Skip to content
Dalton + MichaelDalton + Michael

How To Get Unique AI Startup Ideas

One of the biggest problems founders (and builders in general) have right now is not "how do I build my MVP?" but instead "what idea should I be building?" It feels like all of the good ideas are taken, and the competition for those ideas is ruthless. In this episode of Dalton + Michael, the two discuss some of the starting conditions for coming up with unique startup ideas, and why it can feel so hard to come up with something that is both: 1) something people want and 2) original. They suggest some tactics you consider trying that are more likely to yield differentiated startup ideas than derivations of whatever The Current Thing is. Dalton + Michael is brought to you by @Standard_Cap Dalton Caldwell on X: https://x.com/daltonc Michael Seibel on X: https://x.com/mwseibel

Dalton CaldwellhostMichael Seibelhost
Mar 16, 202616mWatch on YouTube ↗

EVERY SPOKEN WORD

  1. 0:00 – 0:24

    Why quirky, non-conformist people naturally generate startup ideas

    1. DC

      People that are non-conformist or otherwise quirky, they constantly are generating strange ideas on their own.

    2. MS

      Yes.

    3. DC

      And if anything, society tells them, "Hey, like-

    4. MS

      Ugh. [laughs]

    5. DC

      [laughs] Like, shut up.

    6. MS

      Yeah. [laughs]

    7. DC

      Like, we don't wanna hear-

    8. MS

      Yeah

    9. DC

      ... all your dumb ideas-

    10. MS

      [laughs]

    11. DC

      ... on how to make things work better.

    12. MS

      Yeah, yeah.

    13. DC

      Right? That's like-

    14. MS

      Yes

    15. DC

      ... like the system grinds that out of you.

    16. MS

      Yes.

    17. DC

      Those people tend to do well with startups because they don't actually have any problems coming up with unique ideas.

  2. 0:24 – 0:58

    AI makes building easy—now the bottleneck is having a non-derivative idea

    1. MS

      Yes. [upbeat music] All right, this is Dalton + Michael, and today we're gonna talk about how to get unique AI startup ideas. Here are a bunch of old people telling you how to get unique ideas.

    2. DC

      [laughs]

    3. MS

      Where should we start, Dalton?

    4. DC

      I mean, look-

    5. MS

      [laughs]

    6. DC

      ... the problem right now is the tools are so good-

    7. MS

      Yes

    8. DC

      ... you can crank out anything, and so the bottleneck is less building the thing, it's actually, um, having an idea, and the problem is most ideas are extremely non-unique.

  3. 0:58 – 1:38

    The funding-chasing mindset: picking ideas to impress investors

    1. MS

      Yes, and I- I would even go so far as to say many founders out there are not even being encouraged to come up with unique ideas. Like, they're almost being encouraged to, like, what just raised, what just raised a seed round, what sounds good to investors, just tell me how I can get funded, Dalton.

    2. DC

      [laughs]

    3. MS

      Just tell me what idea-

    4. DC

      Yeah

    5. MS

      ... gets me funded. [laughs]

    6. DC

      It's almost like whatever the current thing is on, on Twitter, oh, I guess I'll make a clone of that. Like, oh-

    7. MS

      Yes

    8. DC

      ... you know, when we're-

    9. MS

      There we go

    10. DC

      ... when we're recording this right now, it's like, oh, OpenClaw, uh, yeah, well, let's just clone that.

    11. MS

      Yes.

    12. DC

      Like, there you go. That's my startup.

    13. MS

      Yes.

    14. DC

      Um, [laughs] pretty low effort.

  4. 1:38 – 3:21

    Why “YC funded it” is not real market validation

    1. MS

      Pretty low effort. I think the other thing that's weird is, like, when people who are trying to start companies think that they can de-risk their idea by s- doing something that they see recently got funded or they see is, like, raised a series B, either one of those is really funny. It's like I've seen so many examples, someone's like, "Well, that thing has raised money at a billion dollar valuation, therefore, dot, dot, dot, dot, dot."

