Dalton Caldwell and Michael Seibel just named the biggest danger of the AI era: slop. Products that look great in demos and fundraising decks but don’t actually help users. The same pattern is destroying marketing — AI tools that produce more content faster without asking whether any of it is good. The slop war in marketing is already here. The winners will be the ones who refuse to play.

Key Takeaway: Every generation of platform technology creates a slop wave — SEO spam farms, Facebook app spam, crypto ICOs, and now AI-generated content. The pattern is always the same: metrics go up, value stays flat, then the turkey gets Thanksgiving. In marketing, the slop war looks like more blog posts, more social content, more email — all generated by AI, none of it helping anyone. The way to win is the same as it’s always been: craft over volume, retention over reach, value over vanity metrics.

In a new episode of Dalton + Michael, Dalton Caldwell and Michael Seibel — former YC partners, now running Standard Capital — lay out a framework for understanding why powerful tools produce bad outcomes — and why every generation falls for the same trap. Their analysis of “slop vs craft” isn’t about AI specifically. It’s about what happens when production costs drop to zero and nobody asks whether the output is worth producing.

We’ve written about this dynamic before — why the copilot ceiling means faster drafts don’t equal better marketing, why Jevons Paradox means cheaper content creates more orchestration problems, and why the 20x company thesis requires automating marketing execution, not just content production. Dalton and Michael just gave us the vocabulary for the failure mode: slop.

What Slop Actually Means

Seibel’s definition is surgical: “Slop is products that don’t actually help the user. They can seem great in your mind. They can seem great when you’re demoing. They can seem great when you’re raising money. But when we put it in front of the intended user, we do not see the user getting their problem solved.”

Caldwell adds the self-honesty dimension: “Slop is where you’re sort of actively self-deceiving that your thing is good when you kind of know it’s not that good.”

This is the most important distinction in AI marketing right now. Not the technology. Not the capabilities. The self-deception.

When you generate 50 blog posts with AI and schedule them across three months, do you genuinely believe each post helps your reader solve a problem? Or do you know, somewhere in the back of your mind, that you’re filling a content calendar because an empty one feels like failure?

That’s slop. Not because AI wrote it. Because nobody asked whether it should exist.

The Turkey Graph: Every Slop Era Looks the Same

Caldwell introduces a concept he and his Standard Capital co-founder Paul Buchheit (creator of Gmail) have been developing: the turkey startup. The underlying metaphor is well-known — Nassim Taleb used it in The Black Swan — but Buchheit coined “turkey graph startups” to describe a specific AI-era pattern. A turkey’s life, from its own perspective, keeps getting better — more food, more warmth, more care — right up until Thanksgiving. The graph goes up and up until it suddenly goes to zero.

Every platform era produces turkey startups:

Era	The Slop	The Turkey Graph	The Thanksgiving
Web 2.0	SEO spam farms gaming Google rankings	Revenue climbing as pages indexed	Google algorithm update
Facebook Platform	Zynga-style spam apps	DAUs growing via notification spam	Facebook cracking down on app spam
App Store	Subscription traps with 30-day engagement	Annual LTV looking great	Refund policies and negative word of mouth
Crypto	ICO tokens with no utility	Token price going up	Market correction, regulatory action
AI (now)	AI-generated content farms	Pageviews and content volume climbing	Algorithm updates targeting low-quality content, audience fatigue, brand erosion

The marketing version is already playing out. Companies that flooded their blogs with AI-generated SEO content saw initial traffic spikes — the turkey graph going up. Then Google’s core updates started penalizing thin, repetitive content — not because AI wrote it, but because it wasn’t helping anyone. Google’s official position doesn’t penalize AI content per se — it penalizes low-quality content regardless of how it was produced. The distinction matters: craft AI content survives. Slop AI content doesn’t.

As Seibel puts it: “Graphs that go up where no value is created come down. They have a life of a turkey.”

The Slop War in Marketing

Caldwell’s advice on the slop war is the most actionable thing in the episode: “I think you want to stay out of a symmetric battle for who can slop harder. If there’s a competitor and they’re spamming a lot and you’re like, ‘Well, our only solution is to spam harder,’ and you’re just in an arms race of who can create the most garbage to make the graph go up — I think don’t do it.”

This describes the current state of AI marketing perfectly. Company A publishes 20 AI-generated blog posts per month. Company B sees the traffic and publishes 40. Company A responds with 60. Nobody’s reading any of them, but the content volume graph keeps going up.

Seibel reframes how to think about it: “If you find yourself looking at the graph that allows you to take more from users than you give, I think that’s how you lose the slop war. If you find yourself looking at the graphs that other people don’t want to look at, that’s how you win.”

In marketing, the graphs nobody wants to look at are:

Time on page (not pageviews)
Return visitors (not unique visitors)
Reply rate on emails (not open rate)
Conversion from content to action (not impressions)
Brand search volume over time (not keyword rankings)

These are craft metrics. They measure whether someone got value, not whether someone saw your content. They’re harder to game, slower to grow, and more honest about what’s working.

