The Truth About AI Wrappers (And What Actually Matters)

0:00Personalize a thing that brings value, but there's limits to it. And it's not because of the compute. It's context. What we're finding in the AI tools is context is king. And so when you personalize, do they actually know enough context to actually do a good job personalization? Or is it just spaghetti at the wall? Welcome to the Talking AI Podcast, where we talk AI with both experts in the field and early adopters. I'm your host, Matt Page, and we're here to demystify AI for you so you can get some value from it. Let's talk some AI. Steve Moran is a software and Silicon Valley vet, and the list of companies he's worked at is more like a greatest hits than a resume, but it includes Instagram, Autodesk, Yahoo, Asana.

0:46But today we're going to take things up a level and talk a bit about strategy with Steve, hitting on where's the industry going, how is it affecting things, and the choices that people are making, or maybe not making, better said. And we're going to talk maturity models, defensibility, the AI rappers, it's coming up again, and Steve's got a really good take on this. And then managing large engineering teams. What is that like when you have something as transformative as AI coming into play? Steve, we got a lot to talk about, man, but appreciate you being on the podcast. Thanks for the time, and I'm looking forward to chatting.

1:23Awesome. So let's start here. AI rappers, it's a hot topic. Like most people are, it feels like this universal agreement that, okay, if you get the AI rapper stamp, you're not defensible. The next version of ChatGPT is just going to kill your product. And there's no sense in that strategy. And I used to be in that camp, like where if you're an AI rapper, good luck. But now I've seen companies that are building real businesses and they're essentially creating a experience that works. but i'm curious your take here because when we were chatting you almost mentioned like is it even the right discussion to be having what's your take on the ai rapper i don't think it's the right discussion to even be having like first and foremost and i think one of the analogies i'll give and i'll give you a reason why rappers do and don't matter and i think this is part of the, what makes it a hard discussion is Coca-Cola sells sugar water, right?

2:26Yeah. But like the brand and Pepsi fans too, another sugar water. And there is plenty of on brand. And they sell water, water, water, water, right? It's a brand. Oh, this is yeah. Right. I think everything with lifestyle water brand and what it's just came out. I see there has a masterclass. Yes. Yes. Oh man. I can't think of it. Think of the name. Unfortunately, this is where brands matter. Warren Buffett invests in brands because there's consumer choice. And I think it's a mistake to say there is no moat. There is consumer choice. And I think even if you look at a macro basis, what people are speculating and seeing what ChatGPT has done is that there seems to be just on public sources, right, that there's a stronger B2B business with Anthropic and a much, not that they have a small business on for open AI, but there's a really strong consumer preference for chat GPT.

3:22Your grandma might've heard of chat GPT, but probably hasn't heard of Gemini and Claude. And this is a real thing in the consumer side of things. Yeah. And liquid death, is that the one you were thinking? Yes. Liquid death. That's exactly it. I looked it up. The great example of branding water. I'm almost like feeling this experience myself. Like I've used all of the models from Gemini to ChatGPT to Claude. For whatever reason, like ChatGPT is my thing. It's my go-to. And it's not that like it's better necessarily than Claude or Gemini or any of the others. It's just I've already formed that habit.

4:01That's the one I go to natively when I have a question or not even a question. I think that's the big thing too. Another topic, but like too many people think of chat GBT and AI as like a new form of Google. It's just completely different, but that's all another topic. I think that's, is that kind of the point you're hitting on? It's like this consumer preference that's starting to manifest. I think that's part of it, but I think you also, I think it also is the wrong conclusion just to say that's enough. And as someone, if you're a founder and a startup saying, I'm going to build the most loved version of a b and c and going to become the next brand because it's not that simple right it depends the market right because you look at yes is there ziploc bags but you pick something in a consumer supermarket brand and there's a whole bunch of categories it doesn't matter how good you do your marketing how good you execute your branding how much you try to reinforce it Nobody cares.

5:01There are certain categories where people care. And I think the misnomer is that anything can be branded. And I think you can create a brand on anything, but whether it is sufficient enough to create an affinity in certain areas, it doesn't. And I think this is what makes it a very hard discussion. What separates what can have a brand moat and what can't. and this is, I think you'll get an endless source of arguments on that side of things. Quick break in the pod. If you're listening to this podcast, chances are you've been thinking about how to actually use AI inside your business. And that's exactly why we built the AI Opportunity Finder.

