Marketing to the Machines: AI Agents and the Future of E-Commerce

Transcript generated automatically from the episode audio and lightly cleaned up (filler words removed, misheard names corrected where clear). Speaker labels were added afterward and may contain errors.

Rob May: Hi everyone. I'm Rob May. I'm the co-founder and CEO of Neurometric.

Stephen Messer: I'm Steve Messer. I'm the co-founder of Collective[i], former founder of LinkShare.

Rob May: And you're watching the AI in NYC show. All right. So, this is, so Stephen has been on a couple of times, as a as a guest. Stephen is now a regular host for us, for this season, which is super exciting. And, I actually want to I want to kick it off with, so you have you writeload.nyc. You had a really interesting article today that started with the idea that Bob Metcalfe, the inventor of Ethernet, said never call him a visionary and it went through and talked about why we shouldn't call serial entrepreneurs necessarily serial entrepreneurs might not be the best the best phrasing. So what I guess first like what inspired you to write that? I thought it was really good by the way.

Stephen Messer: First of all, thank you for saying that because it could have gone either way at the way you positioned it. No, look, I it's funny. I've watched for a long time. People think serial entrepreneur is a good word. It drives me out of my mind. It makes me nuts because it is the setup before the fall. And it's funny because Bob Metcalfe talked about visionary. It's got to be what 20 years ago, 25 years ago. He wrote that article and wired and everyone kind of just ignored it. They sort sort of switched the word up and moved on the next one. And I've known Bob for a long time. He's amazing. And for whatever reason, people think serial entrepreneurs should be like a word you like. It shouldn't be. It should be one you fight right away. You can be a founder, you can be a CEO, but you but the idea of like a serial founder is a horrible thing.

Rob May: Yeah. A lot of people like it though. A lot of people introduce themselves with that.

Stephen Messer: A lot of people are stupid. I mean, I can't What is What is it saying? I can fix a lot of things in life, but I can't fix stupid.

Rob May: Yeah. I mean, that's sort of I think it's it sort of self-def. Yeah. How and so what did you propose instead? How they should think about it?

Stephen Messer: I mean, look, look, I think I made the point pretty clearly when I said, "Okay, would you ever call someone a serial investor?" And you'd think, "No." How about a serial CEO? Cuz almost every NASDAQ, New York Stock Exchange CEO has been a CEO more than one time. Yeah. And if I said, "Oh, he's a serial CEO. Serial CFO, serial CMO, you pick the title." They would look at you strangely. And only founders somehow think it's like a badge of honor. I for for the life of me, I can't figure out why.

Rob May: It is weird. Every everybody definitely has their pet peeves. I It's not It's not that big of a pet peeve to me, but I thought the article was really good.

Stephen Messer: To me, it's not a pet peeve. To me, it's a reality of keeping my equity.

Rob May: Yeah.

Stephen Messer: I don't think it's a Look, if I don't care if someone calls me an I'm in New York. It's every other day. What I do care about are things that set up outcomes long before you do. And words have meanings. Like when someone starts seeing you as a serial entrepreneur as opposed to the CEO of the company, Yeah. You have set yourself in a very different role. Yeah. Leave and go on to the next thing. You're the early stage guy. You're the guy who gets it off the ground. He's a little crazy, a little kooky.

And I gave a lot of really good examples of people. Steve Jobs being the perfect example of prior to coming back and saving Apple, he was considered to be a serial entrepreneur who was kind of a crazy. By the way, you look at the movies they put out about him. [clears throat] like he was the guy flying the pirates flag. Well, he did that because the whole Oregon becomes so corporate, it couldn't get any decent product out. And his whole point was, I'm going to make a great product. Nobody will let me do what I need to do. I'm going to go somewhere else. But people treated it like he was out of his mind and they had to move the sick kid out. And I think that's what happens with the especially if you think about professional investors, most of them have not been founders. Most of them are terrified. Most of them are bad investors. They take very few risks. Yeah. They're consensus investors and so they themselves are looking for someone like them. Yeah. Which will kill the company definitively.

Rob May: Yeah. Some someday when we decide to do our final show, we'll call it the name names show and we'll talk about specific [laughter] stories that we tell.

Stephen Messer: I think I do that every time you invite me here. I know I start sharing all the names.

Rob May: Yeah. Okay. So, our guest today is teaches parttime at Columbia University. She's former Amazon exec expert in agentic commerce and the founding principle of MM Ventures and we're going to talk a little bit about a specific project that she founded called the beauty marketplace. So Mary Palmieri welcome to the podcast.

Mary Palmieri: Hello. Thank you. I will not call myself a serial entrepreneur.

Rob May: [laughter] Well, you're involved in lots of other things other than yeah, start I mean teaching at Colia is pretty impressive and [laughter] all these other things, but tell us a little bit about so the beauty marketplace is this thing that you came up with while you were teaching the digital project innovation course at Columbia University. So I'm curious how did how did teaching that course at the same time that you were doing this like how did it shape your approach to like building and launching this from the ground up?

Mary Palmieri: Yeah, so traditionally I was a subject matter expert. I kind of became a technologist accidentally. I started my career in media. I worked at Vogue and then I was producing a product development trade show for beauty. And I thought I'd meet all these emerging brands no one knew about back then because there was no direct to consumer. This was like 2008 2009. So I thought combining my love of like writing and media with you know giving a platform for brands. I created one of the beauty industry's first like e-commerce platforms and that prompted me to go back to school. I'm like all right I know a lot about brands. I know a lot about marketing. I don't know much about technology.

