Andrew Redleaf argues AI agents are dismantling the digital advertising duopoly held by Meta and Google
Summary
Andrew Redleaf compares Meta's current position to Eastman Kodak's historical decline, noting that both enjoyed dominant market positions before technological shifts eroded their moats. He argues that AI agents are disrupting the digital advertising model by reducing human attention, eliminating emotional brand attachment, and altering data privacy dynamics. Redleaf suggests that unlike social media, general-purpose AI lacks network effects and will not consolidate into a duopoly. He contrasts Meta's proactive pivot with Kodak's failed adaptation efforts, while expressing a personal moral objection to Meta's business practices.
Episode sections
- 0:16 – 4:03
Kodak's Rise and Fall
Andrew Redleaf discusses Eastman Kodak's dominance in photographic film from the late 19th century through the 1970s, noting its stock traded in the 40s-50s. He highlights how digital photography eventually replaced film, leading to Kodak's 2012 bankruptcy despite significant shareholder returns via dividends.
- 4:03 – 5:03
Facebook as the Kodak of Digital Advertising
Redleaf posits that Facebook (Meta) is analogous to Kodak, having enjoyed a duopoly in digital advertising alongside Google. He argues that AI agents are rapidly destroying this business model by altering attention, purchase decision processes, and data privacy dynamics.
- 5:03 – 9:33
AI's Impact on Advertising and Data Models
The speaker details how AI agents disrupt the ad model by reducing human attention, lacking emotional brand attachment, and potentially surpassing tech giants in personal data collection. He suggests AI will not form a duopoly due to lack of network effects and the presence of multiple well-resourced competitors.
- 9:33 – 12:22
Meta's Strategic Pivot and Moral Critique
Redleaf notes that Meta recognizes its old model is obsolete and is spending to adapt, unlike Kodak's ambivalent efforts. He expresses a personal hope that Facebook fails, citing moral repugnance regarding eavesdropping and addictive product design.
Assets discussed
| Asset | Kind | Stance | Context |
|---|---|---|---|
| Meta PlatformsMETA | equity | neutral | Today I want to talk about Eastman Kodak and Facebook. |
| AlphabetGOOGL | equity | neutral | personal information that both of the duopolists Duopolis, Google, and Facebook had. That's going to be surpassed shortly by the AI agents. |
“Stance” is how the speakers talked about the asset, not a recommendation.
Claims worth verifying
Kodak was the second bluest blue chip behind IBM in 1978 and had a dominant position in photographic film for nearly a century.
Unidentified speaker (speaker_1) ·0:16
How to verify: Historical stock market records and company history archives; SEC filings from the 1970s. For recent events, verify 2026-08-13 through 2026-11-13. Use another period only when the speaker explicitly names it.
Primary sourceKodak stock traded in the 40s or 50s in 1978, unchanged from a 1965 peak.
Unidentified speaker (speaker_1) ·0:51
How to verify: Historical stock price data from NYSE or financial databases like Yahoo Finance or Bloomberg. For recent events, verify 2026-08-13 through 2026-11-13. Use another period only when the speaker explicitly names it.
Kodak filed for bankruptcy in 2012.
Unidentified speaker (speaker_1) ·1:11
How to verify: Court records or SEC filings (Form 8-K) from 2012 regarding Chapter 11 bankruptcy. For recent events, verify 2026-08-13 through 2026-11-13. Use another period only when the speaker explicitly names it.
Primary sourceKodak's chemical company was spun out in 1994 and remains publicly traded.
Unidentified speaker (speaker_1) ·1:31
How to verify: SEC filings or company history pages for Eastman Chemical Company (ticker: EMN). For recent events, verify 2026-08-13 through 2026-11-13. Use another period only when the speaker explicitly names it.
Primary sourceKodak distributed approximately $30 billion to shareholders against ~$200 million of contributed capital.
Unidentified speaker (speaker_1) ·3:17
How to verify: Calculation from historical dividend payments and capital contributions in SEC filings or financial history databases. For recent events, verify 2026-08-13 through 2026-11-13. Use another period only when the speaker explicitly names it.
Primary sourceKodak squandered between $5 and $10 billion on digital transformation attempts.
Unidentified speaker (speaker_1) ·3:40
How to verify: Analysis of Kodak's historical financial statements (10-K) for R&D and capital expenditures during the digital transition period. For recent events, verify 2026-08-13 through 2026-11-13. Use another period only when the speaker explicitly names it.
Primary sourceDigital advertising duopoly is based on attention, personal information, and narrow targeting.
Unidentified speaker (speaker_1) ·4:43
How to verify: Public disclosures from Meta and Alphabet regarding their advertising business models and data usage policies. For recent events, verify 2026-08-13 through 2026-11-13. Use another period only when the speaker explicitly names it.
Alphabet's AI search led to increased queries but decreased monetization.
Unidentified speaker (speaker_1) ·5:23
How to verify: Alphabet (GOOGL) quarterly earnings reports and investor presentations detailing search revenue trends. For recent events, verify 2026-08-13 through 2026-11-13. Use another period only when the speaker explicitly names it.
