Episode 151: How Publishers Can Get Paid for AI, with Matthew Goldstein

Summary

Episode Overview

Title: How Publishers Can Get Paid for AI
Podcast: Marketecture: Get Smart. Fast.
Host: Ari Paparo (with co-host Eric Franchi)
Guest: Matthew Goldstein (MSG)
Release Date: December 5, 2025

This episode unpacks the critical question facing digital publishers in the age of large language models (LLMs): How can publishers get paid when AI systems ingest and reproduce their content? Industry veteran and “publisher whisperer” Matthew Goldstein (“MSG”) joins Ari and Eric for a deep dive into the evolving mechanisms for content licensing, the current power dynamics between tech giants and content creators, and the nascent marketplace efforts that could reshape the future of publishing and AI.

---

Main Discussion Themes

1. The Death of the Publisher Newsletter and the Shift to Social (03:24–07:14)

• Matthew Goldstein shares why he stopped his in-depth “What I Saw Happen” newsletter, explaining the lack of feedback and high cost through Mailchimp, in contrast with the richer interactions he now gets on LinkedIn.
• Quote:

  “I could do a post today [on LinkedIn] and I get hundreds of comments back. And it feels like it's so much more immediate with what's happening in the industry...”
  — Matthew Goldstein (04:14)

• Conversation on the difference between “owning” an audience via email newsletters vs. “renting” one via social platforms, and why Goldstein just doesn’t care anymore given his career stage.

---

2. AI and the Publisher Content Crisis (09:19–11:45)

• The hosts frame Goldstein as “the most connected man in publishing” and probe his current read on publisher challenges.
• Goldstein notes publishers are less worried about traffic drops now and more obsessed with LLMs scraping their content without compensation.
• He’s “all in” on solving the problem of how publishers get paid for content being consumed by AI.
• Quote:

  “How do publishers get paid for the content that is being stolen by all the big foundational models over the past three-plus years? … That’s all I care about.”
  — Matthew Goldstein (11:36)

---

3. Three Solutions: Upfront Deals, Legal Action, and Marketplaces (11:45–14:32)

A. Upfront Licensing Deals

• A small number of highly publicized deals (e.g., Reddit, New York Times) — often between tech giants with personal relationships.
• Many such deals are expiring or not being renewed; questions about sustainability.
• Quote:

  “We also joke around that they did the first deal and the second deals will never happen… publishers are trying to figure out what do I do now? How do I make up for that lost revenue?”
  — Goldstein (12:32)

B. Blocking & Technical Defenses

• Growing use of tools like Cloudflare, Akamai, and Fastly to block AI bots from scraping.
• Inconsistency among publishers; only about 22-23% of known AI bots are being effectively blocked.
• Memorable anecdote: Goldstein recalls hitting an actual red button at the World Trade Center to start Cloudflare's blocking of bots (15:03).

C. Marketplaces for Real-Time Content Licensing

• The “big idea”: AI/LLM developers and vertical industry players (like banks, retailers) would pay to access specific pieces of content in real time, using an exchange/marketplace model.
• Tension: Lots of seller (publisher) willingness, but little proven buyer demand yet.

---

4. The Future of the Web: Human vs. Agentic (16:44–19:27)

• Goldstein predicts a bifurcation:
  • Human Web: People directly visiting websites (skews older).
  • Agentic Web: Personal “agents” (LLMs, bots) aggregate and deliver individualized content feeds.
• This shift could dramatically increase the frequency and volume of content accessed, though most will be machine-mediated.
• He posits that “quality” will be directly tied to the richness and accuracy of licensed content, giving business momentum to these new marketplaces.

---

5. Deep Dive: Mechanics and Challenges of Publisher-AI Marketplaces (20:06–29:30)

Key Insights

• Number of Marketplaces: Goldstein tracks at least 22 groups trying to create such exchanges; the two biggest are Microsoft and the IAB.
• Buyer Problem: Many sellers (publishers) but few buyers (LLMs, tech companies).
  • Quote:

    “The problem with these marketplaces — they’re like dating marketplaces. Lots of sellers, very few buyers.”
    — Ari Paparo (20:22)

  • Goldstein and Franchi agree that without robust buyer interest, marketplaces struggle.
• Verticalization: Niche LLMs (e.g., in finance, commerce) could be a way forward, with vertical-specific platforms willing to pay for timely, high-quality data.
  • Goldstein’s scenario: Imagine JPMorgan Chase’s app delivering personalized, paywalled news from major financial publishers, all licensed via a content exchange (22:06–23:09).

Monetization Mechanics

• Real-time payments: When an LLM or NLM (niche language model) answers a question, it pays for each article referenced in real time.
  • E.g., “Should I sell my Bitcoin?” triggers the LLM to pull the latest relevant articles and pay per use. (25:32–26:47)
• Comparison to Existing Models: This approach differs from Yahoo Finance's traditional bulk licensing by introducing per-article, real-time compensation.

---

6. Implications for Publisher Business Models and Advertising (29:36–30:49)

• Traditional ad-based revenues will shrink as AI-facilitated experiences replace web traffic and ad impressions.
• Direct payments from marketplaces could offset declining ad dollars—but only if buyers materialize and content quality is demonstrably better than what LLMs can scrape for free.
• Quote:

  “I think [the ad business] gets really, really small. I think it’s like newspapers… it’s probably older people who have not evolved to using an LLM for their news and information. And it’s happening faster than we’ve seen.”
  — Goldstein (29:43)

---

7. Who Will Win: Google, OpenAI, or New Market Entrants? (32:05–35:32)

• Google is in a singularly strong position:
  • Universal access to content (search, Gmail, YouTube, etc.).
  • Massive user base (Gemini’s MAUs up 200 million in 3 months; 44:23–44:48).
  • Execution is the only barrier; Sergey Brin’s reported full-time return improves outlook.
• OpenAI seen as distracted, with recent “Code Red” moments indicating recognition of Gemini/Google’s competitive leap.
  • Quote:

    “Chat really started to know me, got really good. In the past week or two, something happened… personalization went away. The answers were not as good… I think OpenAI lost control of the core LLM and something changed.”
    — Goldstein (42:20)

---

8. Strategic and Legal Uncertainties (31:37–32:33, 46:35–46:41)

• If paid answers become the new model, it’ll impose a radical cost of goods on AI players (especially Google).
  • “It basically creates cost of goods sold for Google that don’t exist right now.” — Ari Paparo (31:37)
• Legal battles and competitive blocking remain unresolved—expect shakeout, consolidation, and a reduction of publishers as only the best content gets paid for and accessed.

