Jeffrey Campbell

I am a technical leader, currently CTO at a company called Sierra Interactive.

Sierra Interactive is a CRM and what they call IDX in real estate. It's basically a website for your business. It's a SaaS company targeting and supporting real estate brokerages and even individual agents.

So you manage your leads in the CRM, and you get your leads with the websites, like organic leads.

Yes, HubSpot plus Salesforce, whatever, exactly. I have been there for almost exactly four years. Technical leader for 12 years total.

Before Sierra, I was at Stripe. I was a technical program manager, which is sort of leading programs but not leading people. I went to Stripe specifically because I think that's a great company. The Collisons are amazing. Before that, I led an engineering team at a company called Seasoned, which was just acquired. Seasoned is in the hospitality space; it's a B2C, B2B play. Before that, I was at a company called Alibris, which is a two-sided marketplace. And it goes on.

I've been around the block.

No. The closest equivalents to what we're experiencing now are two things. There's the app marketplace when that happened in the 2010s, and then obviously the internet in general. I was not in the space when Google appeared and Facebook appeared, but it's similar. This is even different, though, because it affects the productivity of engineering, but also consumer-facing products as well.

Well, there are a couple of things. There's sort of downward pressure: executive-level folks, board-level folks expect unrealistic output because it's magic. All you've got to do is do the AI and you can produce things magically. Internally, when you have senior engineers and you tell them to do something different, they resist it. They don't trust it. It's like teaching somebody a new way to do something they've done forever, and they always say, no, I don't like this, I don't want to do this. So it's a constant, repeated refrain: we're going to do this, to the point where you say you have to do this. So I think it's been hard for my team. It changes your job. You're no longer writing code. You're almost sort of managing a producer of code. And it's different. I got into development because it's creative. This is not as creative. It's different.

Awesome question. I am a little bit on the opposite side of the current momentum. There's a belief now, and Anthropic I think is a really good example, that you should just let agents generate code without human involvement. Let the agents generate the code, let the agents review the code, it's going to be okay. I'm skeptical of that. I think what's going to happen is, down the road, we're going to have software that nobody understands, and that to me means lower quality software. So I believe you have to be more methodical about it and make sure that a human understands that code, that humans are involved in generating that code. It's slower, but it's going to be more stable, more reliable, and more secure. So I'm trying to straddle both of those worlds and please my masters, but also be responsible about what kind of product we generate, so somebody down the road can understand it and use it.

So you have two audiences to please. You have to manage up, which for me is the CEO and the board: hey, I understand you want this faster, this is why it can't be faster than this. And you have to manage down: hey guys, I want this to be faster, and this is why you need to use these tools. In between, you're balancing expectations. On the upside: yes, we can go faster, but not that fast, and this is why. Managing down: we must go faster, but still make sure you uphold this set of standards. I want you to review the code and do things the right way, but you have to use these tools. And what I do there is I give them unrealistic deadlines, like this has got to be done in a week, to put that pressure on them. But I sort of buffer that upwards. There's a translation layer between both, and neither gets the full story, unfortunately.

That's awesome. Yes, yes, yes. Like when the internet came along, say in 1998, when your parents were dialing up to CompuServe or AOL, and the kids are all over the place. That's exactly right, and you have to please both. And for the parent in this case, they have the purse strings, so you have to make sure they feel like they're getting what everybody else is getting.

I'm good. I am excited. I am building more things personally than I ever have before. I've built multiple apps just for personal fun. I'm building things in my company now that I could not have done before; I just didn't have time. So that's exciting. Pressuring engineers to use new tools when they don't want to use new tools is not fun. I would like to just, you know, maybe replace them or do something different. So that's not fun. But big picture, I'm excited. I think it's generally positive. I think you have to be responsible about it.

A hundred percent, and it's super exciting.

Mm-hmm. And that is happening a lot for sure, for me too. I think what's happening at the executive level across all companies, SMBs, enterprise, is there's this pressure from the investors down to 10x, 20x everything. That's just not realistic. But if you tell a CEO or a board that you can get something for free, they're going to get really excited about it. So there's managing that.