There are five AI companies in the US owned by well-resourced companies.
Unidentified speaker (speaker_1) ·8:14
How to verify: Market analysis reports identifying major US-based AI developers (e.g., Google, Microsoft, Meta, OpenAI, Anthropic). For recent events, verify 2026-08-13 through 2026-11-13. Use another period only when the speaker explicitly names it.
Facebook's business model involves eavesdropping on conversations and selling information to corporate America.
Unidentified speaker (speaker_1) ·11:16
How to verify: Public reports on data privacy practices, FTC investigations, and academic studies on social media data usage. For recent events, verify 2026-08-13 through 2026-11-13. Use another period only when the speaker explicitly names it.
Do your own research
- Review Eastman Kodak's 10-K filings from the 2000s and 2010s to verify dividend payouts and R&D spending on digital transition.
- Analyze Meta's (META) latest quarterly earnings reports for trends in ad revenue, user engagement time, and AI-related capital expenditures.
- Examine Alphabet's (GOOGL) recent earnings calls and SEC filings for data on AI search query volume versus monetization rates.
- Research the technical architecture of current AI agents to assess their ability to collect personal data compared to traditional social media platforms.
- Evaluate academic or industry studies on 'brand attachment' in non-human AI interactions to verify claims about emotional neutrality.
- Check Meta's investor presentations for details on their 'AI-first' strategic pivot and associated cost structures.
- Compare network effect metrics (e.g., daily active users, cross-platform interactions) between social media platforms and AI assistant ecosystems.
Full transcriptexpand
- 0:00
- Unidentified speaker (speaker_0): Welcome back to Ask Andy with Andrew Redleaf.
- 0:04
- Unidentified speaker (UNKNOWN): As
- 0:04
- Unidentified speaker (speaker_0): an investor, entrepreneur, and independent thinker, Andy explores the ideas, trends, risks, and opportunities that have captured his attention.
- 0:16
- Unidentified speaker (speaker_1): Today I want to talk about Eastman Kodak and Facebook. When I became an investment professional in the summer of uh 1970. nineteen seventy-eight, Kodak was probably the second bluest of the blue chips behind IBM. And for nearly a century,
- 0:36
- Unidentified speaker (speaker_1): Kodak had had a dominant position in photographic film. Nearly everybody of my age would have had a Kodak Instematic in Stimatic camera, which they filled with uh Kodak film
- 0:51
- Unidentified speaker (speaker_1): The stock at the time traded mostly in the 40s, maybe in the 50s, essentially unchanged from a peak in uh 1965, along with the rest of the Dow which had struggled through the stagflation of the 70s. Now, of course, not too
- 1:11
- Unidentified speaker (speaker_1): many years after, 20 years, film photography. had been completely replaced by digital photography. And by 2012, Kodak had filed for bankruptcy. Its chemical company was spun out in 1994 and remains a
- 1:31
- Unidentified speaker (speaker_1): publicly traded company today. Buying the stock in the 40s or 50s in 1978.
- 1:39
- Unidentified speaker (speaker_1): actually would have been okay. Not great, but not awful, as the company returned approximately $75 a share in dividends over its prosperous years. Now, Kodak's story is unusual in the length of time for which
- 1:59
- Unidentified speaker (speaker_1): it maintained a dominant position. But it's uh technology in dominant position. Being obsoleted by a new technology is of course fairly commonplace. In uh 1975, heading off to college, I would have brought my uh
- 2:19
- Unidentified speaker (speaker_1): Smith Carola electric typewriter, also uh obsoleted tennis 10-ish years later by uh the PC and Word processor. One thing that to note about digital photography, replacing filmed photography, is it's hard to
- 2:39
- Unidentified speaker (speaker_1): identify visible winners and companies that profited from that switch, you know, as opposed to consumers. Now, you know, everybody today carries a digital. camera in their pocket, but it's hard to say that digital photography was anything other than a
- 2:58
- Unidentified speaker (speaker_1): barely noticed attachment to your smartphone. Kodak, over its history, as best I can tell, had approximately 200 million dollars of contributed capital and distributed approximately $30 billion to
- 3:17
- Unidentified speaker (speaker_1): shareholders. Again, it was founded at the beginning of the 20th century, so not a magnificent return over the years, but okay. In its final years, you know, basically from 08 to 2012, but uh for some time before,
- 3:38
- Unidentified speaker (speaker_1): Kodak. made
- 3:40
- Unidentified speaker (speaker_1): somewhat uh begrudging, half-hearted attempts to transform its business to meet the digital age, and in the uh process squandered somewhere between five and ten billion dollars. I said at the beginning, today's talk was
- 4:00
- Unidentified speaker (speaker_1): on Kodak and Facebook.
