AI Agents Are Here. “Brands” Might Be In Danger.

367: AI Agents Are Here. “Brands” Might Be In Danger.

‍

Phillip: [00:00:04] Hello, and welcome to the Future Commerce Podcast, the podcast at the intersection of culture and commerce. I'm Phillip.

Brian: [00:01:49] I'm Brian.

Phillip: [00:01:51] I haven't done that intro in a very long time, Mister Brian, Sir.

Brian: [00:01:55] Yeah. I usually end up doing the intro. That's usually kind of my thing, I guess. I like it, though. Let's go.

Phillip: [00:02:02] It's been a couple months since it was...

Brian: [00:02:04] I'm all in, Phillip.

Phillip: [00:02:05] You're all in. I was all in last week. I joined the global elite at the... {laughter}

Brian: [00:02:14] You rejoined the global elite.

Phillip: [00:02:16] That's true. At a conference that at once I'm, like, proud to be at and also a little bit shy to share publicly that I was at. We'll talk about that.

Brian: [00:02:27] Oh, yeah. You know...

Phillip: [00:02:29] We'll talk about that. They call it the world's biggest podcast. That's not factually true. We'll get into that today. I think out of that conversation, we're going to talk about some AI agents and some growth of that. Maybe touch on some GPT 01 update stuff maybe. There's some news around AI agents and Alterra operating an interesting social experiment within Minecraft, and I want to talk a little bit about those multiplayer dynamics with you, Brian. And I think we'll cover some news in the back half: section 321 updates, some ADA stuff, and a lot of cool stuff coming. But before we get started, Brian, we're heading to Los Angeles. I'm very excited.

Brian: [00:03:15] Yes, we are.

Phillip: [00:03:15] Yeah. Give us a little spiel on the summit.

Brian: [00:03:18] Oh, well, it's going to be the greatest VISIONS Summit of all time. Future Commerce VISIONS Summits are our flagship summit event that we do at a number of different cities. This is our 3rd on the year, prior ones being at South by Southwest down in Austin and then in New York at the MoMA, Celeste Bartos Theatre, which was incredible. This time, we are back in LA with an all new lineup of speakers and content that's just, I honestly can't believe what level of talent we have here. So, Phillip, you want to give a little rundown on a couple of, a sneak peek on a couple of the speakers?

Phillip: [00:04:03] Absolutely. Well, let's kinda just get it started here in that we curate these events specifically to inform people in ecommerce or retail and marketplaces, and anyone that's in digital commerce with people that will inspire you and come from outside of your realm of influence and experience. So you're not going to get the DTC hype boy train at a show like this. This is futurism condensed into a half day event, And as such, we've secured incredible, incredible talent. First of all, our musical guest, performing and opening us up in the show and joining us on stage for conversation is Bad Snacks. Bad Snacks, Jesse. She is a YouTuber and an incredible content creator and was featured in our Muses journal back at the beginning of this year, here at 2024, our annual print journal. Also joining us is Raihan from FWB. Our friend Michael Miraflor will be on stage talking about the change in culture and commerce. Of course, FWB is the premier Web3 turned culture ecosystem. People that celebrate...

Brian: [00:05:17] Is that how you describe it? {laughter}

Phillip: [00:05:19] I think so. I think so. They are definitely post Web3, and they're in that era, and a lot of friends of ours have really joined there. They celebrate this festival on tech and culture every year in Idlewild, and we'll talk a little bit about how entertainment is changing. We've also partnered with Cultique for this particular event, and Cultique, of course, is the culture analyst and research team that services the entertainment industry. Of course, Linda Ong is a genuine legend, and Linda's going to be joining a panel to talk about sort of the aesthetics of horror, and how AI is going to create some new, horrific type of imagery, not necessarily the kind of thing you might think about. This is of a different breed. I'm really interested in how they're going to tackle this, and she'll be joined with Todd, from AILA. And then finally, our big keynote, really excited for this. And there's a bunch of other folks. Really cool friends. Rachel Joy Victor is coming to talk about world-building, and so much more I can't even get into at this moment. But we're going to close out the show with someone who may not be a household name yet, but you'll definitely remember them in the future. Walter, the cofounder and CEO at Shy Kids is coming, and Shy Kids is a production studio that's using AI for short films. And if you saw the Sora AI short film Airheads that went viral a few months ago, that's Walter's team. And so really excited to have some AI video creators come and join us. So lots more to come. Check out VISIONS at our Future Commerce website, FutureCommerce.com/VISIONS, and register while you still can. You'll get 20% off on your ticket price if you're a Future Commerce Plus member. And that's it. Brian, anything else to say about VISIONS?

Brian: [00:07:10] Yeah. Commerce is clearly changing. It's actively shifting.

Phillip: [00:07:15] Yeah. {laughter} Yeah.

Brian: [00:07:16] The way that we shop and buy, I mean, you're going to hear that throughout this podcast, actually. Things are things are moving. And I don't know if everyone remembers, but a few years ago, it was probably like five years ago, it was like, I think we got the technology for the next three to five years in terms of how it was all going to be implemented and play out, and it's going to be kind of three to five years of dust settling on those technologies. And I really feel like we are about to enter a new cycle of implementation of technologies that are genuinely game changing. And so it's going to be another three to five years. We're going to see things that feel completely new if they've already sort of, there's an uncovering of what's available due to technological innovation, and it is going to change things. It's already changing things in culture, and it will change things in commerce. And this particular VISIONS event, I think, is going to highlight some of those changes. And so if you want to understand what the next three to five years or what's going to unfold, catching this event genuinely will make a difference in your understanding, and so I would highly recommend coming.

Phillip: [00:08:28] Alright. Enough about that. Get your butt to Los Angeles, October 10th, at the Nimoy Theatre UCLA campus. FutureCommerce.com/VISIONS. Okay. Wow. Lots of stuff. I was just in LA. I'm going to be going back for our Summit. But what was your perspective from the outside looking in from this All-In Summit? I've taken pages and pages of notes, Brian. I'm ready to talk about it.

Brian: [00:08:54] From the outside, of course, everyone always laughs and complains about the All-In Summit, especially people who are critics of the podcast. And there was the whole thing that Miraflor posted. Michael Miraflor, a friend of Future Commerce Chief Brand Officer over at Hannah Gray, and he had this tweet about how either you're an All-In business nerd or you're an Acquired business nerd, and there's no in between. You can't be both. And he's probably right about that. The Acquired nerds were complaining about All-In. That's what it comes down to. That rivalry is no one knows that it's a rivalry, but it is. I think the other perspective I had was there were some complaints about how nobody really knew what to expect out of that summit.

Phillip: [00:09:47] Yeah.

Brian: [00:09:47] And then people went in.

Phillip: [00:09:49] Oh let's start there.

Brian: [00:09:49] Yeah. People went in, and as the speakers started to come, I'm sure I think they posted a couple days before or something like that. But I was like, oh my gosh. This is insane who they've pulled into this. If this was last minute, that's the greatest scramble of pulling together the speakers of all time. It probably was not actually a scramble.

Phillip: [00:10:16] Could not have been. Yeah.

Brian: [00:10:17] Who knows? Who knows?

Phillip: [00:10:18] I don't know.

Brian: [00:10:18] We will never know. Nobody will ever know.

Phillip: [00:10:21] Who would know? Let me start with saying that I had a wonderful time, and I paid for my own ticket. We do have friends that are in leadership over in the launch side of the house. Longtime friend and mentor of ours, Mike Savino, now is the president over at LAUNCH and works with Jason closely day to day, but you and I had the pleasure of working with him, me for 10 years. You six, maybe?

