AI changes *Nothing* — Dax Raad, OpenCode

Talk transcript

1. 0:00

   So I was told that this could be between five and fifty-five minutes long, quite a range, and that I could yap and go on whatever side tangent I wanted, and it'll be fine.

2. 0:10

   So I'm just prepping you and giving you some explanation for why this is the way it is. Uh, so my name is Dax. I work on a project called OpenCode.

3. 0:18

   It is one of the most used coding agents out there. Uh, it's fully open source, works with any model. Um, every other company likes to say, "We have the smartest coding agent."

4. 0:29

   Well, we've got the hottest coding agent. You go take a look at it. It's a great experience. Uh, awesome UX, totally themeable, personalizable, fun, fun project to use. Uh, that's all the shilling I'm gonna do for now.

5. 0:42

   Um, but let's get into the meat of this talk. So what I wanna talk about today is why AI is not gonna make you a winner. Um, we're at an AI conference.

6. 0:50

   There's a ton of AI content out there. It's all anyone's ever talking about these days. You got lots of people saying that this is the future. It changes everything.

7. 0:59

   The way, like, software development is done is gonna change. Software development is not gonna exist. You've got all these crazy proclamations about how suddenly everything is different, and if you hop on, you're gonna be able to get the benefits.

8. 1:12

   Um, and some of that is, you know, there's some truth to that to some degree. But I wanna focus on the stuff that hasn't changed, the stuff that AI is not gonna help you with, uh, the things that are critical to make a successful product that people use.

9. 1:30

   These things haven't changed. They're not gonna change anytime soon. And to be honest, AI has not really helped us with them. So I'm gonna talk about things in terms of funnel.

10. 1:40

    If you think about I'm making a product, I've got users that I want to start using my product. They start from knowing nothing about you, to hearing about you, to trying your product, to liking it, to then, like, fully converting.

11. 1:52

    Uh, if we think about that journey, there are critical moments. I'm gonna identify three critical moments that you need to focus on. And it's-- this has been true since for, I mean, my whole career.

12. 2:05

    It hasn't changed. It's the things that I think about every single day that I struggle with. Uh, despite all the experience I have, it's still very, very challenging. And the AI stuff really hasn't done much for me.

13. 2:17

    So let's, yeah, let's get started with the first one. So the absolute top of the funnel is marketing, and I'm gonna say off the bat that what I'm about to say, you're gonna feel a ton of resistance to.

14. 2:27

    So I'm gonna preface that off the bat 'cause I remember when I felt that way, uh, and it took a lot of effort in getting over, like, my own psychology, uh, to do this part well.

15. 2:38

    Um, the idea behind marketing is simple. You have someone that's going about their day. They got their problems. They got their worries. They're focused on their life. They do not give a shit about you or what you're doing.

16. 2:51

    You have to figure out how to get them to stop in their tracks, drop everything else that they have, and look at you and think about you and look into what you're offering them.

17. 3:02

    This is a really, really high bar to clear. Um, and I think when I talk about it in this way, the reason people feel resistance is because most of us aren't good at this.

18. 3:12

    You know, it's a skill like anything else. You are not gonna be amazing at it from day one. It's way outside your comfort zone. Unfortunately, a lot of people's reactions to that truth is that they feel like they shouldn't have to deal with this problem.

19. 3:27

    Uh, there's a lot of thoughts like, "Hey, if my product is good enough, it'll sell itself. Um, I don't need to do marketing. I don't e-need to do things that are gimmicky, you know.

20. 3:36

    I'm above that." The truth is there's people that make products that are better than anything you're ever gonna do in your whole life, like the absolute best products out there.

21. 3:46

    They still have to think about this problem. So if they're not too good to think about, "Hey, how do I get people's attention?" Neither are you, neither am I.

22. 3:53

    Like, we all just, just gotta do it, so I might as well start to get good at it. Um, the way that I like to think about it is you need to do things that have a shot at being shared.