So that was my intent with going back to school. That was my whole thesis. And I thought it was going to come out just finding people I would want to build with. Like the whole intention was an e-commerce platform and I took this course it was called at the time knowledge product innovation and it's all about manipulating the use of data to create a competitive differentiator in the market and we were studying business models like Netflix and Pandora and I had an aha moment I thought oh my gosh like I don't want to just build an e-commerce platform on Shopify like I really want to solve the problem of all of the overwhelmed readers that would write into the blog and the number one question was what product do you recommend and I thought that's such a unique answer for me or someone with blonde hair or you guys

so I thought let's build the Pandora or Netflix for beauty and I built a recommendation engine and I learned a lot about metadata taxonomies And that's what we started working on in our class. We built, you know, kind of like Pandora has their music genome project. We built like the beauty marketplace. It was our proprietary taxonomy using our database of over 1,900 emerging beauty brands. And so we built the recommendation engine in the program. Never thought it would take off. I thought I'd end up working at McKinsey or a management consulting firm. And then when we got venture funding, I'm like, "Oh, wow. This is real. Let's let's go. And so that's how it all began. And then of course they invited me back to teach the course as a practitioner. And to your point like be the real world like you know the context of this is what it takes to start a company versus just like invest in one or talk about it.

Rob May: So does it work for men? Can it help Stephen and I be a little more

Mary Palmieri: that's the number one question I get from guys. But I feel like men are simply you kind of like have your products pretty much but I don't know shaving cream.

Stephen Messer: [laughter] Okay.

Mary Palmieri: It's changing with the Gen Z market though now.

Rob May: [laughter] so this is an AI-driven personalized marketplace that matches brands and consumers. And one of the things you have is you have the selfie quiz that is on there. And so I'm I'm curious. You've been in the beauty e-commerce space for a really long time. And so what was the specific friction point in online shopping you saw that convinced you that like you could now take AI and that matching technology through this was the solution as opposed to, you know, all the other areas of business that you could have innovated on.

Mary Palmieri: So I think when I got a little bit of seed funding, I ended up moving from New York to Silicon Valley and that was sort of the best environment to be in because you're challenged to think bigger. And so when I was in H

Stephen Messer: are you? I mean, if we're being honest, ask for more money.

Mary Palmieri: It's [laughter] ask for more money.

Stephen Messer: Is Silicon Valley where you really think bigger anymore?

Mary Palmieri: Well is it this was in 2017 so okay arguably that was like the uber like that was the pinnacle to your point it might have changed

Stephen Messer: if you said if you wanted to find people with group think I would have absolutely agreed [laughter] but we haven't met cuz group think doesn't

Mary Palmieri: [laughter] no I mean for me it was a great environment to be in because I hadn't been around so like I feel like venture capital is almost trending because it's so common on social media now, but back then like you'd ask like how do I raise venture capital and it wasn't like widely known unless you went to like a business school program or you're an investor especially coming from like the beauty space. I mean, I had Emily Weiss of like Glossier who was like a little bit before me and she raised some money and she had worked at Vogue. So, I kind of talked to similar people and I'm like I really had to figure it out for myself kind of.

So for me it was the right environment cuz I'm like okay I really had to think about what is our IP and what is our competitive differentiator and that's when I really like leaned into being more of a technologist cuz I was never an engineer by training but I think it's just a beautiful thing to find a way to really develop intellectual property and set yourselves apart and that was kind of our focus. And a lot of people were kind of doing online quizzes and I thought I wanted to be more than that. I really wanted to be like a recommendation engine. And I actually met my partners in the incubator. They had they were doing like virtual reality technology like virtual tryon. And we ended up partnering and we added the selfie quiz into the mix. And with our technology, we kind of pioneered like extracting 53 data points from like a photo. And that was even more data, like quantitative data we could use to create like these matches. And you know, that's how it sort of evolved. And we ended up selling it as a service as opposed to trying to compete with Amazon at that point. So yeah, and that's when it kind of took off.

Rob May: Cool. How do you feel having worked at Amazon and now done startups like big techs versus big tech versus startups? You know, they have different pros and cons obviously like in this AI world. What is the biggest mistake that you see the legacy retailers make? It seems like they're trying to wedge it into their existing workflow. Sometimes it's not working very well for them.

Mary Palmieri: I feel like, and you kind of touched on this before we even started filming, but it's just the more cooks in the kitchen, the harder it is to really align on a recipe and you just kind of get the kitchen sink, so to speak. So you in Amazon I was surrounded by some of the smartest people who've had so much knowledge and worldclass executors but to really own an innovative concept from start to finish I mean it it really is an aversion to risk especially when you have something that's working so well. So I think you know and that sort of validated my belief in the whole entrepreneurial spirit. You need not serial entrepreneurs but entrepreneurs to you know introduce the solutions and take the risks because yeah like once you're once you're doing something so well especially like Amazon operationally you don't really want to disrupt homeostasis there. So it is no matter even if you are a tech company you're kind of riskaverse to introduce anything extremely new.

Rob May: Oh, I have a great story related to your kitchen sink analogy, which is that in my first startup, two of my investors had this engineer that worked for them who was like, they're like, "Hey, he's really smart guy. He wants to join a startup. He wants to come work with you. You should give him a shot. He wants but he wants to move into product management." Everybody wants to move into product management because everybody's like, "I can design products better."

Mary Palmieri: That's [laughter] how it is like the hardest job in most companies, right?

Rob May: Yeah. So we bring him in, we give him a shot and the very first thing that he you know goes through a couple of cycles and produces something and he's demoing it to me and I'm like this is a piece of None of it ties together. It's all and he's like well the he's like but the sales guys said they needed this to sell it and the engineers said it had to have like this thing is built this way and like the marketing team said if it didn't it did this and then like and the product we wanted and I was like okay man like I get that but like your job is to say yes and no to the things that like make it a consistent product and give it a vision and it like you have to make trade-offs and you have to make and that gets harder the bigger the bigger group that you have.

It's why Amazon was famous for like their what they call them the two pizza teams, right? It's like your team stays small such that they can live on two pizzas you know as a meal until such time as like your product's really taking off and generating significant revenue then you can think about scaling and I think startups are the same way

Stephen Messer: and I was going to say I think also for for the people you're talking about especially in fashion I think what's hard for them is that their whole organization is built for the customer they have and you're essentially trying to argue to them there's a new customer that's out there and they're looking at the same brand they're looking at the same organizations, same warehousing, same products. And I think they struggle with understanding that on online they're buying, their customer is buying differently. Their customer might be a different age. They might want a different experience. And so the hard part is you've got a group of people that are really great at their call it legacy business, call it whatever you want. And you're asking those people who are operationally great at what they do, who do know their customer, you're trying to point out something that's new. I almost wonder sometimes if they would have been better doing spin-off brands with new teams, taking the same product lines but just marketing it differently, essentially having an internal disruptor.