    2. DC

      Well-

    3. MS

      [laughs]

    4. DC

      ... what's so funny is we inadvertently were part of that problem, and so what would always happen when I was doing interviews for YC-

    5. MS

      Yes

    6. DC

      ... is someone I was interviewing would flag, oh, YC funded a company two batches ago doing XYZ, so that validates the market. And I'd be sitting there being like, I funded that company. I did not fund it for the idea. I actually funded a different idea, and they pivoted during the batch.

    7. MS

      [laughs]

    8. DC

      And so as the person who funded that, this is not validation.

    9. MS

      [laughs]

    10. DC

      Like, I can't stress enough, as the source of this validation thing you're talking about-

    11. MS

      No

    12. DC

      ... there's no signal here.

    13. MS

      You involuntarily validated it. [laughs]

    14. DC

      But it's just... Yeah, so I can't stress enough that trying to read the tea leaves of what kind of stuff is launching in the YC batch-

    15. MS

      Ugh

    16. DC

      ... it's not that it's, like, exactly wrong, but it's certainly not, um, a marketplace of picking ideas that have been de-risked.

    17. MS

      I think it's exactly wrong.

    18. DC

      Okay. [laughs]

    19. MS

      I think they're li- [laughs] Like, I'm sorry-

    20. DC

      Okay. [laughs]

    21. MS

      ... like you're gonna decide on what you're gonna work on for maybe the next 10 to 20 years of your life by looking at a list of what other people are working on?

    22. DC

      [laughs]

    23. MS

      Like, come on. There's certain things that you're picking that you should think about that way, like maybe where you go to college, like, or m- what you major, maybe, but, like, what you spend the next 20 years-

  5. 3:21 – 4:17

    Look in the discard bin: pursue ideas others considered bad

    1. DC

      Yeah

    2. MS

      ... of your life on, like, let's have a little bit more conviction than that. So, um, one of the ideas that you brought up that I really like is this idea of look in the discard bin.

    3. DC

      Yep.

    4. MS

      So, so let's talk about that. If, if you shouldn't look at what's cool, how do you look in the discard bin?

    5. DC

      Yeah, the way to think about it is everyone has a filter for startup ideas, and because we have very similar filters, if, if your, if your ideation process is to find an idea that is both good and no one's ever thought of it before, you're gonna have a bad time.

    6. MS

      Yeah. [laughs]

    7. DC

      You're trying to discover something that no one has ever thought of before.

    8. MS

      Yeah.

    9. DC

      And there's a lot of people in the world. I don't know if you heard-

    10. MS

      Yes

    11. DC

      ... Michael.

    12. MS

      Yes.

    13. DC

      There's billions of them, and so most likely what you actually need to do is to find an idea that has already been discarded by others-

    14. MS

      Mm

    15. DC

      ... and it's been discarded for some reason. Do you get what I'm trying to say? Like-

    16. MS

      Yes

    17. DC

      ... like instead of trying to find something no one's ever thought of, find something that everyone's thought of and thinks is bad.

  6. 4:17 – 4:57

    Use controversy as a signal: friends/family worrying can be a feature

    1. MS

      Yes. Classic examples of that, Instacart would be maybe a very classic example.

    2. DC

      Yep.

    3. MS

      One of the things that I see a lot of people do when they're in this kind of ideation stage is talk to their friends and look for things that all of their friends thinks are good ideas.

    4. DC

      Yep.

    5. MS

      And I might advise the exact opposite in this situation. Like, what happens if you talk to your friends and, like, you got maybe not universal bad-

    6. DC

      What if they told you-

    7. MS

      ... but, like, controversial.

    8. DC

      Yeah, what if they're like, "Look, I'm worried about you"?