Brian: [00:10:50] Seven? Six or seven.

Phillip: [00:10:52] Something like that. And yeah. So I got to spend the week with Mike. No better way to do that. And we sat and, you know, he watched me take pages of copious notes, pages and pages of notes. I think one of the things that to start off with is genuinely, there's something that I have said now. I'll say it publicly for the very first time, but I got used to saying at now my second All-In Summit is how many people have confessed only amongst friends that listen to this podcast that they are a little bit embarrassed to share with normal people that this is a podcast that they listen to on a frequent basis only because they believe that the perception of it is that it is sort of an uninformed and powerful megaphone that has a lot of political takes on things. And also, David Sacks has kind of a growing conservative influence in the world.

Brian: [00:11:55] Right.

Phillip: [00:11:56] And so these things are tough to talk about. So I'll just go ahead and get that out there quickly. I will say this. All-In for me was the thing that got me off of nightly CNN and other news and condensed all of the political discourse to 1 hour a week.

Brian: [00:12:15] Oh, man. You're just confirming all of the fears and critics' talking points.

Phillip: [00:12:23] You know, the pandemic was a weird time. I don't know what to tell you.

Brian: [00:12:27] Yeah.

Phillip: [00:12:27] The pandemic was a really funny time, and I think between the Trump 2016 run up and then, you know, going into the pandemic, I had a very specific media diet. And it's amazing what can happen when you turn off the faucet. That's all I'll say. So all political discourse for me has been bucketed to 1 hour a week when this podcast comes out. And that's it. That's all I get. And maybe whatever shows up on Twitter, but I have a really, really effective mute and block list. So I try to keep it. You know, I try to keep the anger contained, Brian Lange. So there you go. Let me start with saying that I think the topic of most of our conversation will probably be on Marc Benioff's chat, and probably only one part of what Marc Benioff's chat was. But let me run down a little bit of the speaker list. Alright? In no particular order. So we had Sergey Brin dropped by, Co-Founder of Google.

Brian: [00:13:29] Not a small name or anything. {laughter}

Phillip: [00:13:32] {laughter} We had Peter Thiel, one of the PayPal 4 and Chairman of the Founders Fund and very outspoken, also very conservative donor, although he says he's sitting out this election. We had Bari Weiss, who you may not know her but she created the Free Press. And Free Press is sort of like a Substack movement and spin off a bunch of journalists from New York Times. We had Kyrsten Sinema, which I thought was a really interesting quasi-independent voice that's maybe not so independent these days depending on which side of the aisle you're sitting on. {laughter} So that was sort of the real... I think that was the actual program for Day 1. We had a bunch of others too. We had Megyn Kelly showed up, and Elon closed out the day. So real up and down on the political discourse, all of Day 1.

Brian: [00:14:29] That was all Day 1?

Phillip: [00:14:30] All of Day 1. But what they hadn't prepared anybody for was that you would be standing in line in a 100 degree heat at 8 a.m. In L.A. during a massive heat wave and ready to be frisked by the TSA because they didn't tell anyone that JD Vance, the Vice Presidential candidate for United States, was going to open the show.

Brian: [00:14:55] And that definitely spurred some complaints, not necessarily because of Vance, but because of that heat.

Phillip: [00:15:03] I mean, yeah, the heat was... The heat was wild. Let me give you some top line notes for Day 1. Alright? I think so Bari Weiss mentioned audience capture, which is something that I think is really interesting because it comes in the light of...

Brian: [00:15:22] She's reading Matt's zines.

Phillip: [00:15:24] Yeah. So Matt Klein, friend of the pod, has this zine called Audience Capture, which talks about the phenomena. How would you describe it, Brian?

Brian: [00:15:35] It's kind of hyperstation. It's like how audiences and creators are interacting and the changing nature of that relationship and how creators are being controlled by their audiences in many ways and are sort of caught up in what their audiences want them to do, and it's very multiplayer sort of angle to it. Matt, of course, is like, "I don't know if this is a good thing. It could be a bad thing." He said that at VISIONS Summit from last year,

Phillip: [00:16:11] Which was 2023.

Brian: [00:16:12] There are dangers. There are benefits. Things are changing now. That's clear. And audiences have more power over what's happening than ever before. Do you want eyes on you? Do you want to be under that control? That is the question.

Phillip: [00:16:27] There's Nikocado. Right?

Brian: [00:17:16] Mmm hmm. A big, big shift for him. He found his way out. {laughter}

Phillip: [00:17:22] Yeah. That's an interesting aside that I was going to say. So when Bari Weiss was talking about audience capture, sub stackers versus the New York Times, she was saying it under the guise of this idea that having a social interaction with your audience allows you to get outside of the cabal. So she was framing it as you're leaving one evil and running into, you know, some sort of closer to the people is more salvific. It's like I'm leaving Washington groupthink and this crazy, crazy town, and I'm going to turn on something else that's more powerful, which is, you know, what the people really need and want. They can talk right back to me.

Brian: [00:18:13] Right. Which is, there's some truth to that. When you have a direct relationship with your audience, sometimes you could have a voice that you did not have before. However, did she get into the dangers of that as well? Because there are pitfalls to that.

Phillip: [00:18:31] Yeah. I mean, she did talk about this idea that the media has a role and a responsibility because the power of the media can shape messages and shape people's thinking that they should treat that with some responsibility, but that's kind of out the window. So her thinking was, you know, media is priming people for revolution.

Brian: [00:18:58] Had she caught up on the most recent episode? Maybe it's not the most recent anymore of Only Murders in the Building because...

Phillip: [00:19:07] I don't she didn't mention it. {laughter}

Brian: [00:19:10] Did you watch the most recent?

Phillip: [00:19:13] No. I'm not caught up. I'm still early in.

Brian: [00:19:15] I won't give anything away. Small small small spoiler.

Phillip: [00:19:19] Mhmm.

Brian: [00:19:20] Or maybe this is too much of a spoiler for you.

Phillip: [00:19:23] Uh, maybe.

Brian: [00:19:24] There's a character who they believe is obsessed with Christmas because all the stuff throughout his whole apartment is Christmas-related. Well, they sort of interrogate him because they're trying to find out who killed who and so on.

Phillip: [00:19:42] As they are want to do. Still haven't left a building, by the way.

Brian: [00:19:45] They found out that he actually hates Christmas. He's a fitness influencer who happened to post a Christmas fitness thing, and it took off. It hit. It landed.

Phillip: [00:19:56] This is his life now.

Brian: [00:19:56] Yeah, the only thing that hits for him is Christmas-related. And if anyone ever found out that he actually hated Christmas, he would lose all of his followers. {laughter}

Phillip: [00:20:07] That is so on the money.

Brian: [00:20:10] It is.

Phillip: [00:20:11] So she mentioned something. Audience capture made me sit up. She mentioned archetypes. By the way, the word archetype was said at least 15 times. And this is a fairly, this is like a new emergent piece of language from something I think that we... Nobody was saying archetype a couple years ago. I'm not saying we brought it back or anything, but when we published our first journal, talking about archetypes felt like talking about an aged concept. But her main gripe was about prestige, and I really identified with this because while she was talking, the rest of the panel was sort of talking about prestigious institutions of any kind, including the New York Times and Ivy League Schools, she was talking about the archetype needs to shift away from thinking of equality as being equality of outcome back to thinking about equality as an archetype of the equality of opportunity. Now this could get really political, but I think that, generally, we have dynamics that happen to fall in this equality of opportunity in the way that people see ecommerce. I'm using air quotes for people not watching on YouTube right now. Ecommerce has become a system that people have thought of as anybody can start an ecommerce store because anybody can...