I feel for them because they have to hit quarterly returns and I mean I watch this at LinkShare all the time. You had some the direct marketers were the best because they just cared about numbers, but a lot of the traditional a lot of the a lot of the people in your sector cared about brand and the brand would kill every opportunity to in innovate. It was it was horrible. They just couldn't change for fear of changing the brand to a new audience.

Rob May: Yeah. What brand do you think has done the best job of reinventing or migrating itself over time? Do you have an opinion on that as a as a person who pays a lot of attention to brands?

Mary Palmieri: Oh, that's a really good question.

Stephen Messer: It's a really hard thing to do.

Mary Palmieri: Yeah. Well, and especially in the luxury market, to your point, there's a huge resistance to, you know, Chanel

Stephen Messer: Mhm.

Mary Palmieri: Like what will the Chanel customers think if we digitize and but I think this isn't a luxury brand, but I think a brand that does it really well and was willing to adapt technology right from the beginning is Nike. Like if you go into any of their stores, they've just embraced all throughout the customer experience value chain. They've embraced innovation. Like everything from going in the store and working with the sales associate, I think they were the first ones to have like the point of sale technology, even something like that to the whole customer experience online. You know, they might need to be a little stronger from a merchandising perspective lately in comparison to what's out on the market, but arguably I think they're winning on customer experience. And yeah, to your point too, I mean now these brands do have another customer now. It is the agent.

Stephen Messer: So I it's almost like the worst customer on the planet.

Mary Palmieri: [laughter] Exactly.

Stephen Messer: And one you don't want, right? I mean like the one that only cares about price, the one that gets no upsells, the one that gets no advertising. It's literally the worst potential customer. It is at the bottom of the barrel.

Mary Palmieri: Yeah, it's crazy. Yeah, it's like the Debbie Downer. Like [laughter]

Stephen Messer: although I would go back to your point, Luxottica has done an amazing job. I think they've done an amazing job at figuring out how to keep their old customers and attract new ones. Their partnership with Meta was great. I thought they've done it probably of the brands. Unbelievable.

Rob May: Yeah. Yeah. I don't you know I don't remember this because I you know it happened when I was probably too young to follow this but I believe that Target was a new brand started by an old retailer, right? Where they were like we got to reinvent ourselves but we can't do it as this brand. The brand started with an M or something like that and like we're dying but we have this cash flow like let's build a new thing from the ground up. Which is hard to do cuz things grow so slow in the beginning and people think, "Well, it's not going to work." And you're like, "No, no, no." But give it 5 years. And it was a new model cuz they were selling their own stuff that they were manufacturing, trying to do high quality, high design in a relatively low-end market. I mean, they they did a great job. But I want to pull on the thread that Stephen brought out about agents are the worst customers. That's a really good that's a really good line. And I'm curious what no do do the retailers and the e-commerce companies realize that yet and do they think about it that way or do they think about it at all? Are they just figuring out what to do? So that's like part one. And then number two, what should they do? Can you do anything?

Mary Palmieri: I based on the conversations I've been having this year, I think this has been an educational learning year for a lot of people, especially people who are traditional marketers. I think if you're a marketer now, you have to have some crossover with tech. But it's just been such a quick pivot. So, a lot of my conversations anyway have been educating. But I think to succeed in getting discovered by this boring agentic customer, yeah, there's like a couple of levers, companies can pull. I think people are pretty much aware of that framework now, but it's just controlling your narrative and what's out there in the market. So, a lot of content optimization. It's also what the beauty marketplace was built on your product catalog optimization and matching what how like how customers are searching your products are shifting and so figuring out those use cases and thinking through that. So yeah there's a lot of learning happening and I think fundamentally too in order for all of that to happen there needs to be a baseline of AI readiness which I think those are few people have the yeah even with the big tech companies some of the projects I'm working on now are just AI readiness and scratching the surface and making sure the right data is available when and everything's sort of integrated and protected and all of that.

Rob May: So yeah, I mean I mean there is some like there are certain brands that even when you know you only feel that way when you wear them or use them or whatever because they have spent massive amounts of money associating neurons in your brain with their brand and other things, you know, famous people and fancy things and whatever. You know, so at some level you are going to be too frustrated with your agents if they always optimize on price for you.

Stephen Messer: Would you would you I think anyone who's using agents thinks if I was going to start making them multi-variant they will blow up pretty quickly. I mean look I think here's here's probably what's going to happen. Amazon already set the goal right which is block muse completely.

Mary Palmieri: Mhm.

Stephen Messer: So you know the debate it's probably going to take place over the next 5 years will be do each do each one of the brands try to have their own their own agent. So, it's like when you go into a store, you have your salesperson, you deal with it, you get you get a little loyalty program, you get something else. I think knowing the retailers the way I've known them, that that'll probably be their first gut reaction. It won't work well. So, then there'll be a service that will try to provide, you know, like a Spotify like agent service. Then they'll hate that and it'll end up going to Amuse or someone else just cuz, you know, who wants to use a separate agent for 16 different things?

Rob May: Well, there's what do you guys what do you guys think about this potential future? So, I one of the email groups that I'm on for AI, there was a discussion that came about this week around is Muse EV is Muse going to be the iPhone moment what what iPhone was for corporate IT is Muse going to be that for like corporate agentic frameworks or whatever, right? So, the corporate

Mary Palmieri: Yeah. So, I mean I know that's what they're trying for small businesses at least.

Rob May: Yeah. So the idea will be well we have our own agents that we take job to job and for the years that we're there plug in and have access to some corporate thing but any knowledge that they get that's more general about how to do things or whatever we can take with us and so if that world happened like the way you would you know not you it used to be if you needed a cell phone at work they issued you one right and then you know everybody wanted their iPhone and all that so IT departments learned how to deal with it but are we going to live in a world where we have our own agents and they can and then the brands will have some agentic plugin where the agent can like plug into that and learn about that brand or whatever and come back and report back maybe. I mean I don't know.

