Loading summary
Transcript56 lines
- [00:00]
A
Welcome back to Product Therapy. Today I'm here with Mike Fisher. Mike is CEO of MyFitnessPal, former CTO of Etsy, and one of the most remarkable product leaders I have ever met. Today we're gonna talk about AI, but not from the usual angle. A lot of people assume that if AI makes engineers, data scientists, and product teams more efficient, companies will simply need fewer of them. Mike makes the opposite case. He argues that when technology makes work cheaper and easier, the world tends to demand more of it, not less. This episode is really about what changes and what doesn't and what product that engineering leaders need to address and prepare for in order to lead well and for the future. Mike, welcome to Product Therapy.
- [00:53]
B
Christian, thanks for having me. And I want to say that most of what I know about product comes from you and Marty Kagan and the team at svpg. So right back at you. I'm thrilled to be here.
- [01:03]
A
It's an honor to have. Have you. And, Mike, you started writing a newsletter, which I'm saying you started writing too late, Mike, you should have gone on this much sooner. One of my favorite newsletters, if not probably my favorite now, just because of how thoughtful you are about the research and the mindset and how clear you're able to bring some concepts. One of my favorite articles this year was the one called More Efficiency, More Demand, because it kind of framed as all of this AI conversation has been going on, really a concept that I have struggled to articulate well, and you did it so eloquently. So let's start there, particularly with your main argument. Mike, why do you think many people are getting AIs impact on technical work wrong?
- [01:45]
B
Yeah, no, thank you for that. And I love the writing because it helps me clarify my own thoughts. And so that's why, you know, like most other writers, it forces me to, like, sit down, put pen to paper, and really think about these things. And I know obviously, everybody's terrified of AI and what it's doing to product leaders and engineers and what it means for everyone's future. I have a son who's a computer science major about to graduate college, and everybody's afraid of this. But if you look back at technologies that if this is the first technology that we as humans cannot stand on top of and be more creative, it will be the first one in history. Right. Since Fire and Wheels, we have always been able to take a piece of technology, and. And while it replaces some work, we stand on top of it and become even more creative. And one of the things that I argue in that article was There was a 19th century economist named William Jevens who watched the coal consumption rise as steam engines became more efficient. And everyone thought, of course as steam engines become more efficient, coal demand is going to decrease. And he saw the exact opposite. And he said, I realized that when you become more efficient, it actually creates the demand for more services. So those steam engines became in more demand because the cost of running them was cheaper. And so that is the, you know, and then that became Jeevan's paradox. And so that is the premise that when things become more efficient, they don't wipe out jobs, they often create more demand for that. And I think we're going to see the same thing around product is that as this creates lower and lower barriers to entry and cheaper software, we're going to start applying software to things that you never would in the past because it was too expensive.
- [03:32]
A
What do you think is behind this assumption that people make that efficiency means fewer jobs? It's almost as clear in every economic readout or as an excuse now to lay off people. What's behind this?
- [03:46]
B
We can get into some of the cost cutting that companies are feeling the pressure of. Absolutely. That the markets think this is a demand. And we're seeing that. Right. We're seeing the pressure in companies to cut some costs because this is increasing demand. And so you're seeing that that's going to swing. I think we're going to see even more demand, you know, soon because people are now, you know, the number of apps in the app stores has risen dramatically and people are finding that they can make software for cheaper and cheaper, smaller things that they would never have done in the past. And so I think people are seeing this initial reaction and afraid of it. What they're not seeing is all these great benefits. My prediction is we are going to have single person, billion dollar companies, you know, in a couple of years. It is, we're going to see that magnitude of change here because of this. And yes, some of the old roles and things that we did are different. Our jobs are going to be different, our title is going to be different. That is no different than has happened just in our careers. But across time, yeah, I have, I've had jobs that entitles that didn't exist when I was growing up. This is just natural. Humans have always worried about technology. Quick story. The founder or inventor of the modern mailbox, you know, the, that you see on the street corners, we could drop letters in, was a civil servant in England who invented this and after he developed it he was terrified that he had just ruined the moral fabric of society with this technology. Because what those mailboxes allowed people to do is women could mail letters without giving it to their brothers or fathers to mail for them and they could screen their letters. And he thought this has got to be the worst thing in the world that like. And of course by the time he died he saw that that's, that was not the case and that it actually was, was great. People are afraid of technology and the change that happens. But like I said it, it happens time and time again and this is going to make opportunities that we couldn't even dream of.