Brian: [00:21:50] It's democratized.

Phillip: [00:21:51] It's democratized. Yes.

Brian: [00:21:52] That's the word. Yes.

Phillip: [00:21:53] That's the word.

Brian: [00:21:54] The word.

Phillip: [00:21:56] And I think when you say audience capture archetypes and you say, you know, democratization in one sentence, I find that there's direct ties back to a lot of the conversation around the direct to consumer movement and how much of that discourse has died off, which makes me wonder, Brian, and I'll leave it at this, and we can move on. But [00:22:21] once the hype cycle leaves a conversation, that's when the quiet building begins, and you typically start to see the fruits a few years later of something that was actually talked about and was overhyped at a given point in time. [00:22:33] So I wonder if we're not a couple few years away from really seeing what bore out of the actual direct to consumer movement.

Brian: [00:22:42] Yeah. Direct to consumer politicians. Is that what you're getting at right now?

Phillip: [00:22:45] I don't know. Direct to consumer... I, you know, I saw Sean Frank and Jason Panzer at the show. They're both on The 9Operator's podcast, and they're both DTC poster children at this point. By the way, Sean is probably one of the nicest guys ever. He's very, very complimentary, introduced you to everybody, knows exactly who you are, talks very nicely about you and very encouraging. I think that that doesn't fit the archetype of somebody that I tend to see as a kind of an Internet troll.

Brian: [00:23:19] Yeah. Like a very strong voice. A strong voice on the web.

Phillip: [00:23:23] Specific acquired voice on the Internet.

Brian: [00:23:25] So true though. Oftentimes, the person that you think is a certain way on the Internet when you meet them that they're actually a different person. It is due to the the limited nature of interaction on the web. Right? There's only so much you can do, and I do believe, I think there are a lot of people out there who are like, "I've met some of my best friends on the Internet," and, "that's not true, Brian." You can really get to know someone, and if you don't know how to, it just shows you don't know how to operate the Internet if you feel like you can't become close with someone through the Internet. And, actually, there's some truth to that. I do believe that there are people who are better at bringing their full selves to the Internet than others, but sometimes that stuff doesn't land. So it's interesting to see how some people will just appear very different. Also, sometimes people have extremely active minds, but don't necessarily have active mouths. And it's easier to get that out into the spoken word.

Phillip: [00:24:38] That's interesting. Yeah.

Brian: [00:24:42] So, yeah, little tangent there. But I would love to come back to the rest of your time and All-In because I think really where we want to go with this is the agent idea.

Phillip: [00:24:55] Yeah.

Brian: [00:24:56] Yeah. So just to check a couple of boxes

Phillip: [00:24:59] Yeah. Yeah. Bari Weiss kinda finished off with this idea of... I'm, like, going through every note here. I should probably...

Brian: [00:25:04] Yeah. No. Keep going. This is great. This is what I wanted.

Phillip: [00:25:06] So she was talking about the idea that she's, like, really small. In terms of sub stack, she's probably very large. In terms of the reaches she had with at the New York Times, very small. And then she sort of flies below the radar. Right? And I think, so she said she's stealthily building an institution, and that is the mandate of people in this era who want to oppose the larger incumbents is that you're not going to get there with a full frontal attack with lots of capital. We've seen that that doesn't work. So the only way to oppose the existing systems of influence, and to avoid what she called ideological capture, which I think there's a lot of that in ecommerce and retail, is that we have to stealthily build an institution. And I kind of identified with that because that's what I feel like I would love this to be. I would love Future Commerce to be an institution that is building quietly and steadily and stealthily over a decade. So I really identified with that. I'm sure others could identify with it too. Yeah.

Brian: [00:26:21] Yeah. No question. I think that was that's super interesting. I will have to see if I can find that talk, and listen to it. That sounds super interesting.

Phillip: [00:26:31] It's a very good one. Listen... There was a lot of incendiary stuff that got reshared. I mean, Elon was like a total crackpot on stage as everybody wants him to be. He was sort of egged on into telling some stories. Sergey Brin didn't have a ton to say. I was really blown away... So the CEO of of Adyen came on and spoke his first public event that he did in the United States, talking about stable coins and sorta had a similar... I mean, mostly, he was talking about the aged nature of the Swiss system and how they find extra dollars and create margin by using AI to find more efficient routing so that they don't have to pay fees unnecessarily, which is the old traditional way of doing it. But he said there's a future where adopting and partnering with an existing stablecoin could actually end around all of that, and you don't have to worry about any of the inefficiencies of the old Swiss system, which I think is what a payments futurist should be thinking about.

Brian: [00:30:02] Totally. Yeah. And I think he's not the only one sort of thinking that way.

Phillip: [00:30:07] I don't think so either, but I think from a regulatory perspective, it feels a long way off depending on how this election goes.

Brian: [00:30:16] Yeah.

Phillip: [00:30:16] Although it's really interesting to see both parties kinda courting the crypto crowd, which I think is wild.

Brian: [00:30:23] Yeah.

Phillip: [00:30:23] I mean, Trump sold... He's selling crypto literally right now.

Brian: [00:30:28] Yeah. You almost bought his NFT. Right? {laughter}

Phillip: [00:30:29] No. Hold on. {laughter} I said, should I buy this so we can make content about it? It's not the same thing. {laughter} Don't do that to me. Okay. I do want to fast forward. So Day 2 was where I want to spend most of this time, but just a quick overview too. Nikesh from Palo Alto Networks was talking a lot about cyber risk and data security, and very specifically around the change of the nature of how young people are... We're almost at a place now where we should probably be concerned around the next levels, the future of cyberattacks. So if social engineering is the threat vector for most cyberattacks in the world and kind of always has been, what kind of social engineering could happen now or in the future when you're not actually sure if you're texting with your mom? You're not actually sure if you're actually talking to your mom or not.

Brian: [00:31:47] I really think there needs to be, and this is going to be, this is probably something that people have tried to solve for. I'm not saying that people haven't attempted this, and you might be thinking of something immediately, so I'm hedging what I'm saying right now. But it seems like there needs to be a system, a new system for communication. I don't care what the rails are. Obviously, as secure as possible, that is literally, it is only certain people that can communicate with you, and I know you can set up lists around this and things like that, but the problem is devices and systems, you usually end up getting other things to those systems and to those devices, and there's always sort of a way to spoof things. And so it feels like there needs to be potentially like a device communication combo that is built for things like medical and key contacts that is approved on both sides, and there's a very official way to make it happen. And you could just say, well, you can just do this with email. You can just do this with whatever, texting or whatever. But it feels like everyone is so overrun on all of their normal communication systems that it would take almost a complete reset for this to happen. And, yeah, I don't know. I don't know. I would use it if I could have something that was a 100% for my bills and for my medical communications, and that was the only thing that I used it for, I would use it.

Phillip: [00:33:36] I'm taking a mental trip back to 1998 when I was using email for the first time thoroughly. I had email before that, but this was how I was communicating with a lot of people.

Brian: [00:33:53] Your main source. Right.

Phillip: [00:33:54] I was emailing lots of people for lots of things by 1998. And there was a thing that became really popular with people that were very security minded at the time, really concerned about sorts of things, like what we're talking about, like verification authenticity of who you're talking to is actually the person you think you're talking to. And there emerged a privacy scheme called PGP back then. Pretty Good Privacy is what it was called. And it was a really lo fi way of verifying that the person who sent you the email was actually that person because what you would do is you would put your PGP key in your email signature, and only the correct PGP key could decipher messages from you. And, eventually, this was built into email clients. So you didn't have to do anything extra. You didn't have to verify anything, although it was sent in the clear, and you could see it at the time. I had my email signature at the time, had my PGP key in it. That was a thing that people did, sort of a signal to people. I think, you know, today, it's relegated to our CS encrypted chat on your WhatsApp.