5:43It's a free tool that helps you uncover high impact, tailored AI use cases based on your business, your goals, your pain points, and your industry. No fluff, no generic use cases, just real ideas that fit your business and they're ranked by ROI potential. It takes about three minutes to run and it's like having your own personal AI strategist for free. If you want to try it for free, check out the link in the show notes or go to hatchworks.com backslash AI dash opportunity dash finder. So do you think that's the trajectory here for some of these like frontier models? It is the brand mode. I think what's interesting too, like it's the, you mentioned earlier, OpenAI specifically, and Sam Altman's made this known, they're going after the quote-unquote consumer side of things.

6:30They're very intentionally trying to build a very user-friendly product in a sense that's going after that. So they're not just the commoditized new version of transistors or the internet or something like that that people don't care about, right? From the public announcements, it feels like that's exactly what OpenAI is doing, right? Even the fact that you have a health benchmark that was released, announced, I think, a day ago by OpenAI, right? It shows you they're going after the same markets, broad markets that Apple goes after, right? They are going on that consumer bent. And I think part Part of it is your margins are higher, right?

7:13When you're going to enterprise and commercial, there are slimmer margins, right? Because cost becomes more of a factor and you get stickier. You can't be displaced as easily. And, but I think it matters who you are, right? They know they have that lead. They're taking that. They're taking that bent. If you're a startup and a founder, right? If you're building something for an industrial, you know, industrial B2B use case, the company probably cares of the solution the best and does not care about your brand. You may be able to have a brand to get yourself out there and have your name known. But at the end of the day, that becomes commercial reasons, right?

7:52So how much is something a consumer product and how much is something a B2B product I think really matters in this space? And not every consumer product is going to have that win. Some people, I think there's an interesting part on the cursor versus Winsurf and other things in these wars is that, or even developers, they're switching daily to the model that is winning. So people have feelings that one model is better than the other. But what we see the behaviorally happen is that people bias towards one, but a lot of developers are like picking the one. Is this going to get me 5 % better today?

8:29And it's not all mathematical, but if it feels like it's 5 % better, that's the one I'm going with. Yeah. Yeah. And there is a lot of feel in there too, in a sense. I love it. You're like the business focused engineer in a sense. So take a step back and, you know, we could frame this around the frontier players or even just Joe Schmoe starting an AI startup. Like, how do you think about defensibility in this new age? Are there like attributes or vectors that you think do make sense to focus in on? What's like your take there? I think you need to know how much you're a consumer. I think if you want to go stereotypes, I think the easy vector is you look consumer versus business.

9:12Right. And then there's consumer like, like building product at Asana, like a lot of people like, oh, that's not an enterprise play. No, it is an enterprise company, but it is a prosumer. You look at Figma and things like that. It's a prosumer thing. the everyday user likes that experience and using it part of what works when you look at your business segment are you consumer are you a prosumer are you a full enterprise solution which you just need to have consistency and reliability and like quality to what you deliver yeah and i think that's the first easy pass question i would ask right and because there's been a big movement and there's been like lenny's newsletter is big in the product management world and there's been some posts and some things have been going around in linkedin taste matters right and they're quoting figma they're quoting you know what asana was early known for when project management was like way more difficult there's more competitors today but there's a number of these things that taste is winning b16 is quoting taste is winning are you a tastemaker and i think that doesn't lend itself to all categories but it's important and can you it's what's nuanced it's there too.

10:22It's funny that it's coming back to taste because that feels like a very human thing. Like the, the attribute of taste, which is a cool factor there. And like you're hitting on something too. Like, I guess one point though, before we get into that, I think one of the biggest things I think of is like this blank canvas problem, right? When I go to ChatGBT, I can do almost anything, but if I don't know what to actually do, I do nothing or I'd say, okay, I'm rewriting my emails. So I think there's an element there too, where even if it has an underlying just API connection to an LLM, but you're building an experience that makes something easy for the end user, it can be inherently valuable in a sense, even though it's just connecting into an LLM and you could call that a wrapper, but it could have true value.