23. 4:06

    So if you think about standard marketing, people will write a blog post about a new feature, or they'll hire an influencer to, like, talk about the new feature or just kind of the things that pretty much anyone can do.

24. 4:19

    Uh, you know, everyone just buys billboards these days for some reason.

25. 4:23

    The thing with these efforts are they make sense. You know, you're, like, telling people about the thing you built, but there's basically zero chance that somebody sees that and they're like, "Oh yeah, I gotta show my friends this blog post about this seventy-six feature, this product release," right?

26. 4:42

    No one's getting excited about that. They're probably gonna scroll right past. I mean, you know what people are like these days. You're all on TikTok. I know your attention spans are just basically deteriorating into nothing.

27. 4:52

    Uh, no one can focus on something that's boring. Um, so none of these things really count as top-level marketing. They're not stopping someone who doesn't know about you to start thinking about you.

28. 5:04

    Uh, the key here is to come up with something that someone's gonna see, and they're gonna think it's so crazy or so funny or so, uh, deep or so...

29. 5:15

    You know, it, it strikes a chord where they're like, "I gotta show this to my friends, to my coworkers. Uh, can you believe that they did this?" You know, you wanna elicit that type of, uh, emotion from someone.

30. 5:27

    The-- As simple as that bar is, you know, it's very easy to tell whether something's gonna clear that bar or not, it's extremely hard to come up with ideas that do clear that bar.

31. 5:36

    Um, so that's a challenge. You gotta be creative. You gotta come up with ideas that work. Uh, and you're gonna miss a lot. It's just like swinging at products.

32. 5:44

    You're aiming for a hit. You're gonna swing a lot, and a lot of times you're gonna miss. That's true of marketing as well. Don't bother taking shots that never have a chance of being a hit, right?

33. 5:54

    That's just a waste of time. Um, it's okay to try nine times on something embarrassing or crazy and doesn't work. But that one that hits, it's gonna be seen by a million people, and your target audience is gonna be in that group or connected to that group.

34. 6:09

    Uh, and if you repeat this over and over, that compounds. You get better at it. Top level marketing, amazing. The reason AI doesn't help here is it's all about creativity.

35. 6:17

    I have not had a single good idea come out of AI. Even when I use it as, like, a brainstorming partner, it, like, just gets super corny. Uh, it just c- it just can't come up with anything that's cool.

36. 6:28

    Like, it, it can't do cool. Uh, and top level marketing is about, is about cool. And that sounds like an innate thing, but it is a skill that, that you can learn.

37. 6:37

    Okay, so now that you have someone that is interested in what you're doing, you got their attention, the next thing to figure out is how to get them to the Aha moment.

38. 6:48

    So if you think about great products, there's a moment where you've had a certain interaction with a product where it clicked for you. You got what it was about, you were sold on it, uh, and you're like, "This is awesome."

39. 7:00

    And even if it wasn't for you, you might think, "This isn't for me, but this is cool enough that if I know someone that it is for, I'm gonna recommend it to them."

40. 7:08

    So that is the Aha moment. You need to figure out what is the singular Aha moment in your product and eliminate friction ruthlessly to get someone to that Aha moment as fast as possible.

41. 7:23

    So the challenges here are you might feel like, "Oh, there's a lot of good features in my app," you know. "I want them to know about features A, B, C."

42. 7:31

    You, you gotta be brutal. You kinda have to, like, you know, kill your children here. You have to pick the one that matters the most and deprioritize everything else and look at every single step that it's gonna take for someone to get there and cut it.

43. 7:45

    When I use products, like when someone sends me, like, "Oh, like, I'm building this thing. Let me try it,"

44. 7:50

    the feeling, [chuckles] like, I get this crazy amount of pain 'cause I see all the steps that it takes before I even get to the point where I even understand what the heck they're doing.

45. 8:01

    Uh, and I know that at each point they're losing 10% of their people, 20% of their people. Um, I know that the percentage of people making it to the end are so small before we can even talk about whether, hey, is the thing you're building that useful.