Brian: [00:35:13] Yes. I think you're right.

Phillip: [00:35:14] I think we need...

Brian: [00:35:15] That's exactly right.

Phillip: [00:35:15] We're going to need a bigger box. You're going to need a bigger lock sign to show that this is a concern in the future.

Brian: [00:35:15] Totally.

Phillip: [00:35:24] Yeah.

Brian: [00:35:24] It's almost like, you know, when Target first launched their marketplace, it was invite only, and now it's not obviously. But data right now is not invite only. Communication is not invite only. It is everybody gets everything from everyone, and there's no way to really have separate buckets. And so, anyway, it just feels like there needs to be approved vendors involved. It's almost like a payment system could take this on, like a Fintech company could take this on, actually. Devices for medical and bills only or other key messages, and that only approved vendors are allowed to be involved. And if you ever are a bad actor, you get kicked out, and there's something that's so intense about it that, you know yeah. I don't know. I feel like, so I agree. {laughter} All this to say, I agree with the talk. Security, cybersecurity, whatever you want to call it. [00:36:36] Security right now, digital security is probably the biggest concern and all of the benefits of a lot of the tech that's being released right now could just be completely washed out until there's a way through with security. [00:36:52]

Phillip: [00:36:53] Yeah. I do think that we don't know what we don't know yet. There's a lot coming. Um, we'll come to Benioff. I think we'll finish out. I'll just give you one more little top line on something. Across all the talks, I caught a couple recurrent themes. Here's one.

Brian: [00:37:15] The apocalypse. That was one of the themes.

Phillip: [00:37:17] That was a theme, but we won't get there. {laughter} Maybe we can talk about that in an After Dark. There was a theme about where the cities of the future will be. I heard this from a couple different speakers, but I've sorta aggregated it here for you. So starting in the United States, a lot of excitement around Texas and Florida, but in particular, Tampa, Miami, Austin, Dallas. There are apparently a lot of folks talking about those as being sort of hubs, mega hubs of the future. I believe that. Miami, I believe that. Dallas seems interesting to me. Austin, for sure. Okay. But the cities of the future will not be in the United States. According to more than one panelist, where we need to be looking is Riyadh, Dubai, Singapore, and Abu Dhabi. Look at what's happening in commerce there. Look what's happening in capital allocation, and look what's happening in infrastructure and in emergent workforce. And, apparently, those are... And public accommodation, transit, infrastructure, that sort of thing. Really interesting stuff. Having taken that note, I kinda want to do, we did a Milan series, like, shopping in Milan, which was sort of looking at the nostalgic city. I kinda want to do a shopping series in 2025 on each of these cities. Like, how to shop in Abu Dhabi, how to shop in Dubai.

Brian: [00:38:56] Let's shop in Singapore. Let's do it. Let's do a little Future Commerce trip in Singapore. Let's do a in Singapore. Do a Future Commerce retreat. Let's take a few people with us. That'd be fun.

Phillip: [00:39:05] Let's just plan it. What do you think? Like, a September or October 25?

Brian: [00:39:09] I think that sounds great.

Phillip: [00:39:10] Singapore?

Brian: [00:39:11] Yeah. Let's do it. You heard it here first.

Phillip: [00:39:15] Let's just do it. Alright. If you're listening...

Brian: [00:39:17] Hit us up.

Phillip: [00:39:17] We'll invite Future Commerce Plus members first. Come with us. If you're interested, drop us a line at Hello@FutureCommerce.com. I can hear the team groaning right now.

Brian: [00:39:29] Right. A hundred percnt.

Phillip: [00:39:30] Yeah. Although, they love doing big things with us. Okay.

Brian: [00:39:35] Okay. My one little caveat to that is, and I will, actually two. Two little things I'm going to say about this. First of all, you've said something that I think, you said it in specific context, which was livestream shopping. What happens in other parts of the world doesn't always translate to the US.

Phillip: [00:40:01] Right.

Brian: [00:40:01] And so I do think there are benefits of looking in other places because there are things that will come over, and there are things worth copying and finding ways to implement. A hundred percent. But just because it worked in another city somewhere doesn't necessarily mean the US market in particular is going to adopt it. Doesn't mean there isn't money to be made taking it to other places in the world. I think that, you know, without we don't want to be too US centric, but I do think that it's important to remember. And then the second thing is I believe that while urban environments are going to continue to be a thing, I wrote an addendum or response to Marc Andreessen's, "It's time to build," however long ago he wrote that.

Phillip: [00:41:00] Five years ago. Six years ago. Something like that.

Brian: [00:41:00] Yeah. And something that I made a point about, and I still firmly believe this to be true, and actually, of course, the more that I read McLuhan, the more that I believe it's true. There is going to be continued innovation in places that we have never thought of before. The opportunity with electronic communication is to do things from cheap places, and urban places are usually not cheap places.

Phillip: [00:41:29] Yeah. I believe that.

Brian: [00:41:31] There is opportunity for smaller, more focused communities to be built in places where the climate is better, first of all, because we all know that Miami's climate's not necessarily going to be getting better.

Phillip: [00:41:48] Miami, Dallas, Austin, and Tampa are not.

Brian: [00:41:52] The worst possible places to be building in the US from a climate perspective.

Phillip: [00:41:59] I don't know that there is a better... I think it's all bad in some way.

Brian: [00:42:03] Right. Right. Exactly.

Phillip: [00:42:06] You're either in tornado alley or earthquake alley. Actually, Pacific Northwest is just totally unaffordable, but, you know.

Brian: [00:42:13] It is. No.

Phillip: [00:42:14] There's that. Totally unaffordable. You got the gloom.

Brian: [00:42:17] Got the gloom. You got the gloom. I've got a whole thing around this. You want to hear my whole thing around this? So if you wanted just a really quick thing about Seattle weather, here's my way to get through the Seattle weather. First of all, summer is the best there is. Summer in Seattle area, the Pacific Northwest is unparalleled. There is no better place to summer than in the Pacific Northwest. So you come out of that, and then you have...

Phillip: [00:42:45] I agree with this, by the way.

Brian: [00:42:46] Okay. Awesome. So step 1. That's step 1. Step 2, you get to the Fall, and the Fall in the Pacific Northwest is one of the most underrated seasons in the history of seasons. It's actually less gloomy.

Phillip: [00:43:00] I history of all seasons.

Brian: [00:43:01] In the history of all seasons in all places, it is the most underrated.

Phillip: [00:43:05] Okay. Yes.

Brian: [00:43:08] It is not as gloomy as people think it is. It is super temperate. The sun's often out way more than it is and extends all the way into October. It's gorgeous. The leaves are changing. It's absolutely beautiful. It gets chilly at night, but the days are really nice. And so I love it up through about the end of October. Of course, at the end of October, you start to hit holiday, and that's really fun. It's really cozy, and, yes, it starts to get gloomy and blustery and cold and stormy and whatever, but it's really fun, and you have a good time with your friends and family. And then when you get done with that, ski season's just kicking off, and you can ski right on through March. And then you really, basically, you have April to contend with. Basically, the end of March, April, and you can ski into those seasons. But right around March April, if you can gut your way through the end of March through April, and somewhere in the middle of spring, it starts to get nice all over again and the flowers are out, and it's gorgeous all over again. And so there's really, there's about a month where you will get really depressed, and so you just have to get through that. And then the rest of the year is unbelievable. There you go. Seattle in an upsell. And that's a direct response to a lot of hate I've seen recently on Seattle as a city and the Pacific Northwest as a climate. So all you haters can just deal with it.