    9. MS

      [laughs]

    10. DC

      If that... If you explain your idea [laughs] Isn't that, isn't that what happened with the founders-

    11. MS

      I mean-

    12. DC

      ... of Airbnb, Michael?

    13. MS

      I mean-

    14. DC

      Like, is that the story?

    15. MS

      It's... I would say this.

    16. DC

      Like, their parents were worried about them.

    17. MS

      The-

    18. DC

      [laughs]

  7. 4:57 – 5:51

    Choose “hard” problems and longer timelines to escape consensus

    1. MS

      Many successful YC companies, every one of their friends and family members thought the idea was bad. We keep on sa- the way you get a unique idea is is you have to move away from consensus, right? So we gotta move away from the consensus of startup founders, away from the consensus of investors, away from the consensus of your friends, and it's really interesting 'cause, like, another way that I like to think about how do you move away from consensus is doing things that are just hard. Saying that you wanna do something that takes 10 years, especially when you're a young person, that just seems really intimidating. So it's, like, easy to tell yourself, "Oh, yeah, I know the numbers say it's gonna take 10 years, but, like, I'll figure it out in two." And then you cut down all of the things that might take longer than two years to work, and I feel like if you wanna cheat, just imagine ideas that might take longer than two years to work [laughs] and the, this, the, the, the likelihood of success could be way higher. [laughs]

  8. 5:51 – 6:53

    OpenAI/Anthropic as non-consensus origins—not copyable templates

    1. DC

      Yeah, and when you think about examples, like currently the, the consensus is that, um, being a founder of OpenAI and Anthropic is a good thing. Again, maybe that'll change-

    2. MS

      Well-

    3. DC

      ... um, someday

    4. MS

      ... it took 18 months, right?

    5. DC

      [laughs] And when you think about those companies, A, they were deeply weird.

    6. MS

      Yes.

    7. DC

      Like, can you imagine a weirder start than how OpenAI got started by being a nonprofit research lab and-

    8. MS

      Yeah, during the time of SaaS ascendance.

    9. DC

      You know, a lot of things didn't go well, and they didn't really have a product for a long time. Anthropic was, like, super weird.

    10. MS

      [laughs]

    11. DC

      And a lot of their early money came from FTX. Like, the whole story is, like, bizarre.

    12. MS

      Yeah.

    13. DC

      And now they seem consensus very good, but wow, they're the definition of non-consensus things.

    14. MS

      Yes.

    15. DC

      Right? And so I'm not even arguing those are instructive stories, that if you do what they do-

    16. MS

      No

    17. DC

      ... you're gonna succeed. I'm actually ec- arguing the opposite. [laughs]

    18. MS

      Mm-hmm.

    19. DC

      Which is trying to just copy what worked in the past for other people isn't necessarily gonna give you a truly unique idea.

  9. 6:53 – 8:07

    How VC/podcast content can homogenize founder thinking

    1. MS

      One of the things that actually makes this harder is how much content VCs are putting out there.

    2. DC

      Yep.

    3. MS

      If you were a young founder in the early 2000s, how much VC content did you have access to?

    4. DC

      It didn't exist, right?

    5. MS

      [laughs]

    6. DC

      So the, the way it worked is before TechCrunch, the only people that wrote about technology were, like, the Wall Street Journal, Walt Mossberg-

    7. MS

      [laughs] Yes

    8. DC

      ... and he would mostly cover Apple.

    9. MS

      Yeah. [laughs]

    10. DC

      He'd be talking about the new Mac or whatever.

    11. MS

      Yeah.

    12. DC

      And so if you were a startup, there was nowhere, uh-

    13. MS

      CES. [laughs]

    14. DC

      Yeah, right?

    15. MS

      Yeah. [laughs]

    16. DC

      Um, and so then TechCrunch came out, and it was the first place to, like, cover startups, and then now we have this, like, you know, we have a lot of content out there, and again, I'm, you know, we are certainly part of it, and I realize that our voices have been part of this thing.

    17. MS

      Yes.