Claude Code and Pivot Hell

One of the most counterintuitive warnings in the episode applies directly to marketing strategy. Caldwell describes how AI tools like Claude Code can accelerate “pivot hell” — the tendency to constantly switch directions because building a plausible prototype is now trivially easy.

“Because you can one-shot anything. So basically, if you’re low conviction and you decide to pivot, you can just have the new thing — send it a TechCrunch article and say to clone it. And so it’s even more tempting to just randomly change your idea over and over again and to build plausible prototypes.”

The marketing equivalent: a founder uses AI to generate a LinkedIn content strategy. It doesn’t get traction in two weeks. They pivot to Twitter threads. No traction in a week. They pivot to YouTube shorts. Then a newsletter. Then podcast outreach. Each pivot produces plausible-looking content, but none of it compounds because nothing runs long enough to build an audience.

Caldwell’s punchline applies perfectly: “If it was so easy for you to clone it, why won’t someone clone you?”

If your marketing strategy can be replicated by anyone with a ChatGPT subscription, it’s not a strategy. It’s slop with a schedule.

Taste Is the Moat

The episode’s discussion of “taste” is where the framework gets interesting for marketing. Caldwell defines it simply: “Having a strong barometer, having a strong compass of what is good and what is not good and not producing things that you know in your heart are bad.”

Seibel adds the business dimension: “To recognize great taste in other people, it’s people that have high standards and that they really care about anything that they release. They’re proud of it and they would stamp their name on it and stand by their work — and the user gets value out of the work.”

This is the case for why brand voice and creative judgment can’t be automated. Taste — knowing which content to publish and which to kill, which message resonates and which falls flat, when to post and when to stay quiet — is the one thing AI content generators systematically lack.

It’s also why the solution isn’t less AI. It’s AI with taste built in. The Doorman Fallacy applies here: you don’t stop automating. You automate the right things and preserve the judgment that makes the output worth reading.

A Brand Parent that learns your brand voice, understands your audience, and builds judgment over time is the craft approach to AI marketing. A content generator that outputs 50 posts per month with no understanding of who’s reading them is the slop approach.

Who Wins the Platform Slop Wars?

Caldwell and Seibel make a devastating observation about who actually wins during slop wars. Discussing the Facebook platform era — where countless companies tried to out-spam each other for viral reach — Caldwell notes that “no one won that was playing that game.”

Seibel’s response: “I mean, Facebook won.”

The platform always wins. The people selling tokens always win. The founders making slop never do.

Applied to AI marketing: Anthropic, OpenAI, and the AI tool providers will do great regardless of whether their tools produce craft or slop. They sell tokens. More usage means more revenue. Whether you use Claude to write one exceptional piece or fifty mediocre ones, they get paid the same.

The question is whether you win. And the answer, every single time across every platform era, is that the companies who refused to play the slop game — who focused on genuine user value, retention over reach, quality over quantity — are the ones still standing.

The Bottom Line

Seibel closes the episode with the clearest possible framing: “The core message is to avoid the slop war and instead concentrate on creating real value for users.”

For marketing, this means:

Fewer, better pieces over more, worse ones. One post that gets shared and bookmarked beats ten that get skimmed and forgotten.
Measure craft metrics. Time on page, return visits, reply rates, brand search volume. Not vanity metrics that make the turkey graph go up.
Stay out of the symmetric spam battle. If your competitor publishes 40 AI-generated posts a month, don’t publish 60. Publish 8 that are genuinely useful.
Build taste into the system. Whether that’s a human editor, a Brand Parent with learned brand voice, or a founder who reviews every piece — someone has to be the quality filter.
Don’t pivot your marketing strategy every two weeks. Pick a thesis, execute it consistently, and measure it honestly. Compounding requires consistency.

The slop war is a race to the bottom. The craft game is a race to trust. And trust, unlike content volume, actually compounds.

Building craft, not slop? Get early access to Lane — the AI CMO that learns your brand voice and executes with judgment, not just volume. Or see how it works.

Got questions? Reply to the email that brought you here, or reach out at hello@luminarylane.net.

References

Dalton Caldwell, Michael Seibel — Slop vs Craft (Dalton + Michael, Mar 2026)
Paul Buchheit — “Turkey Graph Startups” (X/Twitter, Mar 2026)
Nassim Nicholas Taleb — The Black Swan (2007) — Origin of the turkey/Thanksgiving analogy
Related: Rory Sutherland’s Doorman Fallacy — Why automating the wrong things destroys value
Related: Anthropic’s Team Hits the Copilot Ceiling — Why faster drafts don’t solve the marketing problem
Related: The Jevons Paradox Is Coming for Marketing — Why cheaper content creates more complexity
Related: The AI CMO: From Marketing Tools to Brand Parent — Our pillar piece on the Brand Parent category