- [05:45]
A
I love that. And you're calling out history keeps showing the opposite of this, that this assumption. And you're saying we'll have billion dollar single person companies in a few years. That's already happened. I don't know if you saw the article couple of weeks ago, is it Matt Graham or something like that? And, and he had this online company selling Ozempic or GLP1, so something like that and all by himself with AI agents. So it's already happening faster than we are even having this conversation. I love your example of the radiology kind of discipline. I actually used that when I was explaining it to my kids and some of the tech in some of the medical fields. But we're thinking about the future of product management, software engineering, data science. You're saying AI tools will improve efficiency, but the field is expanding rather than shrinking. Are you thinking there'll be kind of newer types of jobs we've not imagined around our work and maybe say more about what radiology is going to teach us about the future?
- [06:43]
B
Yeah, that's another great example. Right. So Jeffrey Hansen, who is the godfather of AI, so certainly someone who knows AI incredibly well, kind of famously in 2016 said we should stop training radiologists because AI is going to be able to read all of the scans and the images and do such a better job. And fast forward seven years, by 2023, radiologists had actually grown by 17%. They're predicting that we're going to see, you know, even grow higher for the next 20 years. And the reason is that that same concept, once imaging became cheaper to read, they wanted to use it for more cases. And so that's exactly what we see is like you know, cheapening or making it more efficient a process. What it does is, allows you to use it in places you would never dream of. Quick example, for myself, I run some of these races and often my plans are in spreadsheets. And in the fall, because I was having some people help me and crew, I said, why don't I write an app? I can agentically program an app now that we could all use together, four or five of us. I would have never done that. It would be, you know, it would take me months and months and months to do that before, and now I could do it. And so things that I would never have used this for before, I'm using it now. And I think we're all finding that now.
- [08:01]
A
You know, in that example, you, you kind of called out, like, when routine tasks become easier, the remaining human work becomes more complex and even more valuable. How do you see that showing up in product and engineering?
- [08:15]
B
Yeah, I think that's exactly right. When this, you know, AI automation takes away some of, for instance, coding. And I do think we are heading to a place where humans will not code. We're heading there pretty rapidly. So we're not going to code, we're not going to review code. But what it does is it allows us and forces us to think the level above. We got to be really good at the architecture and the system's level of thinking and asking the right questions. I think you and Marty wrote an article recently about faster AI. You know, creating features just creates more junk often if we don't do it the right way, and that is the risk here, is that teams that aren't asking the right questions about what is the customer problem we're trying to solve, what is the outcome we're trying to achieve. Instead they're just thinking feature delivery. Or you can create a ton of stuff faster with automation. Doesn't mean it's in demand, doesn't mean it's the right stuff. So we've got to train our people even more. It's even more important that we're thinking about systems level and we're thinking about asking the right questions.
- [09:14]
A
In thinking about engineering, you kind of called out coding. Maybe we don't have humans coding anymore. Are there other parts of engineering work that will get compressed and which other parts become more important?
- [09:25]
B
And I do think that programming and coding have been synonymous in the past. They are now not. We still need programming the mindset of how we think about structuring the software and how the coding piece, we don't need to do that, we don't need to type it, but we still need to be programming and thinking. So some of the things that you think about in the PDLC of, like certainly the coding, the code review, that's going to get Compressed code reviews. Other agents will do the code reviews before you know, the agentically coded software. They'll do the security reviews. They'll do all of that for us. But that means we, again, we still have to think about the architecture, what we're solving for, the infrastructure, how it's going to be scalable. All of that is still really, really important for engineers.