46. 8:16

    Uh, this is another thing that almost no companies get right. Like, think about when you're signing up for something. They ask such stupid shit up front before you even know if you want the thing.

47. 8:26

    They're like, "How big is your company?" Like, "How many employees do you have? What's your title?" Like, who cares about any of this? You gotta cut all of that out to get people to the moment that, that you care about.

48. 8:35

    And if you're able to do this, that's how you build a product that, you know, really starts to take off, uh, really starts to grow itself. All that work you did for the top level marketing, if only, like, 3% of people are getting to the Aha moment, uh, you know, that, that's a total waste.

49. 8:52

    The, the hard part, the other hard part is, like, maybe you don't even have an Aha moment. Um, that makes you think a lot about, "Hey, what am I even building?

50. 8:59

    What am I doing? Is, is this, is this worth the effort and pain of trying to build something?" Uh, and I think it's worth looking at some of the most successful products of all time.

51. 9:09

    Uh, some of them have some, like, crazy, crazy Aha moments, right? You think about ChatGPT, which is the most successful product, I would say, consumer product ever that has existed in our, in our, in our lifetime.

52. 9:21

    Uh, they give you an input box where you can type anything and you're gonna get, like, a human-like interaction as a response. Like, literally anything in the world you can type.

53. 9:30

    That's, like, an insane Aha moment, right? The moment you see that, like, the dumbest person in the world can use that, they can be like, "Yeah, this is awesome.

54. 9:37

    I get it." Uh, it's very, very, very compelling. Uh, on the social media side, right, you have something like Facebook. Uh, you have a crush and you can, like, send a poke to your crush where you're, like, kind of flirting, but it's deniable.

55. 9:50

    Instant Aha moment. Like, you are sold on this product. You're gonna keep using it. Um, so those are, like, very extreme examples, but every product should have a singular moment where you're like, "This is the reason I exist.

56. 10:02

    This is the reason why we're doing all of this." Uh, people that experience it should get it. Um, again, the reason why [chuckles] AI doesn't help with any of this is, again, it goes back to creativity and really understanding the space.

57. 10:15

    Uh, spending a lot of time thinking about it, like gaining clarity, like thinking about positioning, uh, thinking about what you're doing unique that nobody else is doing. All very, very hard.

58. 10:25

    You know, just 'cause we have AI that we can talk to, it's not suddenly gonna turn you into a winner at, at making, like, you know, amazing Aha moments.

59. 10:31

    The people that are good at that, very, very few people in the world, uh, are gonna keep [chuckles] doing that and keep crushing the rest of us. Uh, so again, another thing that we have no choice but to try to get good at.

60. 10:43

    So the last part of this funnel is the retention side. So let's say you've gotten the attention of someone. You managed to get them to the Aha moment. They like what you're doing.

61. 10:53

    They're now, like, using your product. How do you retain them for life? Like, how do you get a customer forever? Uh, if you aren't getting customers forever, you're just leaking people through, and you're gonna hit a ceiling on your business, and it's not gonna be fun.

62. 11:06

    Like, it's, it's annoying when you grow on one side, but you're leaking out on the other side. You really wanna get people for life so that number of customers you're serving is compounding over time.

63. 11:17

    Um, retention is interesting because you have to think about it kind of differently than the last two phases because it's all about power users, right? It's all about someone that's using your product for a while, that's pushing the limits of what they can do with it.

64. 11:30

    They want things to be more configurable. Uh, they want more advanced features. You have, uh, teams using it. You have larger enterprises using it. They have very specific needs.

65. 11:40

    How do you build things that support all of that without sacrificing all the stuff we did in the first two phases of keeping things simple, having a very streamlined initial first experience?

66. 11:52

    Uh, it's hard to juggle both of these t- both of these things. Uh, we've kind of been forced to think about this because we have been doing open source work, uh, I think for me personally, it's been, like, five or six years now.

67. 12:03

    Um, open source products are interesting because they are used by both, like, the random indivi- individual hobbyists that just want something that they can use, uh, that's simple and free and whatever, uh- But also open source projects are some of the best choices for like mega enterprise companies, right?