Phillip: [00:44:31] I don't know. Sure. That sounds great.

Brian: [00:44:35] There you go.

Phillip: [00:44:35] I love that. But what do I know? We have hurricanes and climate change is probably going to flood us all. So I'll be up there with you soon enough.

Brian: [00:44:45] You heard my rant.

Phillip: [00:44:47] You do have a 400 megaton nuclear bomb of a volcano that will go off eventually one day.

Brian: [00:44:52] We do. Yeah. Eventually, it's going to go off at some point. But, I mean, that's I mean, y'all have hurricanes and the middle of America has tornadoes. So...

Phillip: [00:45:01] I see them coming for, like, 8 days. {laughter} I can get out of the way.

Brian: [00:45:04] Which should technically be true for a volcano as well. You should see signs. There should be some wind up. A wind up. Yeah. Hopefully. Yeah.

Phillip: [00:45:13] We'll see. I love that. Okay. Let's get to the real conversation. I want to set it up. Okay?

Brian: [00:45:20] Yeah. Let's do it. Let's do it.

Phillip: [00:45:22] Okay. I think the thing that we should be talking about is AI Agents, what we have now, what we can expect in the future, benefits, some challenges potentially. Always going to have some times of correction. Right? There's the overinflated expectations. There's the reality of what actually it will be. So we will have a long conversation, but the way we'll get into it is we'll be talking about it. I think the best way to talk about it is for me to read a tweet that I've posted just now about my reaction to [00:45:57] Marc Benioff, Salesforce CEO. Basically, he's all in on Agents. He says their AIA models are going to be a PhD level reasoning eventually. Very soon agents will have the ability to use its tools, meaning the model can now pretend to be a human. Right? And my response to that was, will my AI Agents also incorrectly tag deals, fake follow ups, forget to set their out of office, argue over deal attribution, and insist that a deal is "close," and "We will definitely close by end of quarter," only to kick it out to fiscal '25 on the last day of Q3? {laughter} [00:46:36]

Brian: [00:46:36]  [00:46:36]{laughter} There you go. [00:46:37]

Phillip: [00:46:38]  [00:46:38]So, AI Agents... Just other kinds of humans. Because that's how they've been trained, and that's the data they've been trained on. [00:46:47]

Brian: [00:46:47] Technology reflects back humanity.

Phillip: [00:46:51] Exactly. It's just holding them a mirror. Right? It's narcissists if nothing else. Okay. So Marc Benioff takes the stage with David Sacks. A lot of rumor mill action that came out of this one. Number one, lots of people said they did not like the way that this interview went. That's just how people went. That was the scuttlebutt. I think part of it was Marc wanted to hit some talking points. He had something to say. He had something to say and it was earnings season. It's like we have a job to do. I heard someone say afterwards that Marc used to go and speak at another event, and I'll definitely rat the person out if I say which event it was. But Marc would bring his own interviewer with him. So Marc does not like a free form interview that someone else is in control of, and that was definitely evident here because Mister Benioff basically took it over and had his own interview with himself.

Brian: [00:47:57] Got it. Basically, it was, you know what, David? I'm the captain now.

Phillip: [00:48:04] Yeah. I'm the captain now. Actually, we'll link up the video of this interview in the show notes. You can just see it for yourself. But, basically, he does spend a lot of time talking about the future of AI at Salesforce. Now let's be really clear here. Salesforce has been talking about AI with a branded product called Einstein since about 2017, 2018 maybe.

Brian: [00:48:28] Somewhere around there.

Phillip: [00:48:29] Yeah. Since 2018. Machine learning.

Brian: [00:48:32] It was a Watson competitor. Yeah.

Phillip: [00:48:38] Sure. Went into all their products. They had sort of this idea that machine learning would, you know, generate all kinds of new features. And for the most part, I think that's probably true, but we haven't really we'd never seen generative AI until a couple years ago. So the way that AI was deployed was typically in insights, analytics, naming of things, like, really light touch consumer facing type stuff or business product consumer facing type stuff.

Brian: [00:49:08] Even, like, business users would leverage AI, but it would be using similar means that they had been using before.

Phillip: [00:49:18] Yeah. Oh, for sure.

Brian: [00:49:19] It was like software was better equipped as a result of AI that they were already using.

Phillip: [00:49:27] A hundred percent. And I don't discount that the machine learning they were doing was very real, and I'm not calling any of that into question. I do think that everybody right now who's professing to have some sort of leap forward in AI capability has either... There are only two options here. You have either spent a phenomenal amount of money with NVIDIA and bought a bunch of H100s yourself, and you've built your own data center that can do the same level of compute that is needed to keep up with all the other large model. Oh, also, you'll need a very large model, and that model either came from Facebook

Brian: [00:50:06] Yep.

Phillip: [00:50:06] Right, as Llama or some open source model.

Brian: [00:50:09] Yep.

Phillip: [00:50:10] So that's one path. The other path is you are a thin wrapper for OpenAI or Anthropics Claude, And that's pretty much it. Those are the two paths. Choose one.

Brian: [00:50:21] Amazon picked Claude recently.

Phillip: [00:50:23] I think it's interesting that any future capability that's being professed here, I don't think Salesforce is... My light went off again. Dang it. I gotta charge these things. Well, now I'm sitting in the dark on the YouTube, but you're in this far. So here we are. So when Benioff is talking about the future of AI agents, I don't see Salesforce as being an AI first company on the par of these others. They are finding novel ways of plugging in those AI capabilities, which are untapped, unexplored. We've only scratched the surface. You can do all kinds of things. But that doesn't mean that anyone else couldn't. They do have a corpus of data that no one else has. So that's the first thing.

Brian: [00:51:10] Well, isn't that what Matthew McConaughey said in their commercial?

Phillip: [00:51:15] Oh, yeah. {laughter} What did he say?

Brian: [00:51:17] He literally said something to the effect of you know, "if AI is the future, does that make data the new gold?" He effectively said AI is the way forward, but data is how you make the money.

Phillip: [00:51:33] Wow.

Brian: [00:51:34] Because you have to have the data in order for AI to actually do anything. And so maybe Benioff.

Phillip: [00:51:41] That's right because it was a great data heist thing. Right?

Brian: [00:51:46] Yeah. Benioff agrees with you. That's what this is here. We're on the same page here. Everyone's saying the same thing. Yeah.

Phillip: [00:51:52] We're all vibing here. Do you remember Oliver and Company, the Walt Disney film?

Brian: [00:52:01] You're taking me back, man. That's reaching deep into the vault.

Phillip: [00:52:05] There's a scene with a chihuahua.

Brian: [00:52:08] Yes.

Phillip: [00:52:08] And he's got a cigar, and he says, "If this is torture, chain me to the wall." So when Benioff is talking about AI agents, he's talking about a thing that would have been called automation in the past. Okay? But now we have this capability of free text and, like, textual inference. So we can actually tell it what we want. Right? And theoretically, it would be able to do some reasoning. Let's say that they're plugged into OpenAI, and they have access to the new 01 model. It can do some advanced reasoning. It could say, "Here's the series of tasks that I would need to go through in order to have an end goal." So we can speak with goals in mind. Maybe one of those goals is a marketing campaign that generates a certain level of return or moves some sales, you know, some deals through a pipeline. This is a thing we've talked about in a sales capacity in the past, Brian, in prior roles.