11:17Oh, no, it's completely. And I don't think people realize if you read a number of the top design books, and I've read most of them that like you'll find and the it's human experience at the end of the day, if you end up having the choice between two things, why is linear win, right? Like Y Combinator's request for startups is like more designers, right? Because I think people are forgetting it's used by human beings. Who's buying human beings? And a lot of things, if the only difference between two things is one is nicer to use, people are going to pick the one that's nicer to use every time, right?

12:03If that's the only difference. and people also pick things that are more expensive all the time by varying degrees because it's a better experience and they're like i know it's more but i'm going to pay for it and i think experience at the end of the day always matters and if you make the experience good enough you could command a premium or just beat out the selection bias on by your competitors yeah i'm actually going to pull this up youtube you'll see it if you're on apple spotify whatever other thing We'll talk through it, but you mentioned the YCs request for startups. So they put this out, well, with every class they do, but this is the summer one that recently came out.

12:43And I think the point you just mentioned, like designers as founders, like there's a reason there because taste matters, right? This one's interesting though. I think this is like a summary of it on Reddit, but full stack AI startups, not tools, but entire AI native businesses. And there's like a whole video behind this, but I saw them talking about law firms and it's okay instead of just going and serving the law firm what if you just like completely leapfrog the law firm altogether and you build a full stack ai law firm that's replacing the dinosaur incumbent that was like one that kind of stuck out to me like oh man they're like this is like the swinging for the fences type of thing i'd be curious like any of these stand out to you or maybe even that one in particular hey i love the ai voice so that speaks to me so much The personal assistance, I think, is going to be noisy space.

13:35There's so much variety. I think we're unlocking use cases. And some people say you've now had AI startups that have gone bankrupt, raising hundreds of millions of dollars, or fire sale, like human, to have a tricorder type thing assistant. I don't know that's going to work. We haven't seen Amazon crack that nut or Google crack that nut with some of these large assistants. But there is a place that we have not been able to tap before. There's like clearly opportunity to augment products and experiences with voice and voice interaction. But I don't think the naive approach, voice everything, it needs to be pairing with the right experience.

14:14Mobile's not going to go away. Desktop is not going to go away. But how do you augment and pair the right use cases with those experiences? And I think that matters a lot. Yeah, I keep going. Do you have a go-to voice? I've been playing a lot with Bland AI lately. That one's really good to me. I'm trying to, there's a couple others that have come out and then you've got some open source ones as well. It's just getting so good. And I feel, I don't know about you, but I feel like it's on the like the prefaces. I can't say the word on the edge of just being like, I can't even tell if it's a human or not.

14:51There's still that little faint piece there, but it's so close. Yeah, very good. And I talked to the Bland founders. I've talked to the Vappy founders. They're hiring for like heads of engineering. And so I've talked to a few of these. And there's some really good technology out there and the gaps are being closed. What I'm most interested is like the experience that you deliver with voice and how does that get packaged? I think the underlying technologies are going to be getting better. And so people that figure out how to integrate it. Look, when you look to full stack startup, who's going to be that full stack startup?

15:25and I think that's interesting an interesting conversation I was talking to the CEO of a two billion dollar fintech startup earlier today and he was asking me the question he's like hey what do you think are you at a big disadvantage if you're not starting a company today as an AI native company and I said again nuance I don't think it's that clear cut and dry there are some industries now that are dying, right? Chegg and some educational companies are like clearly first casualties of a new wave of interaction. So the question I think actually is, are you something that's dead as a market? Are you something that is now an unlock for an AI, truly AI native?

16:13And there are definitely those opportunities. Or do you need to adapt to an AI world or an AI augment, right? There's a number of businesses that can get more productive and more efficient with AI. There's some that you can now up-level the experience and add on. And I think what you really want to do is if you're going to be dead by AI, identify that quickly and pivot that company while you still have cash. I think that's important. And so there's various levels of these opportunities and how do you bet on them as enterprise? And that's the key, being able to see if it's coming. if your near-death experience is ahead or not.

16:47It's funny, you mentioned the AI native. So I was seeing a call like an AI masterclass before this. And we were going through reasoning as a topic. And I like using DeepSeek as the demo example because you can see it's kind of thought process as it's logically thinking through it. And I asked a question like, if you're a CEO, how do you turn your company, transition into an AI native company? Just a very generic thing. It's funny, the first thing I thought about was like, what even is an AI native company? So it like reasoned and thought through, what does this even mean? Which I think it's just like a generic term that's getting thrown around a lot in a sense.