---

9. Predictions (32:08–32:37)

• Goldstein predicts publisher marketplaces will be sending significant checks to publishers by end of 2026 (or maybe sooner).
  • Quote:

    “They might be getting checks sooner, but summer, end of next year, yes.”
    — Goldstein (32:33)

---

Notable Quotes & Memorable Moments

• “How do publishers get paid for the content that is being stolen by all the big foundational models over the past three-plus years?... That’s all I care about.”
  — Matthew Goldstein (11:36)
• “We also joke around that they did the first deal and the second deals will never happen.”
  — Matthew Goldstein (12:32)
• “The problem with these marketplaces — they’re like dating marketplaces. Lots of sellers, very few buyers.”
  — Ari Paparo (20:22)
• “I believe there’ll be two webs. There’ll be the human web and the agentic web… before you know it, Google and Gemini will be doing all that for you.”
  — Matthew Goldstein (16:44)
• “If you think about it, in five years, what do you really need? An agency? Sad.”
  — Matthew Goldstein, on the future of media agencies (37:57–38:28)

---

Timestamps for Key Segments

• Newsletter Death and Platform Shifts: 03:24–07:14
• Publisher Pain: AI Ingestion & Revenue: 09:19–11:45
• Upfront Deals & Tech Blocking: 11:45–16:33
• Human Web vs. Agentic Web: 16:44–19:27
• Marketplaces for Content Licensing: 20:06–29:30
• Advertising vs. Direct Content Monetization: 29:36–30:49
• Strategic Implications for Google & OpenAI: 32:05–35:32, 42:02–46:41
• Predictions and Outlook: 32:08–32:37

---

Conclusion

This episode argues that the publisher business is at an inflection point: ad revenues are declining as LLMs siphon off content value, yet an emergent set of real-time content marketplaces could finally deliver meaningful compensation—if buyers materialize. The path forward includes technical defenses, industry coordination, and reimagining the very mechanics of content exchange for the AI era.

Goldstein is bullish that agentic content consumption and direct payments will take over, leaving traditional web traffic and advertising as “old people’s habits.” The question remains: will publishers get paid enough, and will there be enough survivors, to keep the web’s content engine running?

---

For full context, subscribe to Marketecture and follow Matthew Goldstein on LinkedIn for real-time perspectives on publisher/Ai trends.

Transcript

[A]

Adlib is the DSP agnostic buying platform that helps agencies and marketers scale Programmatic, smarter, faster and more profitably run premium campaigns across CTV audio, geofencing and YouTube all under one roof with zero trafficking headaches, no bloated fees, no complexity, just enterprise grade performance built for growth. And Ad Lib isn't just a platform. They are your partner. Their team helps you win new business, expand budgets, improve ROI that keeps clients coming back. Unlock programmatic and get ad lib.com g e t a d l I b.

[B]

Com.

[C]

This podcast is brought to you by Adelaide Media verification and measurement are undergoing major disruption. Legacy players are pivoting to performance advertising, AI is reshaping brand safety and attention is replacing viewability. Adelaide is leading the shift with au, a new way to assess media quality that scores placements based on their potential to drive attention and outcomes. Before your ads run, think of it like a credit score for media. Finally, a clear view of quality. Before you buy, take the guesswork out of your investment strategy and try Adelaide AU on your next campaign.

All right, welcome to the Market Extra podcast. This is Ari Paparo. I'm here with Eric Franchi and our guest this week is Matthew Goldstein, otherwise known as msg. Eric, do you know MSG or do you know of MSG in his newsletter?

[B]

I know of MSG and his newsletter. I was a subscriber. I guess I am a subscriber. He stopped sending them and he moved over to LinkedIn. But man, those were the longest and most detailed emails around everything publisher that I think ever existed. And he is all in on publishers. But what's interesting is his focus is now on AI and its effect on publishers. So it's fascinating.

[C]

Exactly. Yeah. I had coffee with him maybe a month ago and he said, I don't care about ads anymore, I want to talk about AI. And I said, okay, why don't you come on the POD and talk about AI? He's really involved in this question of how do publishers get paid for their content. So we're going to hear about the emergence of these exchanges where publishers are asking AIs to pay them as well as upfront deals and maybe we'll touch on the legal side of things. So it should be an interesting conversation. He's also a very fun guy and will probably go off topic a lot.

[B]

Can't wait. Anything to announce this week?

[C]

No, not really. If you're not, you should be getting our newsletter at News marketecturetv. You should also subscribe to the Refresh newsletter which is written by AdTech. God. And AdTech Explained had a really good piece this week on.

AI and what was it about? It was about AI and how they are putting ads or in Google AI searches. So just a reminder that MarketActor's got a lot of great content and you should subscribe to it if you're not. So with that, let's jump in with MSG about AI and content licensing.

All right. Matt Goldstein, msg. Thanks for coming back on the show.

[D]

I had no choice, Ari. Thank you very much for having me.

[C]

Yes, exactly. No choice. All right. We have to talk about some logistics first. Okay, first question. Everyone wants to know what happened to the newsletter? The Things I Saw newsletter which had, you know, you were paid by the word. It was long, it was in depth, what happened.

[D]

So it was called what I Saw Happen. Okay. I used Mailchimp, which I despised. Mailchimp would cost me like $90 a month. I sent a newsletter out once a quarter and I just stopped doing it, partly because I was getting no immediate reaction. I would find out who opened it, get maybe I would send it out to three to 4,000 people. It would get a ton of pass alongs and I'd get like 10 or 15 comments at the end of the day. And then I started putting stuff on LinkedIn.

[E]

I could do a post today and.

[D]

I get hundreds of comments back. And it feels like it's so much more immediate with what's happening in the industry. I actually remember it. I wanted to use your profile and beehive because I hated spending $90 a month for a mailchimp and you told me no.

[C]

Come on. That's probably not what happened. I wouldn't say no. I have. Maybe we didn't upgrade. So now we have a like gold plated beehive account. We could probably throw you on there. Maybe we'll just acquire it. Maybe it'll be the newest architecture acquisition.

[D]

The.

$1, it's yours.

[B]

It's true.

[D]

I get $1.

[C]

$1.

[B]

I was a subscriber and I always, I always wondered if everybody had the same user experience as me. It was so long, it was long that like I would have to hit the like. Continue reading on web. You just dumped and it was all good stuff, Matt. Like you just dumped all this content in it and I miss it. You should bring it back. We should make it part of architecture. I will represent you in the transaction. Let's go. Let's do it. $2.

[D]

Okay. Eric. Hello, Eric. It took me so much time to put that together. I would Literally have a running list in a Google Doc, and every day I would add something to it. And that's why it grew so long. It was 90 days of data.

[C]

You know, it's really true, though. As a someone who writes a newsletter, I get very little feedback. Every once in a while, someone's like, oh, that article is awesome. But, like, you know, even. Even when people like it, they don't say anything. It's just not an interactive medium.

[D]

It would. It would drive me nuts. I would send it to people who I know well, and they would not give me feedback, or I would send it to people. And then five years later, they asked me some for advice. I'm like, I sent you a newsletter 30 times in the past five years. You couldn't thank you once. And now all of a sudden, you're asking me for advice, and I'm like.