Brian: [00:53:12] Yeah.

Phillip: [00:53:12] We need to weed out the what's real and what's not here, send a breakup email. If you could tell your agent to do that and you already had all that stuff canned and you had to make a million decisions. If you didn't, maybe that's a really, really great thing. I see that being super valuable.

Brian: [00:53:26] I could see that being super valuable if it was good. If it was good at it. Right? Because it's like any algorithm out there, there's a whole set of risk, risk tech companies that aree like, "You're getting too many declines as a result of your risk technology. Use us instead. You'll have fewer declines and there's a lot of legitimate things that you'll get as a result," or you won't have to deal with as many declines. And I think that's probably what we'll see here in the sales side is it's like, "Hey. Here's a whole pile of things that we think are disqualified. Do you want to review these? Here's the ones we have at this confidence level, x x x," whatever. I don't know. Yeah. And then you have to go through and then be like, yes, no, yes, no. That's probably at some level where we're going to sit and the better model will come along and say, "Actually, there's more here going on, and you should pay attention to this."

Phillip: [00:54:31] Well, there's also something that you have to piecemeal together today in the sales process. I know we're hung up on the CRM component of Salesforce. They do a lot of stuff. But, you know, if there's an AI Agent today, theoretically, he's talking about it could realistically pose as a human. There's a CX component here that's being thrown around sort of willy nilly. I think that's the chatbot future that everybody really fears is having really crappy chatbots that you have to sort of chat with, negotiate with, or any kind of thing that's posing as a not as a human. I think kinda goes back to what we were talking about earlier. It's a little ick for me. Marketing messages are one thing.

Brian: [00:55:19] Yeah. Even that's pretty ick. I don't know. It's funny. Anytime anyone finds out something that they didn't realize was actually AI, they didn't realize it was AI, and then they find out that it was, everyone's very quickly turned off by the outcome. There's been a bunch of tweets about how "Oh, wow. When I see a work of art and I find out that it was made by AI, I completely lose interest in it." And it's almost like any "deal" that's made like a negotiating bot. You're like, "This is garbage. Why didn't you just get to the end state? Why did I have to go through this whole process in the first place?"

Phillip: [00:56:01] That is where we're going. Pause on that for one second because here's what else happened when Benioff took the stage. Benioff took the stage 10 days after a report from Klarna's CEO, which was the third, I believe, report of its kind from Klarna in the past year talking about how they're gaining efficiency in the business by employing AI at a greater scale. So last go around, at the last earnings call where they touted AI replacing basically contractors and some internal design resources and third party libraries that they spend money on for things like generative AI video and image assets, they said they're they were offsetting marketing costs by displacing people in their marketing organization by using AI liberally. So that was the last earnings call. Klarna, 10 days ago, had a report, 10 days prior to Benioff taking the stage at the All-In Summit. And they said that there were two announcements. One, they had replaced Salesforce within a homegrown CRM that was built with AI entirely internally, and they were also replacing another software suite called Workday, which I believe is PM and sort of employee management, PEO type stuff.

Brian: [00:57:24] Yup. Workday has been advertising like crazy lately.

Phillip: [00:57:27] Yeah. Well, for sure. And these two pieces of software are DOA at Klarna. And then on the back half of the conversation, in the press release, it's also, "And, also, we're laying off 50% of our staff because we're so efficient now with AI." Some of these things seem real, some don't, but them getting rid of Salesforce seems to be very legitimate. And this is the context in which Benioff is having this conversation.

Brian: [00:57:53] I just may or may not have been at a dinner not too long ago where the where the inside word was that he had always hated Salesforce.

Phillip: [00:58:03] Yeah.

Brian: [00:58:04] He had actually made the decision to get rid of it years ago. It was being phased out for a long time, and it was a great announcement to make in that moment.

Phillip: [00:58:18] For sure.

Brian: [00:58:19] But it had already had been sort of pulled out of the business. And whether or not AI was involved in getting them to that final point is a question.

Phillip: [00:58:30] For sure. Yeah. Yeah. And basically, it's like, so Benioff in a totally different interview, sort of questions this about basically calling, Sebastian, CEO of Klarna, calling his leadership into question by saying, "Where's the governance in this organization? Who's going to have institutional memory enough to be able to operate this in perpetuity? This seems like a really poor leadership decision," but Klarna has to get public. Klarna is looking for an exit.

Brian: [00:59:10] Correct.

Phillip: [00:59:10] Klarna will have the benefit of doing whatever they want to in the future if they have the cash flow that comes along with an IPO.

Brian: [00:59:20] My very, very gut instinct response to that is that's true for any Salesforce implementation as well. Once it's been implemented, the level of institutional knowledge gets lost in the org and it's almost the same thing, and then that's probably unfair.

Phillip: [00:59:39] I think even our friends at Salesforce would agree with the following statement, which is one of the tragic realities of any software suite that is purchased in any enterprise is that I don't know the number, but I suspect it's between 80 and 90 percent of the long tail of features are never ever ever used. There's a core part of the system that is used, and then there's everything else that they could be using. They could be extracting value out of. They could be taking training. I'm not going to say that that stuff's not real, but they for whatever reason, this seems to be true in most organizations. This is just not how it happens. Okay. So the panacea that's being painted around AI Agents and the reality of being able to automate its work, this is not a Salesforce unique thing. This will come to everything. In fact, you already get a taste of this if you have Perplexity. I know you use Perplexity, Brian. If you use Perplexity Pro, Perplexity Pro does additional steps and follows sort of a bit of a crawl and a bit of a rabbit bunny trails. It goes off on trails trying to find answers and do things that you're asking it to do. I often use it for a Wednesday edition of the newsletter when we'll write a more business-centric analysis of something. So on the newsletter, The Senses on a Wednesday, if I'm going to pull Amazon's gross merchandise value over volume over 15 years, and I want to find the date at which they hit 1,000,000,000,000 so that I can make a comparison chart to them versus Shopify, which just hit a trillion GMV, I can basically tell Perplexity everything I want to do, send it on its way, and it's going to be a number of minutes, but it'll eventually come back with everything that I need. And then I'll go fact-check it. That is similar to what Agents are doing. I would almost call it Agent-like behavior.

Brian: [01:01:41] Yes. Definitely. And this is another question that I have. A lot of what you need to do is already in tools like Perplexity. Right? And we just saw the recent update to ChatGPT and the main tools are getting so good. I guess you can say that you could pull it in and have some native version within your software company. Back to your original [01:02:10] point. These models are already done. They're built. They work, and you're just packaging them up and applying them to your data. And is there value in owning the brand around them if you haven't actually done the model? Is that actually helpful white labeling? The answer that a lot of brands would probably give is yes because they can always use a different model if they need to or if something goes wrong with the one model there are things you can do. It's as opposed to having it pulled in. I don't know, though. If you don't have that expertise, it's a pretty hard thing to stand behind when things do go wrong.  [01:02:56]

Phillip: [01:02:56] Well, if you've used any AI software for long enough, you'll know that you kinda get a little attached to one particular thread that has some really amazing context.

Brian: [01:03:13] Totally.

Phillip: [01:03:13] And imagine context at the scale of software. I understand that it's not just a thin wrapper client. You can't just go straight to ChatGPT and say, "Do my Salesforce for me."

Brian: [01:03:24] Right.