17:25Not enough. Two points there, like we could maybe talk about. One is not enough people are doing the change management to really adopt AI. I think that's actually a big problem. And a lot of leaders are thinking about it and actually executing very poorly. They're like writing like, we'll onboard cursor if we have developers. and then think that's the end of story. And that's like not a winning formula. And the other portion of it is that companies are not like breaking it apart. What is your internal AI strategy? What is your external AI strategy? It needs to be built more nuanced. Yeah. And as an aside, because you mentioned it, I think deep research is the most undervalued thing that everyone should be using.

18:13It doesn't matter. Yes. Even for your, just if it's just for your hobbies. Are you talking about ChatGPTs specifically or reasoning as a general concept? The deep research models that have now proliferated across industry. I like ChatGPTs in particular. It's very good quality. I really like perplexities. I haven't played around as much with the new search capabilities and Claude to have an opinion there, but like there's deep research on all the big models from DeepSeq to Alibaba's Quinn and, and quite often if I'm doing deep research because they're all slightly different, I'll humanize it, call it personalities, which is not true.

18:49They're just different implementations. I'll often throw the same research question at four of them and they'll take slightly different takes, even with giving it a hefty context prompt. And that gives me a very comprehensive view. And I love the citations without the citations, I wouldn't trust it. Yeah. No, and that's the piece too. It's making it more real and tangible. You can trust it more in a sense when you do have that context there. But I'm curious now, like we talked about taste and whatnot. I think, I don't know. I feel like we're going to have this point where our kids are like, oh man, you used an iPhone?

19:29Like you're so old. What is that antiquated technology you had? I feel like something's going to happen in the user experience evolution. Right now, it's so focused around like chat and this concept of chat. I think there's going to be this whole new revolution that's going to happen. But I'm curious, what's your take there? Where do you think the evolution is going to happen? Is it going to evolve? Are we going to see anything dramatic or is it just like small incremental things as these new modalities come to play? I think there's only one thing you can bet on and it's going to look different than today.

20:04But I think there's a much simpler driver to some of these things than people expect. There's studies on just, and I don't know exactly which one's correct, but there are different studies on like first child, second child, third child, et cetera, and how the different personality types go. And there's some theories that I have no idea if these are correct. So someone will yell at me in academia, but there's some about specialization and that there's a behavioral thing when you're trying to get attention and compete for parents, even though parents don't ignore, good parents are not ignoring any of their children.

20:39But as you look historically, that there's a behavioral thing that you want to get more attention from your parents, right? And if you look at the youngest kids, like the toddlers have no filters of being nice and sharing time with all of their siblings. They do try to dominate as much as possible. And that says like, the theory is be a little different. And so that you fill different niches in your family and you have these archetypes of older younger and affect this but it's about being different and i'll go as generationally people are getting off of facebook why because your grandparents are on and so if you're 16 what is the most awful thing so for no reason ignoring technology just stuff's going to go out of favor on the social networks And another generation will grow up and put another social network in place.

21:31So there's always going to be another social network, whether that's magnified or altered due to a technology advancement. Probably like it is sufficient to say there will be another social network, which actually makes it very interesting for VCs investing. Right. Cause you're just waiting for that generational moment. And what about this like element of, I think a lot of people think there's this vision of this like hyper personalization. to where just everything's automatically like instantly tailored to you and there's not necessarily this structured form factor for something. Do you see that actually manifesting into reality or do you think like structure is a good thing in a sense, especially for humans?

22:14Structure is a good thing. And there's design elements that people don't even realize like international treaties would like. People take it for granted, it's stoplights. And for most of the world, they're the same color. Yeah. And people don't even think about it. There's no universal why red should be stopped and why so many countries, but there's actually international standards that went out eventually that has the world to standardize it. I think there's, I forget if it's like Japan or something, one major holdout that has different colors for their lights and things. I don't remember which country it is.