They should know.

Exactly. It's hope, and it's also, I work in a private equity framework, so we're very cash constrained. So if you tell somebody who's used to being cash constrained that suddenly there's an opening here to do all the things you wanted to do, they may not do the proper analysis.

You have to, yes.

A hundred percent. It's a built-in discipline. You have to be rigorous about monthly cash flow, cost, all that stuff. I'm spoiled. I've come from VC, so I'm used to: we're just going to lose money, it's going to be great. You're right. It's more of an adult structure for sure.

That is for sure true. For us, we have not yet reached the point where we're overspending on tokens. That's going to happen. I think in the industry in general, what's going to happen is tokens are going to get more expensive. As the frontier model companies go public, they're going to have to raise their prices, so companies that are just freely spending on tokens are going to be more cost sensitive. What's happening for us, and you touched on it, is: okay, we want to build more. What exactly do you want to build? Just stuff? Do you want to just be more productive? What's the customer value? What's the ROI? It is not free, because there are still humans involved, and you're still paying for tokens. And if you build a product and you're fully embedded with one of these frontier model companies, your cost is going to go up. Have you thought about any of that? That's what I'm really struggling with. Guys, the principles are still the same. The product development lifecycle is still the same. It can be faster, but you're going to get in bed with these companies, and your cost is going to go up. Have you thought about that? What's the ROI? How are you going to grow revenue with that? I don't think that's happening yet.

It shouldn't, at that point.

It's still the same, exactly.

Yes, for sure.

I think you can optimize the SDLC, but if your product vision is the same, your outcomes are going to be the same. You get more output, maybe not better outcomes. And so if I give you just full volume and I'm feature factoring like crazy, look at all this stuff, well, you're adding overhead. Your infrastructure costs are going to go up, and your development costs are going to go up, because you now have this broad surface area of integrations with all these frontier model companies. And outcomes have to be tied to customer adoption and revenue. That principle is the same. It's never going to change.

In some ways, it's almost like a negative reinforcement: oh, we can do things really fast, and maybe you skip over the critical stuff. Now, you can prototype really quickly. If you do that, great. You should prototype and throw away the stuff that doesn't get adopted. My company is not there yet. But prototyping is very much still a first principle of a proper product development lifecycle. I think what happens is, if we can go so fast and it feels so easy, you get sloppy and you lose touch with the proper outcomes.

I'm more tech.

A hundred percent. I would love to do both, honestly, but I'm more tech right now. My domain is tech; I own that. But I think any responsible tech leader has to think about product too, and has to provide input, because we all have a stake. I have a stake. I want the company to succeed, and if we ship crappy product quickly, I lose, we all lose. So I'm definitely invested there, for sure.

TBD, honestly. TBD. I think we have created some better outcomes internally, for internal processes and efficiencies, which does allow you to save cost. But we haven't yet cracked the nut on customer-facing value. Does that make sense?

So what we're working on is assessing intent. There are warm, cold, and hot leads; make sure you identify the hot leads and get those in front of customers. That's one thing we're working on, and that is definitely AI driven. The other thing is, I don't know if you've worked with CRMs much. They're awful. Salesforce is a terrible product. It's very successful for a lot of reasons. So get rid of that interface, and just have a chat interface, something super simple. That kind of thing could work, and customers would be amenable to that. There's still a trade-off there, but that kind of stuff we will do. I'm not going to commit that it's going to be successful, but that's the kind of thing we're thinking about.

But sometimes complex things need complex interfaces. So there's a gradation of what the product should do.

My team is about 22. It used to be bigger, but we've cut costs. It was about 35, and we're about 22 now, all over the world.

Pretty much, yeah.