- [10:10]
A
Does this really mean that teams will get smaller or would they just be different?
- [10:15]
B
That's a great question. I just heard this. I'm going to. I'm going to steal it. Amazon. Bezos. Jeff Bezos famously came out with a two pizza rule that we all knew for the last 20 years of like, we're going to have product teams that can be fed with two pizzas. I just heard this past week the one pizza rule. So they're saying with agentic programming, we probably don't need as many people on the team. We still need the diverse skill sets. We still need design thinking and product management and engineering thinking, but we might need it in a smaller team. And so the one piece of rule might be the way of the future. I think that's probably going to happen. And I think where they come together is on the prompts. You can then together create the prompts that then put, let's call it the agentic pipeline in motion. If we're working together, maybe I bring maybe an engineering background and you bring a product, you know the customer problems and together we can come up with a great prompt to help the agent try to create that.
- [11:14]
A
I love that I actually had from a CP I was coaching. He kind of was trying to explain to his leaders what AI is really compressing in their system. And he was like, look, we're traveling from Chicago to New York. We've always done it in an hour and a half and now we have AI. The flight time is now five minutes, but we are not still getting rid of the traffic to the airport, the long TSA lines, all of the things we have to navigate. The wait for the taxi is kind of differentiating to them. And it's like there's probably some new buttons that AI will introduce that we may not be thinking about.
- [11:50]
B
That's a really great point because you know, one of the famous studies on like, how much time do engineers actually spend coding? And they came out that it said it's only about 10 to 15% of the time is actually coding and even spent in their ides. The other time is all the other tasks that they're doing. Mm. It's the Coordination, it's the thinking meetings. I think there's opportunities there for AI to maybe help with that as well. The agent decoding is only going to take a certain percentage of their work away. And like you said, it adds other things. We now need to think about ethical, the issues that the agents cannot think of. We need to think about that these things are going to create more demand. It's the exact same way that like when we went to the cloud and we went to the cloud, we created new roles. Like all of a sudden if you've moved to the cloud of a company, you can, you know, probably relate to this. We go to the CL and then all of a sudden we get that first or second bill and we realize it's very expensive. All right, we gotta have cost control. And that created a whole new need for finops. And now every company, you know, almost to me size has a finops operation that has an engineer looking at how they do this and making it more efficient. We're gonna have that with AI as, you know, tokens are very expensive. We're gonna have somebody thinking about the optimization of our prompts and our agentic agents over time. And so that'll create a whole new role of someone like w specialize in, you know, in the optimization of this. So there's gonna be new jobs that come out of this. Absolutely.
- [13:14]
A
Oh, I love this. You're calling out on a lot of things that are changing and a lot of things. The, the AI is even changing the economics of things. You know, we, we've talked about building. I want to kind of think about what that does for expectations. You know, if, if building things gets cheaper, what happens to the expectations from the market? Our executives, our customers.
- [13:36]
B
I think it's a really great point. I think we're entering the phase of hyper personalization. What I mean by that is personalization of if I go to a marketplace or e commerce shop and you recommend some things based on everybody else you're seeing. That used to be the way we did it, right? We saw that. Oh, everybody looked for this, you know, this type of lamp. And so I'm going to show that in the top search results. That was the sort of level of personalization nowadays. I want you to know me, I want you to know my history and what I want. And I want to pick out a lamp for me. I think we're entering, right? You can do this with music. I can create music and mashups of, you know, genres and groups that don't exist because maybe I like That I can now read stories that have created sci fi worlds that I dreamed up and I can have it write and tell me a story about this. We're getting to a place where people expect this hyper personalization because it's possible with AI And I think that is what teams need to be thinking about is just generic recommendations are not good enough anymore. We need to know the customer down to the individual. You know, and my fitness pal, we have an AI coach that's rolling out. It's built on top of 20 years of helping people lose, gain, maintain their weight through nutrition tracking. But we've gotta then take your personal journey and put it right on top of that and say, but Christian, for you today, I noticed that you love to eat bananas. How about you eat, you know, an orange today because I, I noticed that you needed some vitamin C. And then I notice you're traveling. Like, we've gotta get so good at knowing the person and knowing that this is the exact recommendation for you. And so that's the phase we're entering.