Mary Palmieri: I think there's a lot of debate around what the tech stack should be and what the right solutions are. So it'll be interesting to see how it all shakes out.

Rob May: And you don't always get the most optimal solution from a technology perspective when the way these things evolve, right? You get like the most economically pragmatic solution is where we'll land. So you're so you're basically do you think it was a good idea to hire the CEO of for Meta?

Stephen Messer: I don't think it's going to work. I think I understand why they did it. It's not going to work. And it goes back to the branding thing like you just like would you use meta email corporate email? Like no you just wouldn't because you just it's so tied in your head. They had they would have to launch it under another brand with like meta AI like say they're a great AI company. They've got phenomenal data. They should be able to do this market but I think they have to do it as another brand or spin or something separately.

Rob May: It felt like they were trying to copy Oracle strategy of or Google strategy of hiring Kieran to like run like their like cloud operation which is really what the guy would have been good for but if they're going after small business which really would be their sweet spot I can't imagine like buy HubSpot or something right like I don't know why to ride that legacy down

Stephen Messer: they tried that with customer that went well yeah I'm just saying like they they they first one I could see where they did that because then I could see them building the whole brand around it that I could see. But why would you try to convert like HubSpot? It's a dying brand. It's it performing the worst of everyone in the market. It at its core is a SaaS business. Keeps trying to come up with stupid AI products that never work. I mean, it's like like Salesforce, HubSpot, if they make it five years, I'll be shocked if they make it two.

Mary Palmieri: But to your point, I think HubSpot does have a certain like mid-tier demographic. Like a lot of the midsize companies would use them over a sales force.

Stephen Messer: Not anymore. I mean, everyone's ripping them out. I mean, look at their look at their company. What's the market cap of it now?

Rob May: I think it's called the cliff. It just drops. The SaaS cliff or whatever.

Stephen Messer: Yeah. No, it's not even the SaaS cliff. I just call it the HubSpot cliff. I mean, it just keeps going down and they have a they have a leader who's not a founder who's there. So, the person's trying to figure out what to do. Everyone's basing themselves off of a SaaS. I haven't seen them make a single move.

Rob May: That's interesting. While they've got cash flow, you think they'd be moving faster.

Stephen Messer: They've done nothing. I mean, at least you got to get Benioff credit for one a $2 billion buyback to try to make it look like there's life, you know, and trying to acquire I mean, technically SaaS companies that have rebranded as AI, you know, but I mean, look, he's got a lot of AI washing going on and it's not bad. I mean, it's his shareholders are desperate that he can save it. HubSpot, it feels like everyone's written off. By the way, they're not the only one. I mean like I mean pick pick your SaaS provider and when's the last time you've heard anyone talk meaningfully about delivery of value?

Rob May: Yeah. No, that's a good point. I mean it is it's one of the things that I think all these companies have to worry about in AI era is like do these does your value just get absorbed someplace else in the chain right everything that we've it should right should be important. Yeah, I mean generally if you have a disruptive technology that is what the disruption does is it just absorbs usually your highest value stuff leaving you with the junk that nobody wanted until you can't afford to do it anymore and you go away.

Stephen Messer: I mean I like I'm curious going back to the agents in general. I mean like I've I've been shocked the way agents haven't had and like I would have thought if like there was an instinct that would have been like this is your agent, you set the terms, you set the tone, whereas Medis basically selling advertising and your data and they would have gone for like a privacy approach. I would have thought a lot of the brands would have said our agent is designed for the Chanel of a customer. No, no one's even thinking about like differentiation agents. Everyone just sort of assumes it's like back office crap for people and back office crap for customers. And I've been shocked that I would have thought, you know, in particularly the markets you've worked in that they would be out in front, you know, instead of dots and you know you know tomagois that people have, you'd think they would have actually an experience that you'd want. So I don't know. I'm just surprised by it.

Mary Palmieri: I think it's just moved so fast. It's like we went from agents can be your front-end website like customer service agent to now this year we've just accelerated so fast and I'm happy to see that like a lot of the key even beauty players are making deals with like the open AIs and the anthropics of the world because they were slow to the

Stephen Messer: but is that a good idea for them? I mean they they it's you have to kind of you don't have to do anything that relationship to say no and I've seen them say no to a lot of other people before. I mean it feels like giving your business away to OpenAI seems like a bad move generally all around. I don't know why in the world they would do that. I do you see an upside to doing that?

Mary Palmieri: I would be afraid I would be afraid if I was a branding person but I think the problem is like you don't have a branding person.

Stephen Messer: I'm just think of the margins. Yeah. How the hell you just basically killed your company.

Rob May: Yeah. I know but I'm but I'm thinking more back to your question of the so part of the problem here I think is unless you're like a deeply technological person so you have this problem where like if if you want to dial it you know build an agent for something whatever it's hard to understand how these agents work and when you have these like longer running reasoning workflows the failure cases of those workflows can be like quite varied and difficult to understand why if you like you know this thing's interacting with lots of systems like it's hard to figure out like where it breaks and everything else it's probably particularly hard to think about the possible solutions to some of that if you're not technical and there's just so few people that are like deeply technical but have good really good design and user experience sense but you could go you could go find those people and put them in charge of like you make this is for the Chanels of the world you know

Stephen Messer: yeah so you're thinking the bigger argument is that if there's a high failure rate you have to have a large brand to protect you sort of what we do with Visa and Mastercard okay there's a high failure rate which takes my credit card in the back, copies down my number, I don't pay anything because they can absorb the losses.

Rob May: Yeah. No, I was separating a little bit from the open AIF piece, right? I was basically saying like the reason that I think there's no good like Chanel focused agents or agents for Chanel types of people is that there's not a lot of people that could build that effectively based on the state of the market because these agents suck from a UX perspective because they're built by technical people who are like, I love command line, right? Like I was doing it and like I'm one of those people like I love it.

Stephen Messer: Yeah, command line for me, you know, all day.

Rob May: And most people don't.