22:49It's a horrible color too, because is it like red, green colorblindness like a thing? Yeah, I think there's like real issues with it. But so there are places that people like hyper-personalization and people say, oh, apps will be generated on the fly. You don't want that. That's actually stressful to a human. If you're going in, even if it's dynamically thing and you have to think about it, it's actually a heavy cognitive load and you'd be exhausted at the end of the day, reinterpreting it things. There is benefits to predictability. And so what should get personalized? Is it always the container that you're in?

23:24Yeah. Yeah. Such a good point there. Yeah, no, it's, and then again, it comes back to taste and structure and humans like we're creatures of habit. So that building habits, having consistency can actually be a very good thing. Even though you can do something doesn't mean necessarily should in a sense, I guess is the context there. And so the last two things I want to hit on. So I want to go back to a point you made earlier, which is I think just spot on. So many people focus on the tools, the use cases, all of this. Not enough attention is on the actual, like enabling your team to work in this new way and the adoption of it.

24:03And like we talked about the blank canvas problem they experience. I think the other thing, it's like where you want to get them to is where I call it like the one ant, all the ants analogy, right? Or like the purple cow thing where, okay, if I see one ant, oh man, now I see a thousand ants. You want them to be able to recognize, oh, I can apply AI here. Oh, I can also apply it here and here. But there has to be a structured way that you drive that adoption because they're not going to do it proactively. So all that to say, what's your thoughts there? Where are folks and companies falling down?

24:38You've led large engineering teams before. What's your thoughts there on, especially a group like engineers? I feel like they're even harder to transition into using a tool like AI in the coding process. I think like all things, human beings all need change management. It's a universal thing. Some things need more handholding than others, depending on their importance and what they look like. And so here I've been advising some of the startups and the CTOs and CPOs on how to approach this. And it's a, where are you today? And where do you want to go? and what is your approach? And so it breaks down to a very specific process of what is culturally relevant and how do your existing processes work and then tailoring a program for where you're at.

25:29If you're already walking, then the focus needs to be on running. And so where is your starting point? What is your culture and your existing company rituals look like? And then how do you design for the next step? So one of the startups that I've been advising they were nowhere. And I said, oh, are you all paying for Cursor and Windsurf? And we went through a stepwise process. And they're like, no, we're not using any of them. One of my engineers is using one. I said, okay, let's start with your IDE. Let's start with any of them. And let's get the team to start testing it out. This is a startup of 15 people.

26:03And so they started out with Cursor, then they did a week on of Cursor, then they tried Windsurf. And let the team start to have a dialogue and compare these tools. Give them some more time in your sprints and story points so you can do two ways. Take some stuff off their plate or inflate story points for a week or two and give people space because they're going to be slower even with AI tools for a while. So not be under pressure because what happens is if you squeeze them for the exact velocity you had before in the short term, they're going to go back to what they know. Yeah. Right. And they're never going to adopt it.

26:36So there's an aspect of space capacity, an aspect of letting them try and make it part of the dialogue. And then you start this discussion of what the team is sharing and you start things like what's working and what's not working. A dialogue, not just raw. This is the best thing ever. Strawberry ice cream is the only ice cream and everyone should eat strawberry. And you start things like this. And then you bring discussions to code review tools. Then do you, like right now, I'm bringing another startup to a migration tool and AI migration tool. They've graduated from IDEs, code review. Now we're introducing migration tooling so they can continually migrate their infrastructure up with another startup in the ecosystem.

27:18And so you start to take these pieces and all of a sudden you can look at on ramps and then there's other on ramps for security teams, there's other on ramps for different pieces of the ecosystem. Yeah, no, that's so good. I almost feel like you need, in a sense too, like this pattern interrupt. Like we've done, we did a whole summit around this. It took people out of their day-to-day because you almost need to do that. Like break out of the regular, same thing with like hackathons. There's a million ways you can do this, but they're good for getting out of the current flow and thinking differently.

27:53I love that meeting people where they are. Like my daughter's potty training right now, and she's almost got it down. So like her focus is just like going pee and poop in a potty. It's not running a marathon. It's such an important thing that people got to focus on. or reset my daughter went potty training about nine months ago and it was not time it was bad reverting back yeah exactly we had to take a step back reset then we did it again three months ago she's perfectly potty trained timing it worked great and businesses have the same thing just because of theory you can start but if you're in the middle of your like busy season might not As much as leaders want to say it's got to do it now, actually might not be the right time.