- [15:19]
A
I was kind of explaining if, if personalization is easier to do, people want more personalization. You know, it's kind of, if automation is easier, I want less errors. I kind of say expectations increase for customers once you start to see things that you deemed magical or in the experience of opening an iPhone box become very commonplace. You know, it's like, why am I having an app without personalization? The standards increase for people too as well. Let's talk about judgment here. Do you think AI reduces the amount of product judgment needed or does it increase it?
- [15:57]
B
I think it increases it. I think, you know, back to that article that you wrote about when we can create 10 features in the time that used to take us to make one.
- [16:05]
A
Yes.
- [16:06]
B
Gotta be even more selective. Most products out there don't lack features. They actually have too many features. And they have features that, you know, fractions of basis, you know, basis points of users of customers are using. We don't need more of that. Yeah, I always think about, I tell people this all the time. Internal employees, we are not our customers. We know way too much about the product. We think about it 8, 10, 12 hours a day. I think about it on the weekends. I think about when I'm not running. Our customers are sometimes in the apps or in our, in our stores or on our websites for 30 seconds or three minutes. They just don't have that understanding. And so we've got to be thinking about, okay, not for me, who knows the app inside now I know all The a hundred features, I know how to navigate it. But a customer that comes in and thinks, I've got 30 seconds, I want to log my food, I want to get an update about where I'm supposed to be for lunch and what I'm supposed to do. We need to make it simple. So the product thinking is even more important now. We've got to be very careful about this great new tool and technology that we have. It doesn't lessen the need for a great product thinking, it increases it. Absolutely.
- [17:14]
A
So let's talk about roles and let's kind of reimagine product and engineering and AI native world. I want to talk about the roles and the relationships of the roles because more and more people are talking about the roles. Blending. I'm like, I'm an engineer, I can design now. I'm a designer, I can push code now. And I'm a product manager, I can do some design and engineering. But maybe define for me what you think the role of a product manager, designer and engineer is in an AI native world and maybe what the relationship between them should be.
- [17:44]
B
Yeah, I do think we're seeing where we can each, you know, each of those skill sets can do more. Absolutely. And you'll see blending of this. But it doesn't mean we don't need, you know, to me, it's that mindset and the skill set that you think about it. Right. And you know, and we think about it even before all of this, you know, LLMs and this big push for AI, we brought to the team, different mindsets and everybody has different levels of skills. You often see engineers become product managers. They can still code, but they, they don't. They're trained and they train their minds to think about these. Asking the customer questions, what problems are we solving? There is, I think, a merging of this. We are likely to start calling them more builders. That's one of the terms that I've heard where product leaders, whether you're a designer or product manager or an engineer, you might be called a builder. But it's still, you're going to have these builders that bring in mindsets from their different skill sets. It's just that, that background that you bring. So the titles I think might merge and you might see some of that. But I think we're going to find is you still need the skill sets. And I often think about organizations and like titles and stuff. They tend to swing on a pendulum and we tend to go one way and then go another. I think about this from a decentralized and a centralized perspective. We do it with organizations all the time. A platform team, you might start off centralized because you want it to serve everybody and you want standard. And then you go, well, that's slowing my teams down. So I'm going to decentralize it and give everybody. You're going to see the same thing. I think we're going to centralize, say, oh, I want, I'm only going to hire builders. And that's a single person. They got to realize that like, well, I actually need a builder who knows product and a builder who knows design and a builder who knows engineering. And I suspect that we'll see this sort of swing to one extreme and then we'll realize that there's problems over there and we'll swing back.
