#172 Meta Muse: The Cutest Trojan Horse in Tech

Okay.

So first of all, what is Muse? So Meta uses the name Muse for a few things, which is sometimes a little bit confusing.

There is Muse Spark, that's the AI model they launched in April.

Then there is Muse, the personal AI agent, which launched in the US on the 8th of September.

And then there's the Muse Charm, the little keychain.

I think that especially the agent is the interesting part.

Muse is built to actually do stuff for you, so it can open a browser, fill out forms, send emails, shop for you, even negotiate for you, and it keeps working in the background when you close the app.

So a little bit like what Claude or ChatGPT does for you, but much more proactive, I would say.

And it also remembers a lot of things.

So you can use it in its own app on your Mac and right inside WhatsApp.

It went straight to number one in the US App Store.

And Meta says it's already at more than 5 million downloads.

And I have to be honest, from a pure design point of view, I think it's, it's really good work.

Uh, do you remember maybe the Rabbit r1? I think a lot of people have missed this, this launch.

I think it was three years ago.

Back then, it was this tiny game-like device out of orange plastic, which was also supposed to be an agent who could take action for you.

For example, like order an Uber.

And three years ago, this was very revolutionary.

[laughs] I mean, in the end, unfortunately, it failed really badly because of, I think, the design.

I think the design was the biggest problem.

Back then, I was so excited about this little orange box.

I really thought this is a completely new way to talk to technology, and it also was because it was...

Or they promoted it not as a large language model, but as a large action model, so basically as an agent.

But then, as I, as I already said, it arrived, and it didn't quite work.

It...