The new AI proxy, Muse, under Meta, became the focus at the company's annual Connect conference. CEO Mark Zuckerberg clearly stated that the parent company of Facebook plans to push AI functionality to every corner.

In the latest episode of the TechCrunch “Equity” podcast, Kirsten Korosec, Sean O, Kane and I discussed how the AI announcement of Meta could steal the show in the context of OpenAI and Anthropic releasing new models within a week.

Unlike other large AI companies that focus on programming and enterprise tools, Meta has chosen the opposite direction: they have targeted consumers and launched a cute device similar to an electronic pet, known as Tamagotchi. Meta insists that this device is only intended for adults. However, as Kirsten pointed out, this may actually be where Meta's advantages lie.

Sean personally tested Muse. Although he was very pleased that this proxy indeed helped him find the unclaimed funds, he described this feature as more of a “party trick” rather than something that would lead to sustained use. Additionally, there is another issue: whether users will trust Meta to handle sensitive information with AI.

Sean said, "The business of Meta is just selling ads. That's right, they would say that the more they know about you, the more accurate and interesting the ads will be – but wake me up when we're in that feverish dream world."

The following is a preview of our complete conversation, with content edited and polished for clarity and conciseness.

Kirsten Korosec:So, what's your opinion on Muse – this new personal AI agent just released by Meta, which is clearly targeting the consumer market? How does it fit into the situation you just described, especially considering that other leading AI model companies are seeing opportunities and business in the corporate market? Since the Meta Connect conference, their annual major event, has just concluded, and they have already made a full commitment to Muse, that is very clear.

Anthony Ha:This is indeed a bit confusing for me, because what we’ve been talking about is the shift towards the corporate market – not entirely, but obviously both the funds and the attention are heading in that direction.

This is perhaps partly due to the different positions these companies are in – OpenAI and Anthropic are in a leading position in many aspects, but they also plan to go public this year or next year, while OpenAI will do so next year. As a result, there is a sense among outsiders that “I think we really have to start making money now.” It’s not that they aren’t making money already, but because the costs are too high and their valuations are also too high; they must achieve profitability on an unprecedented scale. I believe they see the corporate market as a path to achieving this goal.

I've also been thinking that, for various reasons, Meta might have seen different opportunities. Sometimes I wonder, "Wait a minute, didn't they receive that signal?" But to be more forgiving, you could say, "Well, if both OpenAI and Anthropic are moving in that direction, then Meta might have more opportunities to create a version that is more suitable for consumers, and to continue to advance AI on the consumer side."

Kirsten:What I mean is, we can complain about, criticize, or scrutinize Meta all day long, but they are very adept at integrating themselves into the daily lives of ordinary people, and they have a proven track record of doing so. There’s a reason why Facebook has so many users, and the same goes for Instagram and WhatsApp. In any case, I’ve never regarded them as corporate products. So I think it’s wise for them to continue to focus on the consumer side.

Sean, you have already tried Muse for some time now, and it has been live for a couple of weeks. I would like to know your impression of it, as well as whether you think it has the potential to become a part of various aspects of your life.

Sean O ’ Kane :To be honest, it doesn’t quite seem like it. I understand why some people might think it would be successful. I think many people realize that this is essentially a version created from scratch by Meta, similar to your iPhone or the app on Android. Just like we discussed a few months ago with OpenClaw and Meta, this one basically buys the work of that company and integrates it. This is the first time people have strongly felt: “Oh my god, these agents can really do so much for me while I’m away; I just need to send them a message and let them control my entire computer.” There are many similarities between them. Holding it in your hand and operating it through a chatbot-style interface that performs quite well indeed makes it seem very powerful.

One of the first things I did was —— because it gives you a bunch of suggestions about what it can do, and one of them was “I can help you scan for any unclaimed funds.” I think almost no one would actually think of that, and usually, such things are completely overlooked, because there isn’t much unclaimed property or funds that exist in someone’s name. To my surprise, I actually found some.

It helped me make some money on my first day, which is pretty cool. If it hadn’t prompted me, I wouldn’t have done it at all. And the check that was sent to me is already on its way; that’s great. But it’ll end soon, right? It was a one-time opportunity, nothing that can be repeated. It was more like a party trick really.

Kirsten:What I mean is, you’re essentially disproving your own argument here. I can’t see how this couldn’t become extremely popular.

Sean:A more sustainable version, which is also what Meta has been vigorously promoting in the past few days, is to apply this concept to your daily finances. For example, you can hand over your credit card information, Gmail account, and other details to it, so that it can perform tasks we have seen in other companies, such as Rocket Money-type services: helping you cancel subscriptions you no longer use or identifying duplicate charges. To be honest, these tasks should actually be done by the credit card companies in the first place.

At that point, you will encounter the trust wall of Meta. I find it somewhat surprising that I was willing to explore it and even continued to use it for a while – all the way to today. When I downloaded it, I originally thought it would immediately direct me to services like Threads, Instagram, and Facebook which I hardly ever use, and that it would immediately retrieve that context for me.

But it doesn’t. At first, it treated me like a stranger, which made me even more willing to use it, because I didn’t feel like Meta already had all my information. But as you start using it, it really does try to draw you in, to pull that information into the system, so it can learn more about you.

I don’t know if I’ll be able to trust it in the same way I trust Meta. I think this timing is also quite interesting for me: I just upgraded to iPhone and switched to new iOS and Siri, and this Siri really works, and it can control my phone in a surprising and very helpful way; things it couldn’t do before. Over the past week, I’ve been thinking about how many times I’ve used it, and how willing I would be if Muse had a Siri version to access this information. I trust Apple more to handle such sensitive information—not only from a cybersecurity perspective, but also because its business model isn’t about selling me a bunch of annoying ads.

The business of Meta is simply selling advertisements. That's right, they claim that the more they know about you, the more targeted and interesting the advertisements will be. But wait until we're in that feverish dream before waking me up again.

And besides that one-time profit I got, it was indeed quite good. I don’t feel like I’ve found anything truly useful else – except for the fact, as Anthony mentioned, that it is indeed very consumer-oriented in its interaction, which I think is helpful, and that’s also why people talk about it so much.