- [19:30]
A
I love that I've been trying to anchor on the term creator because the second I say builder, you know, engineers feel threatened and people think delivery. And I'm just like, no, we are creating and we're working hands on to create. And there used to be this old distinction between the thinkers or the thinking roles and the building roles. And you know, some of that stuff is starting to collapse. But we talk about product design and engineering, you're talking about skills. And if your thesis is right, the winning move is not fear, it's adaptation and some of this thinking. What skills you think matter most now
- [20:06]
B
for creators or builders, you're absolutely right. I think the product thinking super, super important. Understanding customers. If you serve your customer, that means you've got to understand them. You've got to think, what are they doing? All the challenges, they're trying to get through their day, they're trying to use your product for a few minutes to solve one of their problems. You got to really think about that. Know the customer that's, that's not going away. The other thing that I like to say is, is not going away and is always leadership. We need leaders. We are going to need leaders. And I don't mean, yeah, I separate management from leadership management, of course. You know, organizations need that for structure. Everybody can be a leader. We can lead from, you know, within our teams, we can lead our peers. Leadership isn't going away with this. We still need that. We need that. You know, those people who have the skill sets can help mentor and build other people. That's not going away. I think the sort of systems thinking from an engineer, systems thinking is even more important now. People who can think about how to deploy agents in different ways, but they keep it all in their mind about how the system is going to work together is super, super important right now. That that is going to be a very important skill set for at least the foreseeable future.
- [21:22]
A
I love that. And I want to go back to leaders because, you know, it's a big piece in the competency. We got the builders or the creators, product designers, engineers. But what does good leadership now look like, particularly product leadership, as particularly as engineering throughput increases?
- [21:39]
B
I mean, I would start with. Everybody's doing an AI transformation right now. That takes strong, strong leadership. Companies and people have such inertia not to change. And you know this probably better than anybody, I think. You know, I still quote you where you say, like 50% of our transitions into the product operating model fail. Transitioning is difficult. That requires leadership. That requires just an immense amount of willpower to get it done. Because you're going to run into obstacles. You're going to run into people that don't have time because they're busy delivering other stuff. They're hitting the numbers. You're going to run into pressure from the board to come down and say, but you got to do this. There's so many reasons to. It's easy to say no. Leaders are critical right today for this transformation. And then going forward, as we were talking about, you can easily. The balance or the switch between caring about the outputs to the outcomes was hard enough already because teams in the past were so trained on measuring the outputs. Oh, I put a feature out. Give me a check mark. I'm done getting them to think about the outcomes. No, no, you're not done with that feature until it solves the customer problem. That was already difficult. Now that I can create 10 times the output, it's even more important that these people really lead and go, no, no, no, that is not the measure of success. The measure of success is over here where we're thinking about the customer and we're measuring. We're. What outcome is changing, what problem are we solving. Leadership is gonna be immensely important in all of this.
- [23:09]
A
I call out a lot to people. If you want more creation or innovation, you need to create an environment for it. And you know, a lot of what we teach with context, I'm just like, look, without vision, AI is just noise. Without strategy. It is one of those things that drives a future factory. If you don't empower teams, like the concepts of empowering teams, the concepts of good context, removing ambiguity, providing those things are more important than ever.
- [23:37]
B
That environment, you know, to create the psychological safety, which means you are allowed to fail or to raise issues and not have be punished for that. Even more important. Absolutely. Even more important right now. Yeah, yeah.
- [23:51]
A
AI does not remove the need to think clearly, make good decisions. He called out understanding customers kind of navigate. Obviously you know, AI is not going to navigate your company politics and the egos and the drama and you know, and it doesn't know all the things your customer did not say, all the data that was not given to it. And we kind of call this trusted judgment. Trusted for a reason.