Stephen Messer: So why why wouldn't there be I mean look why wouldn't there be a Y I mean like YC hasn't come up with anything interesting in a long time but why wouldn't there be a YC company that basically says look I'm going to allow every one of these companies to plug in their loyalty programs to have their own version of what they want we'll just run the back end

Rob May: yeah they should I'm shocked be known [clears throat] yeah that would be a good idea but yeah but then and so separate point is you know should they be cutting deals with OpenAI I think I understand why they do it right. They need to show their team and their shareholder like we're AI focused or whatever.

Stephen Messer: And that's like, oh, give the business away to look good. But that's what they're doing.

Rob May: [laughter] But that's great. I don't number one, I don't think they realize like most people don't realize how many other model providers there are out there, how many other options there are out there. It's like they like, you know, even companies like Cohere don't get talked about very much, right? Or Mistral or I mean, you can't trust the Canadians anymore. I get, [laughter] you know, I get it. So, so yeah, so it's tough and I, so I think they're, you know, they're they're they're kind of doing what they what they think is best, but I Yeah, I wouldn't do it in their shoes, but I probably have very different view.

Stephen Messer: So why do you think no one's talking about it to them? Like you would think like, look, I remember when the internet started up, I remember mobile, I remember all these different waves and everyone had the debate, should I, should I not do it. For some reason now there's this debate of I'm so afraid to look like I have no idea what I'm doing that I'm just going to agree to anything that everybody's with. Like if everyone's buying into OpenAI or Anthropic, whatever I need to do to look like the cool kid, I'm going to do.

Mary Palmieri: That's the broader world though. Also, I don't know that that's just tech.

Stephen Messer: So you're saying the world stop thinking.

Mary Palmieri: Yeah. Interesting.

Stephen Messer: Well, now we have AI, so it's going to get it's going to get even worse. It's going to think even less.

Rob May: [laughter] Yeah. You know, I think it's a combination of people don't enjoy it. I think I you know it's so like if you take a step back and you look at like the web and social media and the whole idea was like wow this will be great cuz like you can find your people now and like people will be smarter because they'll have more diverse perspectives because we thought it was purely a supply problem right it's like oh the big editors at ABC News are squashing down the interesting stories about this thing to put their point of view and what you found was like no we went off in our corners and found the people saying what we wanted to say and we only paid attention to them we didn't think about the demand side of it and the fact that yeah if you could give me more of the things I already want to hear. I would just consume that. We were sort of forced by the structure of the media industry to consume some general stuff that we all saw. And we've lost that. And I think it's both broken some semblance of a common bond we had of like, yeah, we all know Monica Lewinsky, what happened like cuz that at that time like you couldn't avoid that story. That could happen today and you would never know, right?

I mean I think Fox News ran a poll in summer of 24. I think it was before the election. It was something like 72% of MAGA voters had no idea there were any lawsuits against Trump. He had like 24 lawsuits against them, right? He had the January 6th, he had the secret files that he took. He had the rape thing. Like he had all these things. They had no idea be like, and it wasn't like, oh yeah, but I think he's innocent or this is political. It was like they just didn't, oh, really? Like because where would you find that, right? And so I think so that has been a big problem with the structure of the media industry today. And I think it's made people less informed, which also I think makes you, you know, sort of second order effective as you get a little bit stupider because you're not or you become a truth seeker.

Mary Palmieri: Like I think for me, I recognize that bias in certain channels. So, you know, I think the other end of the coin is you can find the people who you know, value and trust and then the that's who you tend to pay attention to and you form your own point of view around.

Stephen Messer: Wouldn't every maggot person say the exact same thing, which is they're truth seekers? Literally found the people they trust. I mean, that that usually would be definitive warning signs of someone. It's one thing to say I go to people I don't believe. I like to hear what they say and I like to then compare it to what I'm hearing. That would probably be a little bit more of a back and forth, right? But I'm curious more. Do you think that there'll be a lot of let's say after the new year, let's let's hope Anthropic gets public. Okay. Let's hope at some at some point OpenAI actually doesn't blow up and gets out there. Do you think there'll be regret for all the deals that they did with these stupid deals that put their business at risk or do you think they're just going to yank it quietly?

Rob May: Regret from on whose partner?

Stephen Messer: Okay, let's say you're target and you're thinking, "Oh, I'm going to go with dots and they're going to be my partner and that's going to be great for me."

Rob May: Yeah. But here's what's going to be like, well, it'll be a more exacerbated version of this, but remember when everybody had a spot in the metaverse? Yeah. What happened to all those? Nothing, right? Because it didn't go anywhere.

Stephen Messer: I mean Meta renamed their company for it. Like how stupid was that?

Rob May: Which shows you again like the importance but it's become so important to follow the trend. Like you and I have been in tech long enough like you remember like in the early days of the internet like it was very much a meritocracy, right? Like it was like what is best? Everybody was logical like the internet was awesome and then you know and then again there were still pockets of but like it is very much now like the tech industry works like the fashion industry. There are influencers people follow what they say. They believe them completely. They repeat exactly what's the cool products like

Stephen Messer: okay does anyone here think instinct is worth anything near what it's worth or do you think they're praying to get acquired

Mary Palmieri: probably praying pl praying to get acquired

Stephen Messer: praying every day to every god in every direction east west north and

Rob May: look like like here's a perfect example right look at Tesla Elon Musk is like and he's been saying this for a while so let I don't know the status today but like let's say two or three years ago winner. He was like, "Tesla's not even a car company. Humanoid robots and they're going to be everywhere." Everybody I know who's deep in the robotics industry, right? I know like the founders of KA, I know deep people at Boston Dynamics and they were like, "Oh, he's insane." They're like, "We are so far from that." They're like, "You can't even like the battery packs on these robots are so big and bulky, right? They they they weigh the things down. They last a couple hours best." Like the like so it's like so here he's selling to Wall Street like we're going to be a humanoid company. You have these a couple of startups that have raised billions at billion dollar multi-billion dollar valuations to build humanoids robots and like we're a couple years later and I realize it takes time but like they're not deployed anywhere at any scale. It's not like the thing you're seeing everywhere. And you know and so but that's a good example of like like Elon Musk has become a tech influencer and whatever he says people don't question it.