    18. DC

      My justification is that we've never tried to explicitly tell founders what to think, [laughs] more kinda just giving, telling stories, giving advice, kinda like we're doing office hours but with a wider group of people.

    19. MS

      Yes.

    20. DC

      But if you're making all of your life decisions based on podcasts, that is the definition of coming up with ideas that are not unique.

    21. MS

      Yeah.

    22. DC

      Like, literally that's the definition of a non-unique idea, is something that you heard in a podcast, right?

  10. 8:07 – 9:06

    Bring back “solve your own problem”—and ignore the $100B-only filter

    1. MS

      Yeah. There's a trope around startup ideas that have been around the entire time that we've been, um, doing startups that I might argue s- wasn't followed during the B2B SaaS era, this idea of solving your own problem or being your own user. It's interesting now because it almost seems like this has completely gone out of style, this idea that, like, oh, if I'm working on something for myself, there's no way it can get big, or like, there's no way the market's big enough, or like, why, why... I don't trust my own ability to, like-

    2. DC

      Yeah

    3. MS

      ... choose things. How do you think about this?

    4. DC

      I think one of the big talking points on, on, you know, VC podcasts or VC content is, hey, $100 billion outcomes, trillion dollar outcomes are the only things that matter.

    5. MS

      Mm-hmm.

    6. DC

      And if it doesn't seem really big-

    7. MS

      Mm-hmm

    8. DC

      ... it's not even... It's like a joke. You'd like-

    9. MS

      Hmm

    10. DC

      ... you'd like mock someone of-

    11. MS

      So how could the thing I want-

    12. DC

      And so-

    13. MS

      Yeah, yeah

    14. DC

      ... what that's causing is a lot of ideas that are, like, s- truly unique wouldn't ever fit within that framework.

    15. MS

      Yeah.

  11. 9:06 – 9:45

    Examples of ideas that look ‘too small’ until they aren’t (Twitch, Whatnot)

    1. DC

      And so of course the example, you know, your startup, uh, Justin putting the camera on his head to start Twitch.

    2. MS

      Doesn't work.

    3. DC

      That's... Is that a $100 billion opportunity, Michael?

    4. MS

      [laughs] It wouldn't work.

    5. DC

      Like, you would filter that out.

    6. MS

      Yes.

    7. DC

      Right?

    8. MS

      Yes.

    9. DC

      Something like Whatnot, which is, you know, one of my better investments at YC.

    10. MS

      Yes.

    11. DC

      It's like, okay, so it's a marketplace for authenticated Funko Pops. What's the TAM on that?

    12. MS

      Yeah, yeah.

    13. DC

      You would discard that.

    14. MS

      Yep.

    15. DC

      And so I think trying to take these, like, late stage investor advice things and apply them to startup ideas, only a small number of things would meet that bar.

    16. MS

      Mm-hmm.

    17. DC

      And so you're gonna end up with extremely derivative, repetitive ideas that would meet that bar.

  12. 9:45 – 14:12

    Who is best at unique ideas—and what to do if you aren’t that person

    1. MS

      I completely agree. And I think this leads me to my last thought on this subject, which is I have found that sometimes weird people have an easier time coming up with weird AKA unique ideas. We do have this challenge that now that tech seems like such a money maker and there's so much money invested in early-stage startups, there are a lot of people who are not very weird, who are very attracted to this space. It's funny that I'm saying weird as, like, a positive, because I think in the normal context in America-

    2. DC

      Yeah

    3. MS

      ... weird isn't seen as positive, but, like, a lot of these ideas that worked seemed really weird, and a lot of the people, when you know them-

    4. DC

      Yeah

    5. MS

      ... are a little weird. So, like, how do you think about... You know, I sometimes wonder, is the tech startup wor- world as welcoming to weird people as it once was?

    6. DC

      Yeah.

    7. MS

      How do you think about this?