- [24:16]
B
Absolutely. And I really, I think we are heading into the future of humans aren't writing code. I think fundamentally changes things about what engineers care and how they're trained and so forth. We're going to see, we're already starting to see systems that self correct that if they have the right telemetry and monitoring they can start a B testing for optimization. We're having this wait. Great. That frees up our product leaders to think about bigger bets, bigger ideas, new directions that are important for the customer. What could we have never solved for the customer because it was too expensive in the past? While we're doing that, we gotta make sure that we're overseeing this, that certainly agents are gonna get outta control. And we're already seeing that in some cases this makes our lives easier. It makes it more complex, a lot more fun in my opinion. I love it, it's a wild time. But it doesn't make it easier. It makes it more complex to manage, more complex to keep our arms on. If you think it was busy, you know, let's say we're moving at 50 miles an hour in our companies. AI is gonna make us move at 100 miles an hour. Cause like you said, customers expect it. They expect us to be hyper personalized and have this thing optimized and perfect for them. That's right. That just means we're moving really fast and we got to keep up with it. And it makes it even more complex.
- [25:32]
A
If more efficiency creates more demand, then maybe leaders cannot just cut headcount and declare victory. What do you advise leaders to do instead?
- [25:43]
B
Yeah, that is a real tension right now that you know, what we see is when the headlines come out and you see this already with layoffs, there's a ton of pressure on companies for improving their margins with this. Because that's the first thing that people think about is like oh we efficiency gains. I think you're going to continue to see this. The pressure is real. Back to the pendulum. You're going to see it swing and you're going to see it Swing back to say, well, efficiency isn't really the name of the game here. It is better products for our customers. We could do more without it. So I unfortunately do think we're going to see us continue probably through this year with these, the narrative of efficiency games. But I think you're going to see the swing back when people realize that and you see this time and time again where companies will go in a wave pattern or pendulum back and forth of they hire a bunch because the market's demanding it or something, then they scale back a bit that to regain their efficiencies and then they go back out and do this. I think we're going to see the same thing here. I would argue that cost cutting and your capability expansion aren't opposites, they're just sequenced. You're going to see the cost cutting now and then you're going to see the expansion of it later. But that is again back to history. This has happened time and time again. We see this all the time. You see kind of the boom come, people get, you hire big, big teams, they have to do some cost cutting to get down to some efficiency gains and then they go back out in a couple of years and start doing that again.
- [27:11]
A
So this idea of AI for efficiency gain for productivity, for what I call like this below the line behavior of margins and you think we're going to swing at some point to kind of AI for value creation, AI in that manner and maybe we need to hire different roles, realign differently or re equip ourselves for it. But this is the obvious start of that.
- [27:33]
B
I think that's exactly the case. We're going to see the, you know, everyone's cutting for efficiency gains. You're going to see that swinging. And I think, you know, as leaders that I think the advice is there are cost cutting areas that we can kind of do without regret. You know, things that were kind of marginal before AI and now that it's being replaced, I think, you know, we can replace with that. That should be kind of a no regret move. But the other interesting thing is to think about we're going to also see the over application of AI that right, there's a difference between probabilistic and deterministic task. And a deterministic task is something that we can automate with a script. And a probabilistic task is something we should automate with AI. We're going to obviously see people overusing AI in many, many cases that you don't need it, that actually is going to cost Even more, in fact, on the cost cutting side. Back to this, what we're going to quickly see is if we do cut savings in headcount and so forth, the exponential growth in token usage. We're already seeing this with engineers. Nvidia, their CEO, said that an engineer making a couple hundred thousand dollars a year should be spending a couple hundred thousand dollars in tokens. We see engineers negotiating in their onboarding, how many tokens am I going to get? And so forth. So a rule of thumb that we're kind of thinking about is your compensation for engineers. You basically take that as your token usage. And so there's not going to be as much cost cut as, as we, as we probably think or the market kind of thinks right now, you're going to see a pretty significant increase in the cost of the tokens for this.