Stephen Messer: So well they're afraid to they're Yeah. They question [laughter] it privately. They don't say it publicly because he's been right enough times even if it's 10 years later and nobody wants to take the risk of saying I thought this was crazy. They'd rather say it's later.

Mary Palmieri: Well, do you both think there's a risk, say Anthropic does go public in due time? Just the costs from the utilization costs, I think that's really that's what is preventing the IPO in the first place, right? Like they're projecting the utilization of the tokenization. That is so I think there's going to have to be a cliff where the companies aren't going to be able to afford their usage. Like even my own personal usage of Claude, I'm thinking, "Oh my god, the shoe's going to drop at some point because the amount I use it in day-to-day tasks for the amount I'm paying monthly, I'm worried when it goes public that that's just going to exponentially increase." And I'm curious how both of you see that

Rob May: is a great leadin to some of our stories this week which is API Anthropic filed their prospectus for their IPO and that has been out and so people have been analyzing these numbers they lost $42 billion

Stephen Messer: let's not let's not go down that road the delusion factor is not real okay yes that is a non-cash item that's on their balance sheets to get deals to gave them massive distribution I don't think anyone today holding that shareholder shares are upset about that.

Rob May: Yeah. So, I actually ran an analysis on my blog on my newsletter about what would have to happen for Anthropic to crash with and who would be hurt like what with the all the deals that they've cut, right, with with various people. And I think the conclusion I'll have to go look it up was about things could be about 15% off to the downside and they'll be fine. Excuse me. And then I think another 10 or 15% after that they could adjust. So it's about 30%. So about 30% of they're they have to be off by about 30% to the downside of the way they see the world and the you know the things they're projecting and everything else before you're like they have that like oh moment of like ooh we're we're not going to make it. But I some of the email lists I'm on have been questioning. They're like, "Okay, so to justify their valuation, most of the value that they need to create is actually coming even from products that they haven't made yet." That's a gutsy bet to take.

Mary Palmieri: It's similar, but that's kind of when Uber went public, what was it, 2017? Or yeah, even Amazon. I mean, that's why AWS was started. Yeah. Because the marketplace, you know, that's why I pivoted to SaaS from the beauty marketplace. It's not a very good margin business. So yeah, I think you know there's an ethical debate there too. Kind of

Rob May: well the hard thing is a lot of the things that people criticize them for today. So for so perfect example is like hey they can't charge these token prices forever. There'll always be some frontier that you can charge for but that frontier is going to get smaller and smaller as far as you go. Think about it as like it's like the market for crazy supercomputers. Like it's still there for weather simulators and stuff, but like you know lots of people use supercomputers for things until laptops got powerful enough to do it, right? And you don't [clears throat] need it anymore. It's going to be a similar thing. And so you could still build a crazy supercomputer, right?

There'll always be people that need that that edge. But there's lots of things they can do in the meantime, right? And they've got they've got good distribution and they they've got all these things. They've got a lot of technical talent. And so like there's no reason they can't launch swarms of models and other types of agents and all this kind of stuff. And so I you know I think the hard thing to evaluate about that business or I think the way you have to evaluate that business is kind of think about there's there's a lot of fluctuation in all parts of the AI stack and market right now. So I think you have to think about like where does this likely settle in and as people see that it's starting to consolidate around certain things like where where does Anthropic play and what bets will they make based on like we have these assets the market opportunities are here so like it's a complicated analysis to do but people are doing it I think I mean look if you really want to think through the business the bet you're making is that the token prices don't collapse I think it's that's probably a dangerous bet right

Stephen Messer: One, I think they drive you to do it, right? So, if you bought, let's just say you're the average user, bought a $20 plan, if you try to use it every day, how many times does it stop you? Five times a day. And each time you think to yourself, I need to get my job done. How long before you go to Kimmy?

Mary Palmieri: Yeah. Know, I already use multiple models because of that.

Stephen Messer: We all do, right? And Kimi's a fraction of the price. You're like, it's really good. Why do I care? And eventually, you start thinking to yourself, why am I even spending all this money for 20 bucks when I'm not getting anything from them? I think the downside risk where everyone's like, "Oh, they're obligations. They're obligations." Look, you have to argue that there's no use for those GPUs because they they could resell those you those GPUs to somebody else.

Rob May: That's true, right?

Stephen Messer: They essentially license those license huge mount from SpaceX or xAI or whatever they call themselves now. Like they went and needed those GPU. If you think about the applications that are coming behind them, literally from world models to construction models for I mean like you start thinking about the need of GPU there is there is a huge der of GPU that is needed. Oh yeah. So you could easily see them at one point saying okay the frontier is kind of like not worth it anymore. I'm going to take the deepsek approach of finding a way to create inexpensive ways to scale my model and I'm just going to sell my GPU rights in an auction model or or a futures model and I could see them making more money from that alone.

Rob May: That probably could be true. And I do think you made an interesting point which is I think most people don't realize that like like when we think about these models we think about LLM's like such a small part of the AI market and the smallest market the smallest sm small data set.

Stephen Messer: Yeah. Like when you start getting into world models where you're talking about 3D renderings of a room, I mean that is a vastly different compute.

Rob May: All right. So that so that leads us to another one of our articles today which is AMD acquired FEL's world model startup for $8 billion something whatever. I'm assuming you think that was a good move for AMD.

Stephen Messer: I thought for AMD it was I think the only thing they should have done is bought Yann LeCun and Fay at the exact same moment and try to become the chip for world models.

Rob May: Yeah, it would make sense and you can see it as a way of them going to Nvidia and saying like okay you got hugging face you got a little text models we've got the next thing.

Stephen Messer: Yeah, good luck with commodities. Go have fun with that. Yeah. Yeah. I thought they should I thought it was a great move. I just worry now that it drove up Yann LeCun's model valuation. If he if they had done it simultaneously Yeah. They would have been basically the hugging face for alternative models.

Rob May: Yeah, I saw that deal, man. That deal was something like 8 billion.

Stephen Messer: No, no, no. I saw Yann LeCun's seed round deal was like ridiculous. Like it was billions of dollars post money for the first round, but it's yan. It's worth it.