    8. DC

      I think it's a great point, and again, maybe this is just, you know, we're getting old or something.

    9. MS

      Mm-hmm.

    10. DC

      That's, that's an option.

    11. MS

      Sure.

    12. DC

      But people that are non-conformist or otherwise quirky in ways that you would describe as... Uh, quirky is the nice way to say it. [laughs]

    13. MS

      [laughs]

    14. DC

      But like, like, they, they tend to do their own weird stuff, and they don't usually need to ask permission. Like, they, they constantly are generating strange ideas on their own.

    15. MS

      Yes.

    16. DC

      And if anything, society tells them, "Hey, like-

    17. MS

      Uh, uh

    18. DC

      ... you know. [laughs]

    19. MS

      [laughs]

    20. DC

      Like, shut up. Like, we don't wanna hear-

    21. MS

      Yeah

    22. DC

      ... all your dumb ideas-

    23. MS

      [laughs]

    24. DC

      ... on how to make things work better.

    25. MS

      Yeah, yeah.

    26. DC

      Right? That's like-

    27. MS

      Yes

    28. DC

      ... like, the system grinds that out of you.

    29. MS

      Yes.

    30. DC

      Those people tend to do well with startups because-

  13. 14:12 – 16:21

    Brace for disbelief: commit for years, don’t seek approval (even from them)

    1. MS

      I think that like in closing for me, um, it is hard to feel like you're working on something where a lot of people think it's a bad idea. That's hard, and I like talking about this because you should brace yourself for that. Like, that is one hundred percent the experience of everyone who's succeeded. So it's like [laughs] -

    2. DC

      [laughs]

    3. MS

      ... so like I know it's hard, right? But like you should feel like you're there, and you should feel maybe like you're gonna be there for a couple years, and like if you're working on something that you're so excited by, you're willing to be there for a couple years. As someone who funds startup founders, I'm excited by that, right? When I meet someone who I get the vibe, mm, if this doesn't work in a couple months, they're moving along, I'm, I'm always like, "Oh."

    4. DC

      Yeah.

    5. MS

      I won't, you know... I always said this phrase, it's like, "You're ten percent in, but you want me to be a hundred percent in?"

    6. DC

      [laughs]

    7. MS

      Like, that's like a weird line. [laughs] So I don't know. So for, for, for the people watching, like it's totally cool to be weird and to work on something people think is weird. If you're gonna use it-

    8. DC

      Including us

    9. MS

      ... including you.

    10. DC

      Like, basically-

    11. MS

      Including us

    12. DC

      ... like if your filter is-

    13. MS

      Yeah, including us

    14. DC

      ... do the people on the, on these videos seem like they'd approve of our ideas, we're saying no.

    15. MS

      No, no, no.

    16. DC

      [laughs] Like, don't be approval seeking-

    17. MS

      No

    18. DC

      ... from authority figures-

    19. MS

      No

    20. DC

      ... if you wanna actually have unique ideas.

    21. MS

      Well, a- and I'll extend that further. We have regularly funded ideas that we thought were bad. We have regularly worked with founders who pitched ideas that we thought were bad, and like to be completely honest, like sometimes thinking of my job as like I'm a doctor, I'm treating the patient walking in the door. [laughs]

    22. DC

      [laughs]

    23. MS

      I don't get to choose. Like, I wanna help these folks succeed, and who am I to know whether this is gonna work?

    24. DC

      Yep.

    25. MS

      I'm not gonna, you know... Let's not put ourselves in that position.

    26. DC

      Anyways, great chat, Dalton.

    27. MS

      All right, thanks. [outro music]

Episode duration: 16:21

Install uListen for AI-powered chat & search across the full episode — Get Full Transcript

Transcript of episode Eht6XjI-p64

Get more out of YouTube videos.

High quality summaries for YouTube videos. Accurate transcripts to search & find moments. Powered by ChatGPT & Claude AI.

iOSAndroidClaudeChrome