Apr 24, 2023 · 34m · another-podcast
Buzzfeed News, Reddit and OpenAI
0:00 / 0:00
gold bands on the timeline = statements, start to end. Hover to read, click to jump. CC turns on captions
Benedict Evans and Tony Cowanbran examine the seismic shift in digital media economics and copyright caused by generative AI, contrasting legacy platform vulnerabilities like the collapse of BuzzFeed News with the emerging challenges of AI data scraping and monetization.
How this conversation actually went
Every chapter scored 0–10 on four independent dynamics. Hover any point for the reasoning behind the score. The hosts hold 83.7% of the talking time here. How this is scored →
The hosts as informed peer 7.7 Guest teaching 0.7 Guest disagreement 0.2 The hosts pushing back 1.3
speaking balance: gold is the hosts, purple is the guest (3 minute bins)
Sharpest disagreement ▶ 14:30 Tony questions the absence of citation and source attribution
Tony challenges the emerging paradigm by arguing that traditional intellectual rigor requires citing sources and asking if we are moving toward a world where attribution is abandoned.
Hardest push from the hosts ▶ 15:33 Benedict refutes sourcing assumptions as a fundamental category errorBenedict firmly rejects Tony's inquiry regarding source lists, explaining that LLMs generate plausibility patterns rather than retrieve source records.
Biggest teaching moment ▶ 31:42 Tony points out Reddit's sole discretion pricing clauseTony educates Benedict on the exact legal language in Reddit's updated API terms, highlighting how platforms are leaving pricing deliberately undefined.
The host holds their own ▶ 26:01 Benedict details historical monetization misconceptions from radio to iPodsBenedict delivers a masterclass on historical business model confusion, quoting David Sarnoff and Jeff Zucker to contextualize modern AI monetization struggles.
the scores for every segment, with the reasoning behind each
| Chapter | Topic | The hosts as informed peer | Guest teaching | Guest disagreement | The hosts pushing back | Why |
|---|---|---|---|---|---|---|
| From Newspaper Link Taxes to LLM Content Expropriation | 8 | 0 | 0 | 1 | Benedict delivers an extensive breakdown comparing the flawed logic of newspaper link taxes with the genuine content expropriation occurring in LLM training. Tony completely agrees and validates Benedict's reasoning. | |
| The BuzzFeed News Post-Mortem and Platform Dependency | 8 | 1 | 0 | 1 | Drawing on his background at Andreessen Horowitz and familiarity with Jonah Peretti's business model, Benedict explains platform dependency risks across BuzzFeed, Facebook, and the iOS App Store. Tony contributes collaborative analogies regarding independent creators and Clubhouse. | |
| Generative AI in Creative Media: Music and Visual Art | 7 | 2 | 0 | 0 | Benedict discusses Drake AI voice clones, UMG streaming restrictions, and derivative art vs genuine artistic sensibility. Tony chimes in with an anecdote about an AI generated artwork winning a competition. | |
| The LLM Sourcing Problem and the Nature of AI Answers | 8 | 0 | 1 | 5 | When Tony asks if AI will eventually need full source attribution, Benedict firmly rejects the premise, explaining that asking an LLM for sources is a category error because language models synthesize probabilistic patterns rather than query databases. | |
| Historical Parallels in Tech Monetization and Attention Shifts | 8 | 0 | 0 | 0 | Benedict demonstrates deep tech and media history, citing Bill Gross's invention of search ads, David Sarnoff's early radio manufacturer royalties, NBC's Jeff Zucker demanding iPod cut, and early cellular roaming models. Tony listens attentively. | |
| Foundation Model Economics and an Era of Industry Uncertainty | 7 | 1 | 0 | 1 | Benedict evaluates the economics and high compute costs of foundation models, while Tony notes Reddit's ambiguous terms of service, leading to mutual agreement that the entire industry is currently operating without clear pricing models. |
Statements from this episode (14)
Evans: Publisher link tax demands against Google and Facebook are flawed
“And this is, in many ways, just a kind of, it's one of those arguments that just sort of disintegrates as soon as you poke at it, in any way at all, because they don't argue with, they don't, they think they're the only people that should get paid, not anyone …”
Evans: AI summarizing content without referral traffic makes publisher expropriation claims valid
“And it's not, now they actually are using the content and not sending the traffic to the newspaper. And so that whole newspaper argument that you are using our content goes from being dishonest nonsense to being actually true.”
Evans: BuzzFeed News never made any money
“BuzzFeed News did a lot of great stuff.
They won Pulitzer Prizes.
People like to sort of sneer at it.
It was a great journalism organization.
it never made any money.”
Evans: News is a low single-digit share of social platform posts
“And for them, news is a very small and peripheral thing.
You know, it's whatever it is.
It's low single digit shares, even all the posts that have links, nevermind all the posts that don't have links.”
Cowanbran: Creator businesses lacking direct email access are unstable
“Like if you don't, if you want an actual connection and you want an actual audience, if you don't actually have that email address and have that privilege of getting into someone's inbox, you probably have an unstable business if that platform ever disappears.”
Evans: UMG is demanding streaming platforms block AI training on music catalogs
“UMG are saying to streaming music platforms, you can't, you have to stop people straining or training on the library. You can't Let someone train a music generation AI on the Spotify corpus.”
Evans: Midjourney does not threaten art because tools cannot replace artistic sensibility
“Anybody can buy a camera, but that doesn't make you a photographer. You know, anybody can fire up mid journey. That doesn't mean you're going to, you know, doesn't mean you're going to produce something great in the way that somebody who has the right sensibil…”
Evans: The nature of art is that if it can be copied, it is not art
“The nature of art is that you can't copy it. If you can copy it, it's not art.”
Evans: Demanding citations from LLMs is a category error
“Asking and asking chat GPT for a source. That's not what it is or how it works. Because one answer is the source is the internet. Two or 300,000 things. The second answer is that when you've said, please give me a source, what these systems do is they give you…”
Evans: Layering search logic on LLMs like Bing Chat is unscalable
“They've layered this, all this logical structure on top of it. So you've got the statistical engine and then they've tried to layer on top this other stuff where it isn't an LLM. So it actually does give you a source, but also that of course becomes unscalable…”
Evans: Generative AI does not destroy search advertising economics
“I mean, this is kind of, you know, like the kind of the, like the first five minute reaction to GPT, 3.5 is this is a catastrophe for search advertising, which you kind of pull back and think, well, it's already not actually, you know, if I get, I can go to ch…”
Evans: Google did not invent search advertising
“Search advertising wasn't invented by Google, it was invented by somebody else, and they presented, I think it was at a TED talk, and everyone started booing, and they thought this was evil, like, you're corrupting search results. You're letting pay to people …”
Evans: Tech industry creative energy has left social media for generative AI
“Where is the, like, where is the creative energy at the moment? And it's not in social media anymore. Like, obviously there's a, yeah, absolutely. Yes. Yes. That otherwise absolutely all attention is now totally maniacally focused on on generative AI.”
Evans: ChatGPT breaks consumer internet economics with compute costs per query
“The idea that it actually costs money to press enter on a consumer internet site, like posting a pic, you're actually like a Facebook if it'd cost money to do a post. And you had to wait 30 seconds for it to appear. Well, that's kind of what ChatGPT is now. Li…”