Rob May: Do you do do you see do does the beauty industry know about world models? Do they think beyond LLMs and image models and some of that stuff today? You know, are they thinking about models that drive the physical interactions in stores and stuff like that or

Mary Palmieri: I'm sure the CTO's and the CIOs and hopefully the CEO are like that conversation's happening in the sea suite is it really trickling down and I think you know so those conversations are being had it's just how do you implement that into the organization and you know what are even the organizational use cases is for a lot of this stuff is kind of where the rubber is meeting the road right now, you know.

Rob May: Yeah. All right. And then last story for this week, which I just thought this was amazing, but they had a White House accord SI, super intelligence, not artificial intelligence. And on the sheet that they signed and held [laughter] up, it did not say the United States. It said the United States.

Mary Palmieri: Comes with the spelling.

Stephen Messer: [laughter] do you think this is a statement on humanity versus cuz machines would not have gotten that wrong?

Rob May: [laughter] There is a there is a segment of the population that says humans are going to start introducing errors intentionally to show things were human and not made by machines. In fact, someone said that to me on Twitter and I was like, not on documents like this, right? You're not going to write your constitution and be like and let's put some errors. So, let's hope not. This is just I mean the funny thing about it is like how many people must have looked at this document? How did nobody catch it? Right.

Mary Palmieri: Well at least you know it's humanmade.

Rob May: Yeah. [laughter] but you know I mean was this just ridiculous to rename AI SI? Like does anybody do either of you guys feel like no no no there's a there's a good like do you make the argument that like no no this is a good thing to do and here's why.

Stephen Messer: I mean look I'll say this. You don't have to like the president or not. He is an amazing marketer. And if he if his gut said, "Let's stand out from the rest of the world and try to get a premium." I don't think if I mean what is Sam Altman or or Dario, these are not great brand people. They are not great. He is he is he knows the market and so they should lean in and take it. They should trademark it.

Mary Palmieri: It's definitely a branding argument, right? But arguably we have to set ourselves up for the next decade. And so I guess that's where [clears throat] SI comes into play.

Rob May: Well, I posted that we're we're not going to adopt on the show cuz doesn't go far enough. So super duper is better than super. So we're going to call it super duper intelligence.

Mary Palmieri: That's what we're super duper intelligent. Got it.

Stephen Messer: The greatest intelligence. I'm surprised he likes the word greatest. I'm shocked. But he went with greatest intelligence. And he even he even gave the it's sigh is the way to pronounce it which I loved. I think he knows his audience. He's a Twitter man. Yeah. And the world lives on Twitter. So, yeah, why not?

Rob May: Yeah. Unfortunately, the world does live on Twitter. Twitter and Amazon are the two things I've tried to give up. I actually, so coming into 20 25, so almost two years ago, my New Year's resolutions was I was like, I'm going to live without Amazon for a year. And I started to try to untango and I was like, oh, there's no way. Like just [laughter] they they've got me. Like Alexa is addicting now that you can talk to her. Yeah. [laughter] And also I don't have so I don't have I don't have an iPhone so I don't have Apple Music. I don't use Spotify and I don't use Netflix. So I'm Amazon Prime and Amazon Music is my like music and so they've got everything and I was like ah I have to move it all over. But yeah when you hear like who are these people that still are signing up for Netflix because I've never had an account.

Stephen Messer: Wow. So it's all about the different streaming. By the way, I can't believe you haven't brought up the battle going on with Venode.

Rob May: That wasn't on my list to talk about this week.

Stephen Messer: I know that. I was shocked. It happened today. I mean, everyone's going crazy.

Rob May: I was surprised. Yeah. We can if you want. Do you have an opinion?

Stephen Messer: I just can't believe an investor who's got money in both players is picking sides. You know, it's I just thought it was a crazy battle.

Rob May: There there are people who I mean there's there's two types of VCs, right? Some are like, "We would never invest in a company that is competitive." And some people are like, "Hed our bets."

Stephen Messer: Hedge your bets.

Rob May: Yeah. Because he picked his baby because in the in the public markets, people wouldn't care so much, right? You could you could buy AMD and Nvidia stock and nobody would care. But it's Yeah. It's very

Stephen Messer: And the fact that he did a Sophie's choice of picking one of his investments over the other. I mean, the reality is he should allow he he should be putting up for secondary, get out, and no longer be an investor there. You're not a neutral factor. You're hurting your investor and you're hurting your LPs. You know, I know he likes to be in the center of the attention, but yeah, still I think you know, that's a as a founder makes you want to think, do I ever want to take money from this guy? Yeah. No different than Benchmark really threw Travis under the truck. Yeah. You know, does anyone really want to take money from Benchmark?

Rob May: Good call back to your article. You mentioned that in the article, David. Yeah. But would you want to take money from them? I mean, I would think twice. I think I've heard Bill Gurley talk about that. Yeah. About the Excuse me. The difficulty and the decision, right?

Stephen Messer: The difficulty. I mean, he went all on I mean, he went out for them. I mean, Yeah. I mean, hired people to put, you know, Travis's name under, you know, make him look bad. I mean, I just Why do you want investors like this? Who wants an activist VC?

Rob May: Yeah. I don't remember the specific things they were so concerned about, but there were a lot of bad stories about they wanted him to go public and he didn't want to go public and I think he was giving them the option of feel free to sell secondary and they they felt it they would make more money if it went public. Yeah. But I mean they were also concerned about like you know they had god mode where people that work there could track their significant others and stuff in Ubers and things like that. There was there were a lot of concerns in that company

Stephen Messer: and I don't and the company grew in how many years from nothing to something. Name a company that hasn't had problems before. Yeah, you get rid of the founder and CEO. I mean, my god.

Rob May: Yeah, but there I there's still lines that you don't want to cross, right? Think about like the ZITs stuff that you where they were they were selling insurance, I guess, in states where they weren't licensed and they just didn't care. They pretended to be licensed.