68. 12:21

    These giant companies, uh, typically want to run everything themselves. They want to host everything themselves. They're very sensitive about, uh, data going anywhere else. Uh, so open source projects have this requirement of supporting both like the really simple case and both the really advanced case.

69. 12:38

    Uh, the way we go about this is there's kind of this, there's like this, I would say, misconception that products can either be simple or they can be capable.

70. 12:48

    And they'll say like, "Oh, we're like Apple. We're building a simple experience." But what they're usually doing is building an incapable experience, and they're saying that, "Oh, we're like Apple 'cause it's the simple experience."

71. 13:00

    There's actually no trade-off between these two things. You can build both a simple and capable experience. Uh, the way to do this, it's just a, it's just more work up front, right?

72. 13:09

    You think about a feature you want to build that, uh, makes a lot of sense for users. Instead of building the feature, you kind of have to build primitives that can be assembled into that feature.

73. 13:20

    So you think about the primitive layers first, make them, uh, pretty wide scoping, um, make them really advanced, really powerful. Then you take those primitives and assemble your simple experience that 99% of people experience.

74. 13:35

    Uh, the w- the nice thing about doing it in this order is when they get more complex, you can give them more direct access to these underlying primitives. They can get more advanced.

75. 13:46

    They can flex and bend your product, and hopefully they never need to outgrow it, right? 'Cause the key here is you don't want to build a product that people outgrow, uh, 'cause you're effectively serving as a funnel for someone else's product, whatever they, whatever they outgrow into.

76. 14:00

    Uh, so I really like building things where think about like everything we ever want to do, and build a set of primitives that can eventually do those things and assemble our product using those primitives.

77. 14:11

    Uh, this is another area where AI does not really help me because

78. 14:17

    it's like so hard, right? The building and trying to build these set of primitives is how you even figure out what primitives to need and, pr- primitives you need and what you're trying to do.

79. 14:27

    Uh, that exploration process, like sure, you can like, again, chat with AI, brainstorm, but this isn't the type of thing that like you... To even like communicate to AI what you want to do, you have to understand it really well yourself, and that's a task.

80. 14:40

    The task is to understand what you're trying to do really, really well. Uh, so this is a, again, a very critical part of making a successful product. Um,

81. 14:51

    AI is not gonna suddenly make you amazing at it. It just, it, you just can't. It takes a lot of work, time, energy, uh, until your brain can wrap its head around the space you're working to design the right set of primitives that can kind of scale to, to all these situations.

82. 15:07

    All right, so the last thing I'll leave you all with is if you go and try out OpenCode, uh, and you go through the full process that I just talked about, you're gonna find a lot to criticize.

83. 15:15

    You're gonna be like, "He's not following his own advice here, here, and here." And the reason that's happening is even though I can understand all this stuff and to the degree where I can talk about it and articulate it, it's just hard.

84. 15:29

    It's very hard to follow this advice. It's simple to understand, I think, but executing it well day to day, there's no getting around the fact that it's enormously difficult, and it takes so much out of a person, a team to line everything up to get it all working perfectly right.

85. 15:47

    Uh, AI is amazing and lets us do all kinds of things that we couldn't do before. Does not save us from that day-to-day pain, uh, in trying to make something that's great.

86. 16:00

    So yeah, I wanted to do a talk that was about AI, but kind of I'm tired of people feeling like suddenly the tables are gonna turn and things are gonna be easier.

87. 16:10

    They're not easier. My life is just as hard as it's ever been. It's just as hard as it's ever been to do something amazing, but it's also where all the fun comes from, where all the purpose comes from, and people are worried about like, "Oh, people are gonna like not have any purpose with, with AI."

88. 16:24

    Uh, no, it's not gonna be a problem because, you know, things are just as hard as, as they ever were. So I hope that inspires you, scares you, I don't know, makes you sad, makes you happy.

89. 16:34

    I don't know how you're gonna feel about this, but that's all I have. Thanks for, thanks for listening.