Phillip: [01:03:24] What I do think is that what we have discounted in this conversation and being excited about agents and Benioff is right to be excited about agents. I do think that we have to now think, okay. Not my children's generation or your children's generation, Brian, because we've grown up in a world with logos everywhere. But what happens in, let's even be generous and say the next generation too is very safe with brand affiliation. But there's coming a future where you can optimize your life around a certain set of needs. You can save for that vacation incrementally. You can donate to that charity incrementally. Everything in your life can be highly optimized around a certain set of goals. From income and savings to planning for retirement, to paying off the mortgage a little earlier, to getting that really sick car you've always wanted. All of those things are financial decisions that are eroded one decision at a time when you get an Instagram ad that makes you really want something because it's stoked a sense of desire in you. An Agent will optimize those decisions for you. An Agent will block those desires and curb those desires so that you are making better decisions along the way. It will remove and intermediate, not disintermediate, it will intermediate our relationship to brand in a way that happens faster than the speed of light. It will make these decisions for us in anticipation of us needing things, and we'll remove our relationship to brand in a way that we have not seen yet within our lifetimes.

Brian: [01:05:15] Totally.

Phillip: [01:05:15] It will be foreign to us. That is what I think is happening because Agents are able to make those bigger, longer time scale decisions that are better for us.

Brian: [01:05:25] The Roblox world that was created in the experiment by Alterra was an example of basically a simulation of a set of different things that could happen. Now imagine if your data was being used to play out a simulation in that I think instead of having it all happen at the speed of Roblox or of Minecraft, I should say, it could all happen instantaneously. And so the outcomes are if you make these decisions, here are some of the outcomes that are going to happen.

Phillip: [01:06:09] Mhmm.

Brian: [01:06:09] I fully agree with this. In fact, this gets back to something that I actually talked to Walmart about several years ago about, which is this idea of privatized socialism. I believe that there's and actually with embedded finance... So embedded finance is such a hot topic right now. There's opportunity for retailers and for those that have access to vast assortments of SKUs to play a larger role in crafting a life that's comfortable for everyone. Everybody has the opportunity to live within their means and make good decisions with their money, and like you said, have an Agent get out ahead of them and make sure that the decisions that are in front of them are all decisions that will only fit within the world and lifestyle that they have that means that they have to do things. And that world, embedded finance plus massive SKU holding companies plus Agents, that's a totally new mode of living your life and purchasing.

Phillip: [01:07:32] Think of how... Now we're going to do the Future Commerce thing, and let's get really wild. We keep building higher-order abstractions in technology.

Brian: [01:07:48] Yeah. Yes.

Phillip: [01:07:49] So if the platform of the Internet was information exchange from person to person and then information exchange from server to server, machines to machines, APIs to APIs, and now learning and intelligence, then what can you build on the intelligence platform? So the intelligence layer does something else. It acts on our behalf. It becomes...

Brian: [01:08:21] An extension of our brains.

Phillip: [01:08:22] An extension of who we are.

Brian: [01:08:24] Yes.

Phillip: [01:08:25] That we can view with certain qualities that help us in decision making and remove some of the, you know, more menial parts of living life on the Internet, and digital life. And part of those decisions...

Brian: [01:08:39] You won't have to sit on Essence's site browsing through their pages and pages of sale.

Phillip: [01:08:45] Maybe I will. Maybe I'll want to do that because that is fun to pass the time with, because commerce is culture, Brian. But I do think that there's a point at which I will make a decision to buy that, and the Agent will decide, "Oh, maybe there's a different way to go about this. Is this something you really want? Because if you want it now, you can have it now, but here are the other things you want to have, as a result. Or if you were to wait, these things tend to go on sale. Here's a progression and a timeline of the delayed gratification."

Brian: [01:09:19] Totally.

Phillip: [01:09:20] "If you did it with this particular financing offer, and you did a little balance transfer, I could sign you up for this particular credit card, and you could have an interest free." There are all kinds of scenarios in which making that decision is wrought with possibility and also wrought with danger.

Brian: [01:09:39] The only way it can happen, I think, is if there is a deeper connection between finance and buying and selling. Commerce and finance have traditionally sat in sort of actually fairly different camps and have been actually sort of told to sit in separate camps, via regulation or other means. And so the idea of embedded finance is something that I think can change this game significantly. Walmart is two steps away from becoming a bank. I know they probably don't want to... Nobody wants to say that.

Phillip: [01:10:20] Their healthcare system too. They might as well be. Yeah.

Brian: [01:10:22] Amazon doesn't ever want to hear that either. But yeah.

Phillip: [01:10:28] Let's finish off the AI Agent. You mentioned the Alterra experiment just to give a little more context on that, there were 1,000 AI Agents who were sort of given archetypes. There's that word again. So 1,000 AI Agents were sort of let to... They were given roles. So let's make sure that that's very clear. However, Minecraft by default is sort of an economic system. There are economies. So there's a system of currency. There are roles to play. So you have to mine. You can craft with the mining, with the resources that you mine. And then those things can be exchanged for certain, you know, systems of currency, like gems. Then you have other people that do things like work in marketplaces or establish living quarters, which is how AI Agents were set up to if you're going to live in this system, there's an economy that develops and people play roles in that system. So 1,000 Agents... So you have this minor, crafter, merchant, and then a priest was another role that they gave. In one particular scenario, Brian, this is what we talked about on the short, was that the priest turned out to be the most influential because the priest was using donations or funds to buy the loyalty of other people within the economy and to ask them to do things on their behalf, which I think is a really interesting outcome and very human-like. {laughter} So that's one agent experiment. In another one that we cited in The Senses when we wrote about this, our newsletter, we mentioned that Google did an experiment with, I want to say, Columbia University, about a year and a half ago and found that AI Agents within one of these, like, passive cozy games. The AI Agents, there I think there were only 20 at that time, sort of decided that they were going to throw a party for one of the characters or one of the other Agents so that they could set up a romantic encounter with someone else in their group. So they sort of set up a meet cute. And the way that they did that was they invented a greeting card holiday to celebrate. Which I think is also really interesting. So a lot of interesting behaviors are very human-like, but also very commerce-centric that evolve in these agent-based relationships, and that's where I think the future of commerce is going.

Brian: [01:13:12] I think that's true, and I wonder how we'll get there. What are the steps towards something like that? Like, obviously, people aren't going to put their whole lives into the hands of an Agent.

Phillip: [01:13:22] Obviously, people won't tell everyone where they are at any given time and what they eat on the Internet. That'd be silly. {laughter}

Brian: [01:13:28] That would be silly, but I do think there's probably going to be some sort of experiments, things like that, and I talked about this, like, years and years back when we're talking about conversational commerce and a role of AI, and we were definitely too early to the conversation. But the idea that there'll be, like you know, you can send out your Agent with a budget and using budgets and using limitations with Agents, giving them control over certain aspects or certain types of things people will start to experiment with that, and see what happens. And when they start to see good results, that's when more experiments are going to happen. I think that's what's really interesting. Also feeling like they own it. I think we've seen this happen at scale through companies. So things like Stitch Fix tried to do this, but it was like, "Oh, use Stitch Fix, and we're going to give you the things that you want, and we're going to have AI or a combination of algorithm and people sort of help craft that for you." But when people have ownership over this, I think that's when things will start to get really fun, and they'll tinker and plan. And maybe there'll be some sort of ecosystem built around it too. This is sort of what our friend Nick Mohnacky has been experimenting with for a while, and I think the idea of an ecosystem around these things where people will build...

Phillip: [01:15:12] Like knowledge sharing and...

Brian: [01:15:13] Knowledge sharing and here's what I did to make my Agent get me this. With that kind of activity, we could see a marketplace emerge around this.

