How to Find Consumer Startup Ideas

Dalton + Michael — How to Find Consumer Startup Ideas

▶ Watch on YouTube · Dalton + Michael · 30:25

In this episode, Dalton and Michael discuss consumer startups. When they were consumer startup founders, just putting something online made it 10x better. The age of “just put it online” might be changing, but the opportunity remains the same: create luxury experiences and give them to the masses.

TL;DR

A 30-minute conversation on the Dalton + Michael channel between Dalton Caldwell and Michael Seibel (who co-founded Justin.tv/Twitch). Their starting point: in their own founder days, putting something online was enough to make it 10x better, because Napster, Netflix, Google Maps and Kayak were “competing with nothing”. Caldwell compares it to industrialisation: once the farmers have left the farm, “you don’t get that over and over again.” That easy gain is used up, which makes consumer startups feel harder today.

Their method for finding ideas now:

• Find what people with infinite resources get, then give it to everyone. “The most reliable source of great ideas that we saw over the years is to look at what the most premium experiences in the world are that people with infinite resources have… and use technology to make them available to everyone.” Their examples: head-of-state healthcare, the private banker you text, the travel “fixer” (Anthony Bourdain’s term), a household that restocks itself, a hotel-grade home, a private chef in place of DoorDash. Uber Black is the model: Uber “didn’t invent black cars”.
• Assume a better version exists that you haven’t seen. “The trap might be to start from what you’re getting and try to make it 10% better… assume there’s an ultra premium version you don’t know about. Go find it.” Ask a rich friend how they get a tutor, how they travel.
• Stretch with the 11-star exercise. Brian Chesky’s Airbnb planning exercise asks for the five-, six-, seven- and eleven-star version of an experience. For check-in, the 11-star version is a parade and an elephant. The point is to widen the brainstorm, not to build the feature.
• AI should enable personalisation, not a cheaper imitation of humans. Caldwell: “I see a lot of consumer AI startups that try to replace things being done with humans with like chatbots or voice agents… the pitch… is that it’s saving money… and not that it’s 100x better.” Approximating “a 95% as good version” of a human will not produce the next big consumer company. Seibel: “what AI allows is for infinite personalization”, and it “makes the business that wouldn’t work before AI work after.” Emails that read like “someone’s OpenClaw” are bad for the receiver and for the sender.
• Attention is the competition. A new consumer product competes with TikTok and X, “a black hole for human attention”, so it has to be far more compelling. The hosts suggest producing joy and in-person experiences over trying to out-addict the feeds.

Dynamic-capabilities reading

• strategic-renewal/business-model — the episode’s argument is about business models that AI makes viable for the first time: premium, personal service that was only affordable at high prices can be delivered to a mass market if AI brings down its cost.

How it connects

• This is the consumer counterpart to Caldwell’s idea-selection advice in his Lenny’s Podcast episode on tar-pit ideas. The “10% better” trap here and the tar-pit trap there are both warnings about ideas that look obvious.
• The private chef in place of DoorDash is what Yhangry, a YC private-chef marketplace running on AI agents, is building.
• The 95%-as-good chatbot argument belongs on automation-vs-augmentation. It is a consumer-side case against automation that only substitutes for a human, and for AI that makes the human-grade service widely available.

What was actually ingested

The full 30:25 episode from auto-generated English captions (no YouTube chapters). Cleaned for names: Brian Chesky, Bologna, DoorDash, TikTok, OpenClaw. The transcript comes in long segments (298 for 30 minutes) but covers the whole runtime, from the opening to the sign-off.

Linked entities and concepts

• Entities: Dalton + Michael, Dalton Caldwell
• Concepts: automation-vs-augmentation
• Dangling (single-source mention, deferred): Michael Seibel (presenter, not a frontmatter author), Brian Chesky

Scope and reliability

A brainstorm between two investors, not research. The examples are anecdotes, several from the hosts’ own wealth, which they admit to. The claim that AI makes premium service cheap enough for a mass market is asserted, not shown; the episode gives no unit economics for any of its examples.