[B]

Oh, I don't know.

[D]

It doesn't feel right.

[C]

The medium is the message. That's. That's the. That's the learning here.

[B]

But apropos the conversation, we're about to have difference between posting on social, which you've been doing great at, and a newsletter is on social. You are, quote, unquote, renting your audience. You are, you know, being reliant on the reach and the algorithms. The newsletter, you're always going into the inbox.

[D]

Okay, Eric, I don't care.

[B]

You're not. You're not a publisher. You're not a publisher.

[D]

Well, I'm 58 years old today. We're. This is December 4th, correct, Ari? And this goes live when we're recording Thursday the 4th.

[C]

It goes live on the 5th.

[D]

Okay, the 5th is my birthday. I will be 59 years old.

[C]

Nice.

[D]

Happy birthday. Happy birthday. You want to come to my party? Ari? Ari, you're more than welcome to come to my party.

[C]

So on Tuesday, December 2, was my 55th birthday also. So we're having a little birthday podcast.

[B]

Happy birthday, boys. Matthew. I'm not invited.

[C]

What's going on?

[D]

You want to come back?

[E]

Where do you live?

[D]

Eric, where do you live?

[B]

Parts unknown. You?

[D]

I'll send you the invite right now.

[B]

So.

[D]

Wait, wait, wait, wait. My point being, at 59 years old, I'm at the end of my career. I don't give a shit, okay?

[B]

Respect.

[C]

It's like Logan's run. You hit 60 in the media business, they just kill you.

[D]

I love that. I love Logan's Run. One of the best shows ever. We're dating ourselves.

[C]

We are. All right, so second housekeeping item. So this week, Spotify wrapped came out. And it was great because a lot of people came out on Twitter saying, oh, man, MarketExter was our number one podcast or number two podcast after Joe Rogan, of course. So I'll take that. But do you guys, do you guys know who our number one fan is? And there's no doubt this is our number one fan because I get feedback from this person every week and I'll. The clue is, you know, this person. MSG Mark Polo. So this is a shout out to our number one fan, Nicole Pruitt, who.

Gives me a running commentary by text on all the things she finds amusing or wrong in our podcast. And so Nicole, I'm gonna get quite a text tomorrow morning when she listens to this. But you h her originally, didn't you.

[D]

At Viacom back in the day, like, you're talking 2001, 2002.

[C]

You're responsible for bringing her into our industry.

[D]

I'd say, well, you know, my job right now is just a big connector.

[E]

And all I do now is connect you.

[C]

Yes, I'm aware of that. So. So shout out to Nicole and shout out to everyone who, who saw us on your Spotify Wrapped. You know, feel free to continue putting that on social. It's great. Great for us. Eric, you got, you got some shout outs on your other podcast, your four other podcasts, right?

[B]

Yeah, yeah, it was really great to see, but my personal Spotify raft was quite enlightening. I don't know if you guys felt the same.

[C]

Yeah, I haven't looked at mine yet, but I'm confident. It's like basically Taylor Swift and Soundgarden, which is a weird combo, but it's the way I roll.

[B]

Yeah, it's fascinating. They aggregate all of your listening and they come up with some pretty interesting findings. I recommend everybody find it. Your listening age is a new one this year, which is kind of fun. People are posting it. I'm 40, but I listen like a 60 year old or totally.

[C]

Yeah. And shout out to Spotify's marketing team because they turned the Empire State Building green yesterday for Spotify. Yeah, that was a good shout out. That's, you know, got to be on the bucket list of most marketers. Okay, we're not here to talk about this stuff. We're here to talk about AI, which is our favorite topic. So, msj, you're known as the publisher whisperer. I put in our show notes. You're the most connected man in publishing. No, no, no, no.

[D]

I am the self proclaimed second most connected person in Publisher AI who's number one.

[C]

Who's that?

[D]

John Roberts from People A. Oh yes.

[C]

Yes, A friend of the pod. Of course he is because I talk to a lot of publishers and they always reference him. He's a real thought leader, PhD in rocket science, et cetera, et cetera. Speaker at the market events. Okay, you're the second most connected publisher whisperer. So what are you hearing? What's, what's going on? What's the vibe?

[D]

Well, you asked me one question about is traffic down or not across publishers and we've all heard anecdotally it is. And I spent a lot of time thinking about this and I'm working with Criteo to try to put together a definitive point of view on that and similar app says traffic down and you hear individual publishers saying they're down 10, 20, 30, 40%, whatever the number is.

[E]

But I'm also going to reach out.

[D]

Or if they're listening to the Talk to Parsley or chartbeat and find out if they can actually tell us if they're down or not. Because they know for sure as a benchmark more than anyone or similar web it's sort of what you have to put a little piece of code in your PC and they sort of capture stuff and they put some AI on top of it. Like I want to know for sure. And I feel like chartbean and parsley are the two bets by Google or GA4, which will never.

[C]

Yeah, it's, it's all non representative samples. Right. So they're all trying to estimate based on that. But if when you're talking to individual publishers, are they coming to you saying oh crap, we're down x percent.

[D]

I don't hear the oh crap, we're down x percent. I hear oh crap, the LLMs, the foundational models are taking our content. What do we do about it? Like I am all in that. That's the conversation I'm having more than oh crap, the graphics down we've ordered. We sort of move beyond that. And, and, and so the, the more thing is like how do my big thing is how do publishers get paid for the content that is being stolen by all the big foundational models over the past three plus years? Stolen Ari in the past two years. That's all I care about.

[E]

I don't care about.

[D]

So if you ask me any ad questions, I will do not know the answers.

[C]

No, we're not here to talk about ads. We're talking here to talk about AI. Okay, so there's three dimensions that I want to talk about. First is the upfront deals that people are making to license the content. Second being the legal angle, which is let's sue them until they pay us. And the third being the concept of real time exchanges for content purchasing. What order you want to take that in?

[D]

Let's talk about the first one.

[E]

Go first. Okay.

[C]

Upfront deals. So a lot of headlines. Reddit, $30 million, New York Times licensing to Amazon for their shopping stuff. What's going on? What's the state of the upfront deal market?

[D]

So whenever we get together, no one talks about the specifics of these deals because we never anything confidential. So we sort of know what we read in the trade.

You know what OpenAI did 10, 11, 12, 13.

With most of the big players. We joke around that Dan Altman did deals with his billionaire friends who he knows to sort of make sure nothing bad happens. We also joke around that they did the first deal and the second deals will never happen. I think some of them are expiring soon and some publishers are trying to figure out what do I do now? How do I make up for that lost revenue? Where does that come from?

[B]

That's a great point. We haven't heard about one of these deals and feels like months because we would talk about them. They were coming hot and heavy and we were talking about it seemingly every. Every week. There's been nothing.

[D]

Well, so what happened two days ago with Sam Altman and the Code Red inside the company? Right, right.

