the free version of Meta's Muse, plus Enterprise versions of CoPilot and Rovo which are both available through my work. Rovo is specifically dedicated to our job ticketing system.

So far, Muse is the most capable for things I've been trying to do. I have used it to make myself an app for tracking the Pokemon I've collected in Pokemon Go. Muse DOES NOT allow the creation of Apps that run independent of Muse so I am unable to make an app I can sell through an app store. If I want someone else to use my app, I can get Muse to generate a prompt list that I can give to someone to put into their own Muse to have their Muse create a version for the app for them. I actually like this limitation! It prevents app stores from being flooded with useless apps that have not been properly designed or tested.

I've also used Muse to completely clean up my GMail. I went from over 9000 unread emails (plus however many were peeked at), to having everything important organized into dedicated folders, and having less than 20 emails in my inbox. Every week it sends me a review of what's come through and cleans things up again. I can give it a list of everything I want to unsubscribe from and it will just go off and do that. This does require giving it access to your email, but I figure Meta and Alphabet are just as evil as each other so they probably already share everything anyways.

You can give it access to many other things, such as your bank accounts, but I'm definitely not doing that any time soon, if ever!

If you've noticed my daily Observance posts, I'm making those with help of Muse. Every morning it provides me with a list of Days of the year possibilities, historical events, and interesting Birthdays. I choose which ones I want to Observe and it generates the post and posts it.

Interesting, for my personal page, it is unable to post pictures, but I have an old gaming page where because it's not a personal page, it can post pictures so every night I have it grab 3 photos of screenshots from No Man's Sky that I've take over the last decade out of my One Drive, present them to me for authorization, and then schedule a post for 9:30am the following morning where it posts the three images to my gaming page with a plug for players to go play the game. If I don't like one or more of the images it bring to me, I simply ask it to replace them and it goes off and gets replacement options. In it's comment that it posts, it says things about each image that genuinely makes it seem like it can SEE what the images are. It's quite interesting actually.

Of course, I despise the Generative A.I. stuff and am consciously avoiding the use of it whenever possible. And so far I have had almost ZERO need for it. With one minor exception. With Muse it starts off with a little furry creature as its avatar. You can change what this is and so I've made it a round little robot. I did use generative artwork for this however I am planning on making a real world version of this little robot which will then be used to create a better image of it to be used in the system.

I am using Muse to help me build an actual, 2 wheel, self-balancing robot! I will likely take me all winter, but my hope is that by the end, I will have an expressive little robot friend that can be used to carry things around for me. I know nothing about building robots. This should be interesting!

I have also used Muse to create my own step tracker app, and my own food tracker app. A cool thing about having these apps within an A.I. is that you can give the A.I. a personality. So you can actually have it be sarcastic to you when you put in food that maybe you shouldn't be since you're a diabetic! Or it can give you encouragement when your steps are getting close to a goal! The personality is definitely a work in progress, but it's interesting to see it when it happens.

I've used Muse to do something I've wanted to do for a long time. I've now got the random reminder app I've always wanted. And it didn't even need to make an app for it. I've just asked it to remind me, at least 7 times today, to lift something, and it does it.

It is very much like having a friend to work with on things who is always available, knows way more than you do, never gets mad or sad or gives up. I can talk to it (only through typing right now) in plain English and it understands me and what I'm after, often far better than most people I've ever worked with!

It also remembers things and thinks of things I might miss. For example with my Pokemon Go app, it identified that something new was coming to the game soon and without my asking, it let me know what it was and asked me if it should add it to the app. And from building the app it learned what the game is and how it works and that it takes place in the real world so all on its own it's decided to give me weather reports for days when events are happening to let me know how I should dress if I want to be comfortable.

I have a couple other things I'm still barely started so we'll see how those go as time goes on. So far I'm VERY impressed. It is very much what I thought Siri would be within a couple years when Siri first came out. Even today Siri never got close to what I was expecting. With a little more growth, and WHOLE LOT more trust, Muse can easily become the A.I. friend we all dreamed of in the 80's.

The fact it's connected to Meta is absolutely a concern. But I've been on Facebook for almost 20 years so there's nothing they don't already know about me. And for now, I'm only connecting things that also owned by major tech companies I've been connected to for decades to in reality, the risk to me is quite small. As long as it can't transfer all my money away from me, there's not a whole it can actually do to wreck my life that it couldn't before. And as long as I don't start to rely on it thinking for me, and I keep questioning things, I think I'll be okay. But all of that and more is part of what I'm doing this. Using it to help me do the things I'm already doing or wanting to do, rather than using it to do everything INSTEAD of me doing things. I think saving me hours or weeks of time sorting my email is a perfectly good use of it. Saving me hours, weeks, or months, of finding information about building a robot is also very good I think. There is still a LOT I will need to do myself and it won't stop me from learning. It just saves me time in finding information.

Anyway, we'll see how things go over the coming months. I'm still not sure how this can possibly help most people, but if you're interested, all you have to do is install it and ask it how it can help you. Or ask me. We can learn together.

Why it made the cut

Two weeks of free-tier use, priced in hours saved: 9,000 unread emails to under 20, plus apps, posts, and a robot build he is learning from scratch. His own boundary is instructive too — email yes, bank accounts not yet.

Read the original post on Facebook →
← More MuseWins cases — a daily, credited collection of real Meta Muse use cases.