- [29:13]
A
Isn't that an interesting way to even kind of frame? It's like, oh, we have less engineers, but we're replacing engineers with more tokens and engineers are probably using more tokens that are more expensive than a salary of a junior engineer.
- [29:26]
B
And just like we saw in the cloud in development of FitOps, we are going to see people developing the need for optimization of our agents and they're going to have to be constantly working it out to keep the cost down. So, you know, we've seen this game many, many times before. Mark Twain, I think said that, you know, history doesn't repeat itself, but it rhymes. Still the same thing. We're, we're going to, this is going to be a repeat of many, many things we've seen, but it's gonna be slightly different. And that's the fun part of it is figuring out like, okay, how do we take these lessons from the past? Apply them here. What are we gonna see with that?
- [29:59]
A
Let's separate some of the signal from noisy. And I just wanna hear your take genuinely on like, what do you think AI changes really for product and engineering? If we're just kind of presenting it clearly to people, this is what I think will change. And this is what I think will not change.
- [30:16]
B
I think it changes the shape of our daily work. So if you're an engineer, you're a product manager, you're a designer, you're a product leader, the shape of your daily work will change definitely. And I think we need to embrace that. And that's what we back to the leadership. That's what our leaders should be talking to our people about, is embrace the change. It's not replacing you. You're gonna learn these Skill sets. This is where it's going. I think the ratio of hands on keyboard to thinking is gonna change. We're gonna need to spend more time thinking about stuff because when you could do this 10 times as fast, it actually makes it more complex. You gotta be thinking about it. All of your little cracks in your systems and your design and your architecture will show up more apparently and faster if you, if you don't. I think the speed of iteration is, is obviously going to increase that. Using this for discovery, using this in delivery, where we could do maybe three iterations in discovery before maybe we do 10, maybe do 15 now. Yeah, that iteration speed is there. Does it? We still need to do discovery, we still need to do delivery, but it means many, many faster. So I think all of that changes. Obviously, as we talked about. The last one I can think of is really the cost of software goes down and when that happens, that demand and what you can apply software for goes up.
- [31:33]
A
Yeah, I was just telling a company, you know, yeah, AI is going to expand what your organization can do. But my question is, are you going to increase your ambition or are you just going to reduce your spec? You're calling out. AI is going to change speed, it's going to change costs. It's going to say with access to capability, with more things you can do, but it doesn't change judgment, the need for leadership, clarity. It probably makes the absence of those things more obvious in several companies.
- [32:01]
B
I think that's exactly right. It'll show the cracks in your architecture or your organization structure, your trust between people. It's gotta show those much, much faster.
- [32:11]
A
Do you think there are any things that will remain stubbornly human?
- [32:15]
B
Understanding the customer. Understanding the customer. As we talked about, having that innate sense of what problem are we solving for them and understand that that is going to be squarely on the shoulders of product leaders. The discipline of problem framing. Again, agents are going to be great at optimization. I really believe that they're not good at this discovery. The thinking about what is the big problem we're trying to solve. We need leaders to think about this. We talked about trust inside teams, our agents, no matter how much we automate and do this. And we are using, within MyFitnessPal, we are using AI to help coach leaders. So we're recording one on ones with everyone's agreements and we're feeding that into engines to give immediate feedback on. Hey, in that one on one you weren't as clear as you could have been with the, you know, the goal that you're trying to achieve this week. So that's great. But that AI agent isn't building the trust between the two people. That is still a human thing that has to happen. And then judgment, I think, is the last thing that's not going to change. You need people who are really good, have really good judgment.
- [33:17]
A
Really great conversation, Mike. I was just thinking in my head, if, if, if we revisit this conversation in, say, three years. Okay, what do you think will have surprised people the most?
