00:00Well, Mandeep Singh, he is our global head of tech research here at Bloomberg Intelligence.
00:04He knows a thing or two about Meta and, of course, about their AI ambitions.
00:08And Mandeep, you've used this Muse app and you said that it's pretty good,
00:12which is in line with what a lot of folks have been saying, which is why the stock soared yesterday.
00:18Yeah, I mean, this app has got a couple of very good use cases to serve as personal agents,
00:25which is what they are pitching for.
00:27So I'll mention, you know, just if you want discount codes for any brand that you're shopping on the Internet,
00:34like you look for discount codes through Google searches and all that.
00:39Just ask Muse a question and it will give you the best discount codes.
00:43But do the discount codes work?
00:45Because I asked Google the same thing and Google will give me a whole bunch of discount codes.
00:49And when you apply it, it says, nope, not applicable.
00:50It does for the most part.
00:52And you bring up a very good point.
00:53So the challenge for this is, in theory, it scrapes all the best information from the Internet.
01:00But what if, you know, a retailer like Amazon blocks you or it doesn't give you the most up-to
01:05-date information?
01:06And that's where the agent may not work as desired because it doesn't have that latest information,
01:13not because it can't, but because, you know, the retailers who own the inventory don't want it to be integrated,
01:20you know, with what they're doing.
01:23And so in this case, Amazon probably will end up launching their own agent.
01:27But I would expect that agent to work in conjunction with what Meta is doing
01:33because there will be a lot of agent-to-agent traffic and shopping where the human user is completely gone.
01:41It's your agent working with that retailer's agent and doing the transaction.
01:47All right.
01:47I've got a use case.
01:48Now, this will not work with Matt Miller, but it works for me.
01:52I arbitrage Lyft and Uber.
01:55And it really works.
01:57You're not loyal to either.
01:58No, I could care less.
01:58I'm just looking for the best price.
02:00This will do that for me, it sounds like.
02:02It will.
02:03Any sort of price comparisons that involve opening three different apps
02:08and doing a price comparison or three different websites,
02:11this can do it very efficiently.
02:14And not only will it pull the links in terms of what you have to do,
02:18it can actually do the full transaction because they want to integrate payments.
02:23They want to do this end-to-end, an agent working on behalf of the user.
02:28So in theory, this is very powerful.
02:30The challenge is, in the ecosystem, you want the payment rails to be integrated.
02:35You want all the retailers who have that actual supply.
02:39If you're booking travel, Airbnb or booking a hotel that's showing up,
02:44you want this agent to be able to do this end-to-end.
02:48What if Airbnb doesn't support it?
02:50Like Amazon.
02:51So that's where this can fall short.
02:54So it can come up with information for you,
02:56but then you have to execute it on your own.
02:59Or it may not even do that part.
03:01Yeah.
03:01So the agent is very capable of executing it all the way through.
03:05But it just doesn't have access.
03:07It doesn't have access.
03:07Well, that's not helpful.
03:08You know, I mean, if I'm looking for a place to stay,
03:11I want it to look up Airbnb along with other hotels.
03:15Exactly.
03:15So what does it mean for Airbnb and booking.com and all that kind of stuff?
03:19All these companies, the reason why you see Amazon resisting this
03:23or any other platform that is likely to resist is because they will lose direct traffic.
03:28All the, like, look at what you do on Amazon.
03:30You go to the Amazon page.
03:32You spend five minutes looking at different items.
03:34Amazon shows you 10 different ads.
03:37They make so much money when you are searching and showing you sponsored ads.
03:41All that time goes away, and suddenly it's MetaMuse whose agent is doing all the work.
03:48But it's a fragmented market that the agent is working in right now, so that's a problem.
03:52Yes.
03:52All right.
03:53So I'm keeping it on the phone, Mandeep.
03:55I'm all in on this AI stuff.
03:57I'm going to surpass you very quickly, my AI knowledge.
03:59We're not going to need you anymore.
04:01How often do you go online shopping, though, Paul?
04:03Never.
04:04That's what I thought.
04:04I've got people to do that for me.
04:05Mandeep, thank you so much for joining us.
04:07Mandeep, he runs all the tech research for Bloomberg Intelligence,
04:09one of the smartest folks out there.
Comments