[B]

Yeah, we'll talk, we'll talk about that later on. Maybe that's the reason why.

[C]

Is that because they're running out of money or they don't want to do the deals?

[D]

My hypothesis is that they realize the quality of chat, which all the kids call it, is not as good as it used to be, that Gemini has starting to be much better in everything. And I think Sam and team got distracted with ADS and Pulse and agents that they realized we're too distracted. We need to make sure the foundational model works really well. And the foundational model was not doing as well as Gemini. When Gemini did their Code Red a year or two ago and figured it out, my gut feeling again, I know nothing. I'm just hypothesizing is that those chat does OpenAI are scraping again to try to get new information for their foundational model because what they had in there just wasn't working. It's not as good. Had they been blocked so much or do they actually need to do deals to get real solid information there with real metadata as opposed to scraping it and try to build the model a lot better. And as we know, Google has one bottle, right, where search and AI are combined. So Google has access to everything. Let me say it again. Google has access to everything. They have a distinct advantage in this entire ecosystem.

[C]

Right. Plus YouTube. Right. So what's the current state of blocking? Right, so Cloudflare put out their blocking a couple months ago. John Roberts, who we talked about, was a part of that initial rollout. Are publishers happy with the state of the blocking opportunities?

[D]

Okay, so I was there when we actually hit the red button on July 1st at the World Trade Center. There's a picture I can.

[B]

You can.

[D]

You can put that on the COVID of this market.

[C]

What red button? Was it an actual red button?

[E]

It was an actual red button. I will send you the picture later. It was like a little small one.

[D]

I think they bought from Amazon, like the Staples one. We actually hit the red button, the.

[C]

Red button that told Cloudflare to start.

[D]

Blocking stuff and really call Cloud Bear. But it looked good, right?

[C]

You know, because probably what happened is you press the red button and then a screen came up that said, are you a robot? You know, and then you had to click the checkbox, and then it worked. Right?

[D]

So what's happening? So more companies than ever are blocking the box. However, I think it's just really inconsistent across all the publishers. Akamai does a great job. Family does a good job. They're out there blocking. They've been blocking for a while. What I've been trying to get is the definitive list of who should be blocked, and that's really not around right now. Someone the other day jokingly told me it's easy to block. Just block any robot, any AI from China or Russia. That's the first step, right? But if you have deals with open AI, you can't block them. There's a site I know called Dark Visitors. My friend, he does a great job of looking at who's being blocked based on his data. The last time I looked at it, said that 22 to 23% of the top bots are being blocked to the AI bot. And to me, that just feels way too small given the number should be much higher. But I think it's very inconsistent about blocking and why they're blocking. And what I want to do, and it's on my list to do, is I want to bring together Akamai and Fastly and Cloudflare and Datadome and Cloudfront, bring them all together, and try to have a conversation about who should we block. Is there one list that's going to help us all block this in a better way?

[C]

Right, so tell us about, you know, how publishers are thinking about this. And you know, they used to get paid with ads, now their content's being shown on different sites. What, how are they adopting, adapting their model?

[E]

So I believe that the foundational publisher model of, you know, building traffic, getting Google to send you links, having ads, having direct ads, having programmatic ads is all fundamentally changing and it's been happening over the past year or two. I believe that there'll be two webs. There'll be the human web and the agentic web. The human web will be people like Ari, Ari or Eric going to a homepage of a website. It'll still happen. It'll probably be more older people who are used to that. And I think the agentic web is Aria. Eric, you say, I like these 10 publications. Here's my list and my agent. Then go to these 10 publications every day, get me my information I want, let me know when it's updated, give me a notification, send me a podcast in the morning, send me texts during the day. And then what happens is the traffic going to these websites increases by 5 or 10x. It's not you as a person, it's your agent doing it. And that's what I think ChatGPT tried to do and they got distracted and I think they might have screwed up their foundational model. But before you know it, Google and Gemini will be doing all that for you. And they have such a distinct advantage given they have access to all your content. They have gmail, they have YouTube, they have Google Calendar that it's going to be your one stop shopping for anything. And then they have to end up paying for this content itself some way. They can't continue to just take it. That is the big.

[C]

You end up in the sort of in the argument that Google News has been having for 20 years, which is Google News is like, hey, we drive you traffic. And then publishers like hey, you're stealing our content. Don't you end up in that same situation, which seems like it's a, you know, both two sides just don't understand each other.

[E]

No, I believe there's going to be a fallout. So if there is a 10 million publishers today, in two, three, four years there's going to be 50, 60, 70% less. And I believe that the content will be so rich and so good that the agents who are a marketplace will find a way to pay for this content. And what'll happen is Everything will be blocked. So well, that if you don't find a way to pay for it and get it with all the metadata, it's not going to be as good and you can't answer the questions well. So one thing that I want to see, I would love to see is here's your five foundational models. Here's some objective third party, could be Jonah Goodhart from Mobian that says, I'm gonna grade a thousand answers a day and how everyone does in asking these thousand questions. And they all get a score. And right now, if you got a score, score would be a C, a D or an F for probably every one of the foundational models. How do you get it up to an 80 or 90? One of the answers is give it access to all the content in real time with the metadata. And everything else instead of everything gets better and better.

[C]

Okay, so let's talk about the emergence.

[E]

Quality and it's the right answer instead of the crap we get now. We accept it.

[C]

Yeah, I know, I get that. But let's talk about the emergence of these potential publisher content marketplaces. Right. Where the AI people would be paying in some way. And I want to talk about the mechanics, this current status of these efforts, like what the ideal state is, because there's a lot going on here, by my, by my estimation, and maybe you definitely should know better than me, there's like two big ones that are getting attention. There's Microsoft and there's the iab, and then there's a lot of like private companies that are trying to create marketplaces.

[E]

So I have, I have a list in front of me of 22.

[C]

22. Oh, my God.

[D]

Okay, 22.

[E]

Because what happens is every time I post on LinkedIn, I get all these people coming to me and saying, I have, I have a marketplace. I have a marketplace. Sure, I'll talk to you. I have a list of 22.

[C]

Okay, so the problem with these marketplaces, they're like dating marketplaces. Lots of sellers, very few buyers. Right. You know, every publisher will be happy to get paid for their content. Right. But you need the demand side to want to pay for it.

[E]

Correct. So a couple things. So, you know, the, the big one out there by the biggest one now is Microsoft. As you alluded to, it's called the Publisher Content Marketplace. Axios wrote about it in September. What's going on? I think there's probably three or four publishers who publicly said they're part of it. It was Gannett in an earnings call. It was people in an earnings Call. And I think it was Axel Springer, might be a couple more that you've heard of, but I are the ones that public. Right, Right. And Nikhil, who used to run msn, just got a big promotion. I send you the link and the in the notes, Ari and Eric, about him being part of it and taking a bigger job. So I think Microsoft is really doing a good job of trying to figure this out and trying to figure a way to pay the.

