Meta’s new AI agent Muse made headlines at the company’s annual Connect event. CEO Mark Zuckerberg then revealed that Facebook’s parent company plans to spread AI capabilities everywhere.
On the latest episode of TechCrunch’s Equity podcast, Kirsten Korosec, Sean O’Kane, and I discussed how Meta’s AI announcement looks set to gain traction during a week of new model announcements from OpenAI and Anthropic.
While other big AI companies have focused on coding and enterprise tools, it was a bit of a surprise to see Meta go in the opposite direction with a focus on the consumer and a cute Tamagotchi-like AI device that Meta claims is for adults only. But as Kirsten pointed out, this could play to the strengths of the meta.
Sean tried Muse himself and, while he was happy that his agent actually found unclaimed money, he said the feature didn’t encourage continued use and was more “like a party trick.” Additionally, there is the question of whether users can trust meta AIs that handle sensitive information.
“Meta’s business is selling advertising,” Sean said. “And yes, they will claim that the more they know about you, the more accurate and interesting your ads will be. If you have that fever dream, wake me up.”
Continue reading for a preview of the entire conversation, edited for length and clarity.
Kirsten Korosec: So what do you make of Muse, the new personal AI agent that was just released by Meta and is clearly betting on consumers? How does that fit into what you just described, at least as other frontier AI model companies see opportunity and business within the enterprise? That’s very clear as their big annual event, MetaConnect, just happened and they’re going all-in on Muse.
Anthony Ha: That was definitely very heady for me. Because we felt that what we’ve been talking about is certainly a move to enterprise. It’s not exclusive, but that’s certainly where the money and attention is.
Perhaps some of that is due to the relative positions of these different companies. OpenAI and Anthropic are leading in many ways, but also, in OpenAI’s case, they plan to go public either this year or next. So there’s a sense of, “I think now I have to actually make money.” That’s not to say they aren’t profitable (already), but the costs and valuations are so high that they have to be profitable at this scale, which is essentially unprecedented. And I think they see Enterprise as a way to do that.
And I wonder if Meta is considering other opportunities for various reasons. There’s a part of me that’s like, “Wait, didn’t they get the memo?” But I think you could be more charitable and say, “If that’s where OpenAI and Anthropic are going, then there might be more opportunity for Meta to create a more consumer-friendly version and continue to advance AI on the consumer side.”
Kirsten: I mean, we can complain and criticize and criticize meta all day long, but they’re very good and have an established track record of being embedded in everyday people’s lives. So there’s a reason why Facebook, Instagram, and WhatsApp have so many users. I’ve never thought of them as enterprise products anyway. So I think it’s smart for them to continue pushing consumer-facing works.
Sean, you’ve already been trying out Muse, which has been out for a few weeks now. And I’m wondering how you feel about your impressions and whether you see it succeeding to be a part of every part of your life.
Sean O’Kane: I mean, it’s not. I can understand why some would think that would be the case. As I’m sure many of you understand, this is roughly Meta’s friendly basic version of the OpenClaw app on your iPhone or Android. We talked about this a few months ago. Meta released it and it basically bought and integrated the work of those people. That was the first big explosion. “Oh my god, these agents are going to do all these things for me while I’m out, and they can control my entire computer just by sending me a text message.” There are a lot of the same elements. And when you hold the app in your hands, which functions like a relatively nice chatbot as an interface, it looks pretty powerful.
One of the first things I did with it was because it gives me a lot of suggestions about what it can do. One of them was, “Let’s see if there are any unclaimed funds.” I don’t think anyone has ever really thought about this, and it’s often overlooked completely. Because there aren’t that many unclaimed property funds in your name. Surprise, surprise, there were some for me too.
Thanks to this, I was able to make money from the first day, which made me very happy. I wouldn’t have done it if this hadn’t prompted me to do it. Then a check was sent to me in the mail. wonderful. (But) it will end soon, right? It’s a one-time thing, but it’s not something that can be repeated. It’s more like a party trick.
Kirsten: So you just killed your own argument. I don’t see why it hasn’t become more popular.
Sean: A more sustainable version of that, and what Mehta has talked about a lot in the last few days, is applying that idea to real, real, day-to-day finance. For example, give them your credit card, Gmail account, and all their other information and do what we’ve seen other companies like Rocket Money do, cancel subscriptions you don’t use, identify double charges, and so on, which frankly your credit card company should already be doing.
And at that point, you just hit a wall of trust with the meta. One of the reasons I wanted to explore this and kind of stick with it to this day is that, somewhat shockingly, when I downloaded this, I was really immediately prompted and I was like, I can connect directly to Threads, Instagram, Facebook, and I’ve never actually used it, but I can immediately bring up that context.
But that wasn’t the case. And at first they cooperated with me as if I were a stranger. That made me use it more actively because it didn’t feel like Meta was already leaving everything to me. But once you start using it, you’ll find that it’s trying to capture you and pull that information into your system, allowing it to learn all kinds of information about you.
I’m not sure I can trust Meta in the same way. I think this is interesting timing as I just upgraded my iPhone and started using the new iOS with the new Siri. Siri actually works and allows you to take some controls on your phone in surprising and convenient ways that weren’t possible before. (I’ve) been thinking about how much I’ve used it over the past week, and how much I wish the Siri version of Muse would receive that information. Because not only do I trust Apple more with that really sensitive information, I trust Apple with information from a cybersecurity perspective, but the fact that Apple’s business is not to sell me a bunch of crappy ads.
Meta’s business is selling advertising. And yes, they will claim that the more they know about you, the more accurate and interesting their ads will be — wake me up if you have that fever dream.
And I don’t feel like I’ve found anything else that’s more useful than the one-time money lever I got. Other than the fact that it’s kind of consumer-tailored in its interaction with users, to Anthony’s point, I think that helps, and I think that’s why people are talking about it so much.
If you buy through links in our articles, we may earn a small commission. This does not affect editorial independence.