- 4:03
- Unidentified speaker (speaker_1): and to bring Facebook into it, I would suggest that in many ways Facebook is the Eastman Kodak of this generation. Essentially, for its um entire corporate life, Facebook has enjoyed a duopoly
- 4:23
- Unidentified speaker (speaker_1): in digital advertising, which has they're though not totally obsoleted and replaced print and television advertising, has very largely supplanted it. The duopoly, along with Google, is based on three things: attention of an
- 4:43
- Unidentified speaker (speaker_1): audience, a trove of personal information on its audience, and the ability to customize and focus narrowly target advertising. I would suggest that AI, AI agents,
- 5:03
- Unidentified speaker (speaker_1): whatever its other effects, is in the process of very, very rapidly destroying that business model. Perhaps the first indicator. indication of that was in alphabet's earnings, where as they switch from standard search to
- 5:23
- Unidentified speaker (speaker_1): AI search, queries are going up, but monetization is going down. The adoption of AI agents uh works in uh several ways to to destroy. the ad selling business model of both Google and
- 5:43
- Unidentified speaker (speaker_1): Facebook or or Meta. First, on the attention dimension, uh human attention is in fact a limited quantity and definitionally if you're paying attention to one thing, um, it has to come from distract you or or displace attention to another. So to the extent
- 6:03
- Unidentified speaker (speaker_1): that people begin to get their information from AI type agents, attention is going to go down. Second, if in fact we begin to use AI agents
- 6:17
- Unidentified speaker (speaker_1): to make and execute purchase decisions, the whole sort of nature of advertising changes for a variety of reasons. I mean, uh first, advertising to human beings is meant to uh drive awareness, attention, and emotional connection to a brand.
- 6:37
- Unidentified speaker (speaker_1): Presumptively, that won't work with AI agents, you know. First of all, since um their attention is essentially infinite, they can focus on the whole of all the information uh available to them. more or less simultaneously and more or less at equal detail, grabbing an AI's
- 6:57
- Unidentified speaker (speaker_1): attention through advertising doesn't seem likely. Second, you know, as far as we can tell, AIs don't have emotion. AIs won't attach attached to brands and form associations, emotional associations.
- 7:15
- Unidentified speaker (speaker_1): as human beings do. Next, I said at the beginning, Meta's model and the driver of digital advertising and its supplanting of traditional media was in large part based on the supply of
- 7:35
- Unidentified speaker (speaker_1): personal information that both of the duopolists Duopolis, Google, and Facebook had. That's going to be surpassed shortly by the AI agents. If one is in fact using an AI agent in whatever form as a personal assistant or as one's
- 7:54
- Unidentified speaker (speaker_1): primary source of information screening and editing, certainly the AI will have as much. personal information about you as Facebook and Google do currently. Importantly, it seems to me very unlikely
- 8:14
- Unidentified speaker (speaker_1): that the AI agents will evolve into a competition controlled duopoly. Monopoly. The reason being, apart from there already being five in the US owned by very well resourced companies, it appears to be the case that
- 8:34
- Unidentified speaker (speaker_1): creating a good enough general purpose AI does not require crossing, forwarding, or otherwise breaching company modes. And also, it's not obvious that within uh general purpose AIs there are network effects.
- 8:54
- Unidentified speaker (speaker_1): Whereas one was on Facebook because everybody else was on Facebook, or one used Excel because everyone else used Excel. It doesn't seem obvious to me that an AI agent is better simply by market share. In my choice of
- 9:14
- Unidentified speaker (speaker_1): AI. there's no reason for me to be particularly interested in your choice. In contrast to Kodak, whose efforts to transform itself were always somewhat ambivalent, with always one
- 9:33
- Unidentified speaker (speaker_1): eye to maintaining the film business for as long as they could. It seems to me, at least, that both Google and Meta recognize that their old model is in the process of being obsoleted. To me,
- 9:54
- Unidentified speaker (speaker_1): Facebook has decided and is acting as if in order to spend whatever it needs. to spend to be part of the technology that is obsoleting its model. That doesn't in any sense guarantee or
- 10:14
- Unidentified speaker (speaker_1): even suggest that the profits available in the new world will be anything like what was available in their era of digital advertising duopoly. I would suggest that my analogy between digital
- 10:34
- Unidentified speaker (speaker_1): replacing film
- 10:36
- Unidentified speaker (speaker_1): and AI supplanting search and social media is likely to be apt in the way while digital imagery production became ubiquitous. and you know we went from people carrying filmed cameras
- 10:56
- Unidentified speaker (speaker_1): on special occasions to having one on them at all times. The benefits, uh the profits as it were, were um highly diffuse. Now, I admit to um uh sort of bias and here I've always thought Facebook and to a lesser extent Google,
- 11:16
- Unidentified speaker (speaker_1): but Facebook's business model was um morally repugnant, that eavesdropping on people's conversations, effectively selling the information to corporate America was uh sneaky at best,
- 11:33
- Unidentified speaker (speaker_1): that the manipulation of attention was the creation of an addictive and harmful product. You know, I do actually hope that Facebook meets the fate of Kodak and ultimately goes away.
- 11:57
- Unidentified speaker (speaker_0): You've been listening to Ask Andy with Andrew Redleaf. Independent thinking on the ideas shaping today. today's financial world. Thanks for listening and join us next
- 12:07
- Unidentified speaker (UNKNOWN): time.