- [33:29]
B
Yeah, it's a great, great question. I don't know if Bill Gates or, you know, I think he's attributed to saying this of, like, things change slower than we think, but the impact of that change is much greater than we can imagine. And I think that's the case here, that it feels like things are moving pretty quickly, but realistically, you know, it's. This has actually taken decades to get here. We forget that, like, AI has actually been around for decades in trying to get to this point, but I do think it'll change. Such as we can't imagine back to the jobs. You know, I couldn't imagine having some of the jobs that I've, I've had or the titles that I've had when I was a student. We're going to do the same thing. This is going to fundamentally change things that we can't even imagine. But it's going to be great. We as humans are going to stand on top of this technology and build even more amazing things and do some incredible things that we couldn't imagine, you know, a few years ago.
- [34:21]
A
I love that. I love that we always. I've argued where, you know, technology changes fast, human behavior changes slower, but it's like one of those big foundational shifts in our world today. We probably don't have enough electricity and chips and things to scale it the way we want, but it is probably a significant technology shift for us. One of the things that we've been talking about. I don't know if you saw the recent Wall Street Journal article about people choosing to quit or kind of give up letting another big technology shift. Oh, my. I've been through, I don't know, desktop publishing, the Internet, and now this whole stuff. You know, I've had enough maybe kind of wrap up up here by telling me maybe what gives you optimism about the future of product and engineering, given the world we're. We're going into?
- [35:09]
B
Like, I love change and I love technology. So I do have to remind myself that some people don't like the pace of change. Maybe as. As fast as I do, and they're. They're afraid of it. I think what gives me hope is, as we've talked about, we've seen this many, many times, maybe not as dramatically as we're seeing now. I do think this is probably the biggest change that will happen in my lifetime. The Internet was huge, but I think this is, you know, potentially even larger. The move to the cloud was massive for companies, but this is probably even bigger. But every single time, we've been able to do even more wonderful, exciting things with this. And I feel the exact same thing here, that for the people that are willing to embrace it and go through that change, and I know change can be very difficult for people.
- [35:54]
A
Yes.
- [35:54]
B
But if you're willing to do it, you're going to come out of here being able to do some amazing things. It's like that. You know, when I was a kid, I wrote my first line of code and you see a pixel move on the screen and you're like, wow, I did that. And nowadays you can do that through agency programming and see a whole app come to life. Like, wow, I can do this. It's amazing. And I think, you know, even though it is going to disrupt and change people's roles, and unfortunately, people are going to lose their jobs over this, the people that embrace it are gonna really be able to do some amazing things on the other side.
- [36:25]
A
Last question for you. You mentioned you've got a computer scientist as a son. Maybe kind of end with, what advice are you giving him? And maybe advice you give product managers or engineers, if they were your kids today, about how to stay relevant or compete in the world ahead.
- [36:42]
B
My advice to him is embrace this. And he and I have talked about it quite a bit that, you know, you certainly want to learn the fundamentals, right. Because as we talked about the architecture, the, you know, the systems thinking, the design of this super, super important. You want to understand how an operating system and a network work so that even if you're an agent is coding for you, you understand that learn the fundamentals and then. Then embrace the technology. And that combination, I think, is incredibly powerful. And you're going to be in demand for many, many years.
- [37:14]
A
I love that. I love that. Mike, this was a fantastic conversation. Thanks for the honor of being with us on product therapy. I can't wait to have you back. Sure. If I could do some episodes on so many of your articles, I will love to one of my favorite newsletters. I will add it to the podcast notes, but thank you so much. For being here.
- [37:32]
B
Thank you so much. My pleasure.
- [37:34]
A
Wonderful. Want to learn more? Until next time, Please check out svpg.com Sign up for our newsletter that Mary Kagan puts out. Join us for one of our workshops near you and get access to all of the articles and content we put out. And thank you to everyone for joining us. Until next time, have a good day.
- [37:56]
C
A quick Disclaimer While this podcast is named Product Therapy, it is not hosted by licensed therapists or mental health professionals and it is in no way a substitute for professional mental health services. We recognize the importance of mental well being and encourage anyone facing personal difficulties to seek support from qualified professionals. See www.findahelpline.com.