Yeah. I mean, the data I'm seeing is that most people are adding now. Our cuts were around private equity things, and also I had to do a full cultural shift with the team. Sometimes you have actors who are pulling in a direction that's not appropriate, so you have to make changes to the people. I think what's going to happen with AI and staffing is we're still going to be hiring people. You get more productive, but you need people at some level. It's Jevons paradox. When the steam engine was created, they thought we were going to lose all these jobs. In fact, it created more jobs. It's happened over and over and over again.

That's true. And especially in the traditional SaaS space, there was a real panic. It happened where I was too: oh my God, SaaS is dead, don't hire. That's fading. So I'm not sure we know yet how it's going to affect hiring. For us, it was more cost control and just cultural shift. I would like to hire more, but PE rigorous discipline. PE gonna PE.

Yeah, yeah.

A hundred percent. We are very tied to interest rates. That's what it comes down to.

Totally. There's just the reality of boots on the ground, like what actually is going to drive revenue. Just because you have access to AI-generated code doesn't mean you're going to produce something in an industry where there are consumers, right?

So I've seen this with my team. We had one person who we hired laterally out of customer support, and that person was really a champion for the new tools. I've seen this in a couple of different cases. I would do a balanced structure if I could. It's no different than what I did before at Seasoned, where I went from five to 35 in about 18 months: balanced senior, junior, mid-level. I want built-in mentorship cycles within my team, and now often the juniors are more capable with the new tools, and they can put a little pressure upwards: hey, seniors, look what this person's doing. They can do brown-bag lunch kinds of demonstrations. So I would definitely keep it balanced. The seniors are the hard ones to get to use the new stuff. They're skeptical. They don't trust it. They're insecure about it: wait, what is my job? Whereas the juniors don't have that; that's just what they're used to. And all my interviews would be AI-assisted. Show me how you use this tool. Here's a problem; I want you to use Claude or whatever you like to use and solve this problem. If you can't, I'm not going to hire you.

Yes, there's a whole world there in the remote space.

Yeah, or how many jobs do you actually have? Is this your third job? How many jobs are you going to carry?

So you're saying return to office became an excuse? Yeah, I think that's right.

A hundred percent. Oh my God, a hundred percent. And you can get value from somebody next to you that's maybe not clearly associated with your work task. It could be moral support, or it could be just rubber ducking with somebody. A hundred percent.

A hundred percent.

Yeah, there's a balance there, for sure.

I am profoundly optimistic.

I think we're going to figure this out. We're going to increase productivity. It's going to increase GDP. We're going to adapt, as long as we don't screw it up at the foundational level, like can we get enough electricity, things like that. I think this is a very positive thing. I think it's a leap for medicine, it's a leap for general research, it's a leap for everything, if we do it properly. And I think those who are tending negative have maybe been exposed to scare tactics. I think also there are some frontier model companies that want to close the door and own everything themselves, so they try to scare us away. But I'm very optimistic.

Yeah, well, I don't blame people for not liking big tech. I totally get it.

Absolutely, and I totally get it. Number one, the CEOs of big tech: just stop talking. We don't want to hear what you have to say. Just stop talking and just do your job. That would help. But yeah, I understand. I get why people are suspicious.

That's valid. I mean, there are arguments, like Jonathan Haidt's, that maybe mobile phones weren't a good thing.

All this social media is not a good thing. I think you could make a very strong case. It seems like that's the case, right?

Well, and they'll tell you it's an infinite layoff machine.

That's why they should not talk so much, because you're right, they're not reading the room.

I think it's true, but nobody believes it. You're right.

It's true, relatively.

That's true. Whereas there's somebody in Siberia who's just happy as can be.

It's purpose, right? Meaning.

Arthur Brooks talks about that a lot. You've got to find meaning in your life.

Great question. It's something I'm working on. Creativity, number one, for sure. Making things that generate value for people is where I find inspiration. But honestly, my kids are almost out of the house. Empty nest phase: what now? So that's something I'm thinking about a hundred percent. It used to be professional success, climb the ladder, stay healthy, have good relationships. But that's a good question. That's a journey I am restarting right now. So I don't know yet. I don't know right now.

It's exciting and terrifying all at the same time.

A hundred percent.