[C]

Pay the publishers and so let's talk about the mechanics.

[E]

I'll ask you one question first. I'll ask you, Eric, what financial app bank do you use?

[C]

Are you being fished? You're being fished live.

[B]

I don't want to be hacked. Okay, I'm not going to answer the question.

[E]

I'm assuming it's Chase.

[B]

Okay, no comment.

[E]

Okay, Ari, because you already told me this question. It's Chase.

[C]

So, yeah, so my password is.

[E]

So, Ari, how often do you log into your Chase app?

[C]

I'm not answering these questions.

[E]

I know the answer. You already had this conversation.

[C]

Do you have like Coinbase passcode too? I mean, come on, people, it's two.

[E]

To three times a day. And you log in to make sure everything's okay and no one stole your money.

[C]

Okay, that seems, that seems a little paranoid, but yeah, I'll work with it.

[D]

Sure.

[C]

Okay.

[E]

What if chase and a BL marketplace did a deal with deal with the 20 top investment websites out there? Wall Street Journal, Barron's, Investopedia, New York Times, all of them. And what if Ari and Eric, every time you log into your Chase app, you got the financial news that you wanted, okay? And it knew it all about you and knew what stocks you owned and knew your kids are going to college know so much more about you. And what if that became your way of getting financial news? And what if the way Chase did it is they didn't scrape the data, they went through a marketplace and they paid for the content with such great added value to give you all the news. So when you think about that scenario, like that's realistic, that's going to happen, right? Goldman Sachs will do it. Bank of America will do it. Morgan Stanley, they're all gonna, they're all gonna do it. And I think this is what happens across every other vertical to get these content, these, these, these LLMs or niche LLMs, like the small LLMs, to actually pay for the content the right way because they know scraping does not work and they know the board of directors of these big banks won't allow them.

[C]

To scrape anyone okay, so you, there are a couple of questions embedded in your little scenario that you've provided.

[E]

If Eric did not like.

[C]

No Eric, well, he's an investor, so he's not going to give you any money for this thing.

[E]

But like he's already offered me $2.

[C]

I'll let Eric jump in a second, but I just want to point out a couple of assumptions in that little bundle of conversation. So the first assumption you're making is this general. There are going to be a lot of LLMs, not one LLM, which you know, in the search market, it turned out not more or less not to be true. There are other search engines, like Amazon is a search engine and, and you know, Yelp is a search engine.

[E]

I think what I said is there'll be a lot of, there'll be five LLMs, there'll be many SLMs.

[C]

What's an SLM?

[E]

Small language models or NLMs. Niche language models going across articles. That's what I believe.

[C]

Right. So in your assumption this JP Morgan Chase has or some financial institution is able to create a finance LLM that is preferable to just using Gemini or ChatGPT to get your finance info.

[E]

Well, if, if Jonah Goodhart did the score of that and LM, it'll be like a 99 because it'll be right all the time.

[C]

Okay, the second thing I'd like to ask about is like, how is this different from Yahoo. Finance? You know, like basically. And what I mean by that question is like we have a history of content licensing that's in bulk. You know, Yahoo Finance does a single deal with, let's say Bloomberg, gets all this stuff fed into it and then does its own value add on it. Why in this world are we talking about needing marketplaces that are per. Per content piece in real time?

[E]

Okay, so Yahoo Finance, they do what, 20 or 30, 40 different deals? Is that what it is?

[C]

They have a lot of deals. I don't know.

[E]

I have no idea. But they do all the deals through one marketplace or create their own marketplace instead of doing individual deals. And that much easier to do than.

[D]

All these one UP contracts.

[E]

Just have a marketplace that does it for them.

[C]

Right. Okay, so let's talk about the mechanics. Because I believe the assumption under all these marketplaces, the assumption in all these marketplaces is that it's real time and per article, right?

[E]

Oh, well, there's a lot of talk about when do the LLMs pay for the content.

[C]

Yeah, right, exactly.

[E]

So there's a couple different scenarios. My favorite scenario is Ari asked a Question about. Give me a good finance question, Ari.

I want to sell all my bitcoin.

[D]

Okay.

[C]

Okay, sure. Is now the right time to sell my bitcoin? Sure.

[E]

And then you ask the LLM or the NLM or the small lm and it goes out and it says, okay, there are 80 articles in the marketplace I could use right now. Okay, so it sucks up 80 articles in real time, and maybe it pays everyone a little bit for those 80 articles, and then it uses five articles to give you the answer, and it pays a little more for those five. Okay.

[C]

Yeah. So I think the way John Roberts explained it to me, because I asked him a similar question, well, JR is.

[E]

Number one, so I'm a distant number two when it comes.

[C]

Yeah. But this is coming through my memory of his interpretation of things. So it's like, kind of dulled down. So you're taking John Roberts brilliance and putting like a gauzy lens on front of it and making it dumber. But what I. The way he explained it to me was like, sure, you have this model that's been trained in advance with all this content, and maybe it's licensed and maybe it's not, whatever. But when you ask that question about, should I sell my Bitcoin now? You have to get the last, like, seven days of content that's rich to put it in the rag and then. And then do do interpretation on that content to give you a really sharp, fresh answer.

[E]

I think I said the same thing where there's a foundational model, it has all the information. You ask it the question, it looks at a hundred articles and it picks five or six or seven to give you the best answer in real time to answer your question. And you probably should sell it already.

[C]

So, Eric, I think you've been chomping at the bit to jump in here. Do you want to like. Or I guess I'll just ask you a specific question. Have you invested in any of these 22 marketplaces that are on MSG's list?

[B]

I haven't. I'm super interested in this idea in this category, by the way. So at the risk of now my inbox starting to get, you know, pelted by all those people, you know, looking to, To. To raise money. I think this is fascinating. And I guess my question for you, Matthew, is are there any companies that are interesting and like, you know, doing it right and breaking out? Because this needs to exist in the world. I just don't know if this is going to be like Microsoft or this needs to be kind of Rebuilt and there's going to be another one or there's many. Like this is absolutely something that needs to exist.

[E]

Again, no one knows. I believe there'll be many marketplaces out there. I think there'll be some big ones that Microsoft or Google or Amazon controls. I think there'll be some smaller ones, there'll be some niche ones. Right. Snowflake's developing something in the financial sector. You have Tollbit out there. You have Data X which has X financial people and DVR roots. Cloudflare talked about something, you know, Ozone talked about something. Mozilla, there's just Geodisc. My friend who does it from impact.com was doing one on affiliate. Like they're all out there, they're up but their problem is do they have the buyers? Like it's easy.

[C]

Yeah, the buyers.

[E]

Do they have the buyers? Or Eric, will you give them money? When they come with this pitch and you're in a first question, be like who are your buyers? Do you have any buyers? And they're like no. It's like, okay, but if you give.