I don't remember all the details, but I will tell you my my worst and anti-portfolio story is my first company was backed by First Round Capital and so was Uber. And so when Uber was just in 2010, they were launched I think in San Francisco and like I don't even know if they'd launched another city. They were planning maybe Boston or New York or both or whatever. And I met Travis at a first round party. And he was explaining to me how it worked. And I was like I said, "Dude, why would I ever use this? I know where all the taxi stands are." [laughter] Like that was my

Mary Palmieri: There's a lot of Uber stories like that. I feel like why would I get into a stranger's car? Yeah.

Stephen Messer: Yeah. It's funny. I want to go back to this cuz you can see a justification for I don't see why the VC if the VC brings it up and the founder says no there are other issues or other ways dealing it but they also could have just gotten out of the deal. They could have said look I'm going to sell my stake in secondary.

Rob May: No no no I get that but like so how what what's your line for how bad of a thing does a founder have to do before you go the VC should replace them?

Stephen Messer: Legality if they if they've crossed the line of something truly legal. Yeah. And they've said I don't care. Okay, fine. But I'd ask a question of a public company because there have been quite a few that seem to have crossed lots of lines consistently like every bank in this country and I don't see Jamie Dimon getting thrown out. Yeah. And they've had numerous issues.

Rob May: Yeah. I'm saying like I don't know the whole story, but my understanding is that Travis crossed a lot of lines.

Stephen Messer: I think that story was absolutely put out. I don't see I don't think I've seen anyone go after him, including all the states that wanted to go after him.

Rob May: Yeah. No, that's a good point.

Stephen Messer: So I think my question is more why do VCs have to stay in deals if they're unhappy? Please go.

Mary Palmieri: Oh yeah. No, please leave. Please exit the room. I mean I've seen it happen. I've seen I've seen it happen through sometimes it's forced to happen through like CPHAS and stuff, right? I've seen it happen where they have to devest the stake because maybe it's foreign entity relations stuff.

Stephen Messer: So yeah, but I'm I'm asking generally. I think I think the venture world is structured where they're long only and I think they believe their value add and [clears throat] the reality is 99.9% of the times have you ever had a value added investor sitting at a table I mean if we're being honest have you choose cards right I you had a value added or you have net basically net neutral right the discount rate is I had to pay their lawyers to do the deal so there's a discount rate

Rob May: I can tell you yeah I can tell you stories on both sid I mean I've had a few moments where I was like oh wow that was like really so I give you example my you know my first company we did backup for cloud computing applications like Google apps and we got locked out of so you can imagine Google on our biggest signup day ever Google suddenly sees this single API sucking all these people's Gmail out and that's what we did that was the business model suck your Gmail out and put it someplace else and they locked us out so we had paying customers and we were locked out because they they was like fraud fraud and Josh Kopelman who was an investor was like oh I know the VP of Gmail. Let me text him and in like 90 minutes we're back. So this is like great operator, but there are

Stephen Messer: Yeah. Well, that's a good point.

Rob May: And yeah, but I can also tell you stories where I've had to be like, but you want us to do what? That's like sounds terrible.

Stephen Messer: Well, I'm not saying anything bad. There are great investors that are out there and they can be value ad. Yeah, my point is there are exits and what happens is if a VC decides that they you are not the dour thing of the day, their approach is you have to go. Yeah. And they're an investor of the two who are supposed to go. Who is the person who should go who is by their nature going to go at some point as opposed to the one who usually sticks around? And I think that's probably that's a bit of a of a of a mistake both for founders and I think also the model of being long only I think might be the dumbest invention ever.

Mary Palmieri: But to your point, I mean, shameless plug, that's kind of why I started MM Ventures because I saw that even in the early stages, I didn't have much operational support. And it seems very disconnected where it's like here's here's the money and then you have a lot of young founders who really don't even know how to appropriately operate and scale a company, too. So, I feel like more investors should be that. And I know it's very controversial like people say if you're investing you shouldn't be an operator and then I would agree.

Stephen Messer: I would differ. I would actually argue the founder is supposed to run their business. If they need help, go get an assistant. But not from the investor. The investor is not coming like a PE firm with an operating partner. They're not sitting there saying, "Let's figure out how to make this work." They are investors. I'm okay with that. My point is investors change their mind all the times about investments and they should get out. And my challenge that I have is with with the long only venture is they sell you on value ad which is why you're like why aren't they helping me? They're not value ad. In fact, if you read any of their blogs, they say as soon as the way you win is you the way you make returns is you focus on your winners, which basically means they're abandoning everyone who needs help.

The winners do not need their help. The winners are perfectly fine on their own. Like trust me, the venture guy who shows up at the board meeting is not adding an insight that's helping that founder. The one who actually needs help whether they're raising money or they need to get better employees or whatever is they're the ones the VC is publicly saying you just abandon those people which by the way means they'll never raise another round because if they're not coming into the round they've crushed it without realizing like they weren't neutral by abandoning them they've actually pushed their neck under the water so I don't get it at all that's why I said let them go bring in a new group let them sell out yeah and if they if they're not interested in the business take 10 cents on the dollar take cents of the dollar. Your expectation is it goes to zero. Give the guy a fighting chance to save his company.

Rob May: Yeah. So, cool. [sighs and gasps] Well, we'll wrap up on that. We should have a whole show on this topic. It's a really good topic.

Stephen Messer: Look, I'll do it anytime you want because I think I think that I think that market is getting heavily disrupted. I think venture is going away.

Rob May: You think there's a better model, but many times

Stephen Messer: [laughter] many better.

Mary Palmieri: I'd like that question answered in the next show.

Rob May: Yeah. Mary, thank you for being on. If people want to contact you or learn more, what should they do? What's the best?

Mary Palmieri: You could head to mmvententures.ai or find me on LinkedIn or Instagram and

Rob May: All right.

Mary Palmieri: Reach out for all your agentic commerce questions.

Rob May: There you go.

Stephen Messer: Well, I'll say if anyone wants to connect with me, you can go to intelligence.com. You can get access to my network and connect and I'm happy to share my network in exchange for yours.

Rob May: So awesome. Thank you guys for watching. Please leave comments. We love the stupid TikTok comments you guys you guys leave sometimes. And yeah, rate us, like us, tell your friends, subscribe, and we'll see you next week. The next