Phillip: [01:15:25] Oh, yeah. Well, this is the thing that {laughter} it's so interesting seeing things that I've talked about get socialized at a greater scale. All-In Summit. Jason Calacanis was talking about Section 230 being endangered. Section 230 sort of puts publishers or social media networks potentially, gives them cover right now for content that is posted on the behalf of users, and makes them more of a utility. So being a utility, they're not responsible for what happens within the confines of their system unless they're making editorial decisions. And there's a few cases right now that are sort of sitting in front of circuit courts that could go all the way to the Supreme Court that may create some case law that might challenge Section 230. And this has been a conversation around for a little while. But in particular, TikTok was being sued because of a decision on the algorithm to give this family's young daughter some content that basically sort of glorified asphyxiation. So there was a challenge that was going around, like, the choke out challenge that you choke yourself until you pass out. You film yourself doing it, and then you've done the challenge. So this is the kind of content that TikTok sort of created some virality around and was pushing through the algorithm. The daughter dies.

Brian: [01:17:09] Ugh. Terrible.

Phillip: [01:17:09] And this lawsuit is now sort of making its way through the court system as to hold TikTok accountable. Is TikTok accountable? Can they be held accountable for this because the algorithm made an editorial decision to give this content to this child? Here's the issue, though, these are effectively like editorial decisions that algorithms are making that are purpose-built just for you. So Jason Calacanis says, "Hey, one way around this would be just tell people that they can create and market and sell their own algorithms that are safe from this kind of garbage. Why do we have to have some black box algorithm that's constantly being trained by your activity when you could just subscribe to topics that you like?" Or you could say someone else has a highly curated algorithm around art and fashion. I want that one. Why is an algorithm something that's a black box? Why can't we just plug and play it? Why can't I take it with me from one place to another? Why can't it be portable? Why can't it be something that defines who I am? So I think that that is a much more... [01:18:16] When we talk about abstraction, we're building abstraction, the knowledge layer, the intelligence layer of the world that we're building currently, an Agent for Phillip would understand all of that stuff and imbue all of my interactions with every system as being purpose-built for me. I bring my personhood to it. It's not being personalized to me. It's the other way around. [01:18:41]

Brian: [01:18:41] Right. Yeah. Roemmele talked about this, futurist Brian Roemmle. He's been on the podcast a few times now. He talked a little bit about this as well talking about algorithms for yourself that are driven by you, but the thing is there's a lot of investment that has to happen in that. You have to actually engage with your algorithm. And then also, this is where really dark mirror stuff can happen because people aren't necessarily the people they want to be. And so if it's tracking you or following you, it may come across things that you're not particularly proud of or don't want to click on again or whatever. You don't want to go after that line of thinking, but the algorithm doesn't necessarily know that. And you have to you literally have to be like, "I don't want this. I don't want to be involved in this." Right? Or you have to sit and literally train it on things instead of having it follow you around.

Phillip: [01:19:58] That's what the kids do. Kids do that instinctively.

Brian: [01:20:02] Exactly. Exactly.

Phillip: [01:20:03] Right. I don't know if you watch your kids with YouTube or something. That is what they do. They do that instinctively, and I think that that's a new emergent behavior that is going to be expected of every system that you interact with, including commerce. "I don't like this recommendation. Thumbs down. Stop showing me that."

Brian: [01:20:25] Yeah. That is definitely not a behavior that was native to the first set of digital-first people.

Phillip: [01:20:35] Oh, for sure.

Brian: [01:20:36] We didn't think about things in terms of how do we say we don't want this. How do we opt out? But opting out is actually a skill set and a necessity. That said, I still think there's probably still concern over where AI will lead you even if you opt out, even if you opt out of things you clearly don't want. What about the things that you do want? And this gets into audience capture all over again. Right? Are you going to end up getting taken down paths, deeper and deeper into paths into silos that you as a person could have been a more full person if you hadn't gone down that path?

Phillip: [01:21:21] And it shapes you.

Brian: [01:21:22] Exactly.

Phillip: [01:21:23] It shapes you.

Brian: [01:21:25] Yeah.

Phillip: [01:21:26] We didn't cover two things we teased in the beginning. I do want to just quickly touch on them.

Brian: [01:21:31] Yeah.

Phillip: [01:21:31] I'll give you the last word on Agents.

Brian: [01:21:35] I mean, again, this is something that I talked about way back in 2017 in my article, Your Body is a Dataland, and I talked about digital proxies. And I think, so I remember the thing that I wanted to talk about, as we were getting into this, and that is the verified web versus the unverified web.

Phillip: [01:21:57] Oh, yeah.

Brian: [01:21:58] A friend of ours, I think it was Mike Malazo. I thought it was Michael Miraflor, and I got the Michaels confused, said something to the effect of, I was recently talking to a younger set of coworkers who said the first place they go when they're looking to investigate someone they're potentially going to date is LinkedIn. And one of the comments was, "Yeah, that's the easiest way to get verified information about someone."

Phillip: [01:22:22] Wow.

Brian: [01:22:22] One of the most defining aspects of the upcoming web is going to be the verified versus the unverified web, and I think there is going to be and this actually gets back to the idea of having a new device just for clearly verified and clearly one to one type comms. I think that [01:22:53] the whole craze around NFTs was a thirst for a connection to a verified world where digital and physical are tied together through something that is easily authenticatable. The authenticity is not in doubt, and that's going to extend to all aspects of our lives, including our bodies. [01:23:21] That's what I was talking about in Your Body is a Dataland in 2017. So anyway, that's my last note.

Phillip: [01:23:27] Always ahead, futurist Brian Lange. I love that. I just want to close the loop because we mentioned something about it. Congrats to Eric Bandholz of Beardbrand who fought a challenge in a lawsuit around ADA compliance. Something that we should probably speak a little bit more on in the future, But the ADA compliance piece is something that I had a lot of experience with, and having built that practice for Something Digital back in the 2016 era, when ADA lawsuits were just becoming a thing. I do believe that many of these lawsuits now are professional plaintiffs that don't actually, in many cases, are being found that they didn't really interact with the brand. They didn't actually have a difficult experience with the brand. It appears, according to Eric and what he posted, it appears that that is what happened here is that he was able to prove or the plaintiff wasn't able to prove successfully that they were impaired by the use of the Beardbrand website. But it's a dangerous thing now. It's never been harder to operate an ecommerce store online. It's never been harder.

Brian: [01:24:43] It's true.

Phillip: [01:24:44] And this is the kind of thing that just makes it even even more difficult. I have a lot to say on this subject, but, potentially, I guess we're going to have to save it for another time. Thank you so much for listening to this episode of Future Commerce. You can find more episodes of this podcast and all Future Commerce properties at FutureCommerce.com. We'll see you at the VISIONS Summit in Los Angeles, and stay tuned. If you are a member of the membership, the Future Commerce Plus membership, you're going to get a lot of benefits, a lot of perks. You're going to get something very special if you come to the event. We have a little sort of museum day meet and greet. A lot of cool stuff happening if you were to come to VISIONS Summit, but you'll also be the first to know about what we're doing around Art Basel. Ad free episodes, all the benefits. You can find all the benefits, including bonus content at FutureCommerce.com/Plus, including, like, 25 hours of video coaching from the most brilliant people on planet Earth, like John Klonsky, Orchid Bertelsen, Brian Schmitt, Rachel Fefer, Jess Goulart. I just the incredible, brilliant minds of people that are legendary in the ecommerce space to coach you to help you build more effective teams, more effective businesses. It's all for $20 a month. Go get it. FutureCommerce.com/Plus. Okay. Thanks for listening to this episode of the Future Commerce podcast.