[D]

Em a lot of money, I don't.

[B]

Know if they can bring the buyers. Yeah, that's the, that's the interesting question. I've had a few conversations with these companies and all of them are saying yeah, absolutely. It's like so easy to get in the door with publishers and publishers will sign up and you know, obviously if it's a revenue share model they will sign up. But the question is, hey, like how's your conversation with OpenAI going? How's your conversation with Anthropic going? And then there's, you know, at this point there's no answer to that.

[E]

And my gut feeling is the conversations will not be with Anthropic or Open AI. Right. If you're doing, if you're doing a E Commerce one and your conversations are with Target and Walmart and Best Buy and CVS and all these other companies, it's probably a lot easier to do a deal with them than it is Amazon.

[C]

Right.

[E]

But I think if it goes down a level to niche LLMs or small LLMs, that's how the conversations change.

[B]

I gotta ask so that, that makes sense. I gotta ask. So like what does this mean for the future of the ad business of publishers?

[E]

Oh, I think it gets really, really small.

[C]

I think yeah, you're swapping ad for.

[E]

Ads for payments and I think it's like the newspapers. It's like it gets really small and it's probably older people who have not evolved to using an LLM for their news and information. And it's happening faster than we've seen.

[D]

Given everything we've been through.

[E]

Has only happened three years ago.

[B]

Right.

[E]

It all starts.

[C]

It doesn't.

[B]

You're so good. My question was going to be like, what is your time frame, time horizon for. For this? Ari, I'll shut up from here. I know you got.

[C]

No.

[D]

Oh, I think.

[E]

I think as JR says all the time, there's momentum, there's a lot happening behind the scenes. So it's all happening right now. It's just the question of, do you get those right buyers to come in and start changing the ecosystem and buying instead of scraping the. Scraping it harder? Or does someone realize that when I buy through the marketplace, the quality and the correctness of my data is so much better?

[C]

Right. There's a lot of weird strategic implications here if this were to happen, right?

[E]

Oh, no, it's going to happen, Ari. Well, I spent my whole career on it. I got like 10 more years to work.

[C]

I was just joking, you know, I mean, I get it. It's just so easy to spider and scrape. Right. And it's not illegal, so it's just a interesting problem. Like one problem. I could imagine. The example I often gave, especially since I talked to Mr. Roberts at People like, suppose you wanted to ask an LLM, like, who is Brad Pitt dating right now? They don't have their model trained on that. They have to get the information. But then you might say, okay, well, obviously People is going to be a great source for that kind of celebrity info. But it's also on 20 other sites. They've just copied people.

[E]

Okay.

[C]

And if you really just want the fact, how does the price maintain?

[D]

Stop right there.

[C]

Okay.

[E]

There's 20 other sites that copy people. In a year or two or three.

[D]

There'Ll be three or four or five.

[E]

Sites that copy people because other ones will not be.

[C]

All right.

Yeah. And then the other two interesting things. So, you know, it's interesting that Microsoft's doing this, because if you were in a world where the answers that AI delivered had to be paid for as you go, that's just a total dagger to the heart of Google's model. Right. It basically creates cost of goods sold for Google. They don't exist right now. Right.

Yeah.

[E]

It gets really complicated.

[C]

Right.

[E]

You're talking. The new World Order is all happening in front of us.

[C]

Yeah. All right, so let's just close this out with a prediction. So when do you see how many years or when. What's a Date on which you think there'll be, you know, a widespread adoption of one of these marketplaces Next year? No, like when? Like end of next year. You're saying by the end of next year, publishers will start getting checks from marketplaces.

[E]

They might be getting checks sooner, but summer. End of next year. Yes.

[C]

All right, let's call it on that. This is a fascinating conversation. We'll be back in a minute with the news of the week.

[F]

A word from our sponsors. Budgets are shrinking, AI is rewriting the rules, and the marketing playbook is changing fast. That's why possible is where you need to be. From April 27th to 29th in Miami beach, the most influential brands, marketers and innovators will connect to make deals, spark partnerships, and shape the future of marketing. If you're serious about growth, visibility, and staying ahead, visit possibleevent.com and secure your spot today. Again, go to possibleevent.com to secure your spot.

[B]

All right, we're back with the refresh that went longer than usual. So we'll, we'll prioritize our agenda here. But there's a bunch of news this week. We have M and A, we have AI and we have more AI. So a good conversation to have with Matthew here. So let's start with M and A. One announcement and one official deal closing. So Broadsign acquires Place Exchange. Place Exchange, well known of home SSP broadside, I think less familiar, but seems to be you're kind of putting together some interesting out of home assets. Two things here. The second acquisition this year of an SSP and the second acquisition this year of a scaled out of home ssp. So like we're not quite there in terms of the end of year wrap up. But you know, nobody had out of home as you know, the breakout category of 2020. But the deals are happening. What do you guys think?

[C]

I think out of home ads work, right. I think everyone knows that, but they've been resistant to really becoming a liquid market. They are very differentiated by their position, their screen size, they're creative, etc. Etc. So it's remained kind of fragmented. And so that combination means that it makes sense to own one for certain buyers, but it doesn't scale well enough to have, you know, billion dollar outcomes and anything like what we see in display ads, mobile ads, with applovin, et cetera. We're a while away from being able to be like a buyer, have an easy button where I'm like, hey, I want to reach these consumers across the world. Press a button and Suddenly I can do it and get outcomes. We're so far away from that. So it ends up being kind of a business of niches, many of which are profitable.

[E]

I think as digital advertising ecosystem and number of impressions decreases and people know that that out of home becomes even more important. Not going away.

[B]

It's great.

[C]

I mean out of home really works, right? It's just, it's just hard.

[B]

That's so interesting, Matthew. So like you're not the first person that said that where it's just like, hey, you know, if the open web quote unquote, you know, starts to decline, those dollars need to find a place, right? And you know, if you can start to put together some real scaled opportunities in things like out of home audio, obviously CTV is already there and they become liquid to your point, Ari, they become measurable. Like you can start to see maybe some non obvious growth areas in ads. So this is super interesting. And the first step is making it available at scale and liquid. And you maybe start with the supply side which is happening here with these two deals, right. Vistar and T Mobile and Broadside and Place Exchange. You can buy this stuff programmatically on ttd, right?

[C]

Yeah, I think you have to start on the supply side because it's so fragmented. You need to take control over the screen size and the experience and the creative in order to get scale. That appeals to the buy side.

[E]

Yeah, we could have gone on for hours. One other point though. If you're buying out a home, you're never going to buy a bot, right? When you buy publisher traffic right now. Yeah, there are bots in there.

[D]

You buy out at home, you're buying real people.

[C]

Well, you're buying, you might be buying muted people who aren't watching the screen though.

[E]

Okay, but you're not buying bots.

[C]

Okay, Is it better? Would you rather have a person not paying attention or Bot.

[E]

That's the name of the podcast.

[B]

All right.

That'S funny. All right, well we'll continue to watch out of home. That was again, not on many people's dance cards. Okay. M and A number two. This is not an announcement. This is just kind of the finalization. Omnicom acquires IPG officially and is officially now the largest holdco. And a lot of the conversation on the Internet this week was twofold. The first was the inevitable job cuts seem to be happening. There's 4,000 job cuts that were being planned and then this sort of is not my world. But man, there's a lot of people that are pouring one out for all of these legacy names, like agency names that are being sunsetted. Like there's a lot of people that are very upset about DDB and some others. Like it's a, it's a real, it's a real conversation happening out there. You guys paying attention to this?

[E]

I, I think the, I feel so bad for the agency people, the agency world. I think they're just, I think they're getting left behind. Do you really need them long term? You know, do they make great creative. Can you make creative easier? The way they buy just the whole model, it just, just like publishers. Are the models changing? I think the same thing happened the agency model. I think it's been going on for a while and just gets smaller and smaller. Yes, you still need amazing breakthrough creative. But if you think about it, in five years, what do you really need? An agency. Sad. Especially because New York is such a mecca for the agencies to be. To be here.

[C]

Where will all the humanities majors get career options? It's just a real problem.

[B]

All right, what's your take on what Matthew just said? That was, that was almost verbatim Sam Altman from that quote a couple of years ago.

[C]

I don't believe in the agencies going away. I think AI will reduce the headcount requirements. It will be smaller but with the same economic impact likely in the future. I'm of the, the, of the tobacco wallow approach. They're like cockroaches. They figure it out. But you know, shedding a tear over some brand name is not my thing. Like if agencies are good at two things agencies are good at. Number one is coming up with new names for their own agencies. And number two is figuring out how to give some random mid level manager a CEO title.

[D]

Or, or, or.

[E]

The other thing is the agencies still get hired by the cmo. So if something goes wrong, the CMO fire the agency and doesn't get fired themselves. Themselves.

[C]

Yeah, yeah. It's a battle of who gets fired first.

[E]

Yeah.

[C]

All right, I'm going to ask you to move on, Eric, because I just can't take agency talk much longer. I just, it makes me want to turn off the podcast.

[B]

Come on, let's. Matthew, just. Let's talk about more. No, the. I'm seeing the rise again. This could be anecdotal, but you know, it is one data point. I'm seeing the rise of the independent agencies that is like performance centric.

[E]

It's niche. Everything's niche. It's like it's a small things work now. You don't need the big ones, you have the big ones, you have the small ones, nothing in the middle.

[B]

Interesting. Yeah. Specialists, performance based. Maybe they're focused on commerce, maybe they're focused on ctv. You know, a lot of the portfolio companies across Imperium, like I'm seeing it, you know, time and time again, like where's the growth, where's the focus? It's that level down. It's the, you know, independence of all sizes. It's interesting.

[C]

And the big ones buy the small ones eventually, right?

[D]

Yeah, yeah.

[C]

So it's the circle of life in.

[B]

Agency world for sure. Okay, let's talk AI. Where are we going to start? Let's start with Code Red, which we teased earlier on. So apparently there was a meeting, there was a memo. Sam Altman had a lot to say about code red on OpenAI. What is code Red? Code Red is they need to improve the product and that means improving the product and delaying other initiatives, reportedly including ads specifically. I'll pull this out from the article link to in the newsletter. Altman said OpenAI had more work to do on the day to day experience of its chatbot, including improving personalization features for users, increasing its speed and reliability and allowing it to answer a wider range of questions. Why is this happening? This is happening because of Google. The other story I think, you know, folks didn't anticipate, particularly in the back half of this year, is the absolute crazy rise of Google. Latest Gemini Release is beating OpenAI's models on industry benchmarks. Gemini MAUs went from 450 million to 650 million. Not this year, in three months. So this is crazy. Like Google is nipping at the OpenAI heels in a way that again, I don't think anybody would have predicted when everybody was making fun of Google and all of those mishaps they had in the beginning. Go ahead, Matthew. This is something you've been wanting to talk about.

[E]

Okay. Ari, do you use Chat or Gemini?

[C]

I use Chat for the most part. I've used Gemini in app like I've used it in a Google sheet and it's been awesome. But I use Chat for the most part.

[E]

Eric, what do you use theme?

[B]

I use chat. I use Gemini across enterprise but since everybody's been talking about it, I have been using Gemini over the past few weeks.

[E]

Okay, so I use Chat and Gemini. I use Chat for everything work related and Gemini for everything personal. I sort of split those two and Chat really started to know me, got really good. In the past week or two, something happened like the personalization went away. The answers were not as good. It gave me everything that was terse, like three, three words in a sentence. And it's just sort of lost something. And I think what happened is like, I think OpenAI lost control of the Core LLM and something changed and I think they all realized it and it just, it degraded. It wasn't as good. Something changed and then Gemini got so much better and then people started to switch because they realized something happened. And that's why he called the code red again. Just hypothesizing. But I've seen a difference in chat in the past couple of years.

[C]

Yeah, I haven't seen anything particular personally, but I will say that OpenAI has felt like a company that is got too many, too many burners on or too many pots on the burner or whatever the expression is. And some of the product stuff, not the model stuff, the product stuff has been very half baked over the past year or so. If you're a user of some of the more advanced features, they just like, often don't work or links don't work. It's just, it's clearly the kind of product development that you used to see out of Facebook in the early days where shit would be thrown against the wall and it wouldn't work and stuff.

[E]

Like that everywhere and it doesn't work. And we all tried it and I tried Pulse and I tried to get it to work and it didn't work. And I gave up. I tried.

[C]

Yeah, I'm totally unable to use Pulse. The Pulse is like their daily update thing, which sounded great, but then I think it's only available if you pay $100 a month.

[D]

That subscription level never worked.

[E]

I tried so long. I tried to integrate chat into my Gmail. Did not work. Like, nothing worked after a while and I'm like, okay, Google's gonna figure this out. They have access to everything.

[C]

Yeah. So Google is just in an amazing position right now. They have the people, they have the cutting edge research, they have the chips, they have the data, they have the consumer relationships. Right. They have the capital.

It's really quite a position to be in. And the thing that was holding them back over the past two years or three years has been execution and politics, internal politics. And you know, it's been reported that Sergey came back full time and solved a lot of the politics problems. Right. That matters a lot. Right. And now they're just in a really incredibly strong position to execute this over the next couple of years.

[E]

Stock up 50% in the past couple months.

[C]

I bought some stock right you know, I bought some stock before my book came out, which was kind of silly, but it worked out up.

[B]

The. The enterprise tools are really good. And again, I remember beating them up. We were talking about this stuff like, you know, what, what are these? Like, you know, just diamonds that are appearing and everything like that. Just randomly. Yesterday, somebody reached out to me for a c. For a CES meeting, and there was a very visible prompt I had not noticed before in my Gmail, hey, do you want to add this to Calendar? And one click. Like, we had a meeting set for the time in the email in January. And it was so seamless. So not only did this completely change the workflow, make things more efficient, it kills some, frankly, calendaring startups. It's fascinating stuff.

[C]

Wait, did it get local time in Vegas?

[B]

Correct.

[C]

Because that would super intelligence only thing.

[B]

I had to change.

[C]

Oh, you had to change. I knew it wouldn't get the time zone right. And time zone will never be a solved problem.

[E]

Yeah, And.

Gemini is now embedded into Chrome. Wait till it gets better and better. Like, I don't think Gemini is going to be right there and Google Chrome's going to figure it out.

[B]

Yeah, yeah, Major, major L by me. Just, you know, making fun of them earlier this year. They fixed everything.

[C]

No, it was worth it. It was fine to make fun of them earlier this year. They deserved it. Right. Remember, we're only a year out from when they. When they used to do, you know, give us a picture of the founding fathers of America and they'd all be minorities and, and, you know, and stuff like that. So they deserved a little scrutiny, a little joshing.

[E]

Man, have they improved everything? Right? They're integrating everything.

[C]

Yeah, let's give them credit for improving. That doesn't take away from making fun of them a year ago.

[B]

All right, I won't beat myself up.

[E]

So much, but they still don't Pay for content. $100 billion in revenue last quarter, and they still don't pay for content.

[C]

Yeah, right. So that's the. That's the big question.

[B]

Couple of rumors on the. On the ads front. So Trey Tatone from the marketexture team had a. Had a nice piece this week, and in it he showed that there's a screenshot floating around out there with ads in AI mode. I have not been able to replicate it. I haven't seen it, but it seemed to be a bit of a content recommendation box underneath the content.

[E]

So here's what I heard the other day from someone who I believe heard this from a podcast that Sergey or said a year ago, which is that the 10 blue links get a CPM of X and the 10 blue links are okay because your query is one or two or three words. Right back in the day. Your query now with Gemini is a full sentence. Instead of having 10 blue links, you have one link that is so much better. And the CPM will be 10 times what it was before. And you get away from all the crap on the page to just your answer and. And a link that is perfect. It just makes Google and genius better.

[C]

Makes sense. And I think that's what we've been seeing in the evidence so far, which is fewer clicks, but more valuable clicks. And the curves are a little unclear. Like, will you be able to extract. Will Google be able to extract as much value from one perfect link that they do from 10 or 12 bad mediocre links?

[E]

The answer, I'm sure, will be yes.

[C]

You would think so. Yes, you would think so.

[B]

Yeah. Yeah. And then the other one again, a little bit of a. Somebody caught something.

OpenAI seems to be experimenting with some ads in chatgpt. Android.

[E]

I thought they killed all that right now. Now they're super focused on the core product.

[B]

Yeah, well, yeah, it's just somebody found some. Some. Some code, some strings out there, and. And you know, those strings were including things like search ads, search ads, carousel. So it's validating. I think what we've already been seeing, which is.

Their framework is likely focused on search, shopping, product, you know, these deals that they're cutting versus. Versus Anything else.

[C]

What's the origin of Code Red? I think it's Mark Zuckerberg. Like, I think I read a story once that Zuck used to call a Code Red and, like, literally, like, lock the doors of the office and make people work until it's fixed. Well, I meant in tech. Right. I guess calling a Code Red is a. Is something that. Is a pretty new concept in tech.

[B]

There was a Code Red that was from July 2001. It was a computer worm. It was named that because researchers were drinking a Mountain Dew Code Red soda when they discovered it.

[C]

Well, but then you're saying Mountain Dew invented it, not the worm, right?

[B]

No, the researchers, they just said, okay, Code Red.

[C]

Yeah. If I was a CEO and I had a Code Red, I would call it a do it with, like, get over Mountain Dew, lock the doors, and forget the Code Red.

[B]

Hey, two more quick hits before we call it a day. Just one. Again, note from the weekend. ChatGPT referrals to retail mobile apps increased 28% year over year during Black Friday weekend. It's still off of a small base, but we see where this is going and we see where this referral traffic is all going to be originating from. And it's LLMs. We can't have the debate over agentic commerce. Ari, you've been all over that. We should talk about that on its own pod at some point.

[C]

Yeah, yeah. There's just. There's been two market newsletters, guest newsletters, talking about agentic and then. And there's a lot of interesting debate. I put it on Twitter and we'll put the link in the notes. It's a really interesting question. I think it goes to psychology almost as much as technology. Like will consumers let a bot buy things for them? And it's unclear.

[E]

Yeah, as long as it's easy to return.

[C]

Yeah. I would love a bot to return stuff for me. That would be awesome.

[B]

Yeah, exactly. Final one. We definitely need to call it after here. TTD loses Judd Spencer, an engineering lead that had been there for over a decade. Seems like a big loss.

[C]

Yeah, they've been moving around a lot. I mean, Judd, the CRO Jed left recently, and then you obviously have. The Sincera team has been added, and then they brought on Andrew Eifler and product. A lot of people moving around. There's also another article that's not in our notes that we should put in that I think it was digiday broke. That trade desk has gotten somewhat flexible on pricing and they're starting to make deals that are lowering their rates, potentially in competition with Amazon and Google's rates. So that's. That's a big deal if it's true. And we'll see how that affects financials.

[B]

Absolutely. Unfortunately, we got to call it there.

[C]

Great episode. Yeah. Matt, thanks so much. MSG. So I follow you on LinkedIn. I guess since you're not in the newsletter business anymore.

[E]

Unless you give me access to your beehive and Eric, I'll see you in my party. I'll send you an invite.

[D]

Ari, you're not invited.

[C]

Okay, that's fine. I. I like to not come to your event, but I appreciate the advice.

[E]

I literally invite you to, like, seven in a row. Yeah, you don't even say no. You just don't reply, don't show up nothing.

[C]

Is it proper? Should I say no? Should I make up some reason?

[E]

I might do a calendar invite? You just say no.

[C]

Okay, I'll do my best. I'll have my AI do that.

[D]

You're inviting everything now?

[C]

I'm going to tell Gemini. Like when you see an invite from msg, please politely say no. It's my agentic workflow. All right, thank you everyone for listening. It was a great conversation.

[B]

Bye bye everybody. Thank you for subscribing to Market.

[E]

New interviews are added every week at.

[D]

Marketing and your favorite podcasting app.

Foreign.

[C]

Thank you for listening to the marketecture podcast. New episodes come out every Friday and an insightful vendor interview is published each Monday. You can subscribe to our library of hundreds of executive interviews at marketecture tv. You can also sign up for free for our weekly newsletter with my original strategic insights on the week's news at News.

And if you're feeling social, we operate a vibrant Slack community that you can apply to [e-mail].