The Speed of Context: Why AI Changed What Engineers Actually Do

Speaker A: Your Sprint board is full, work is shipping. But here's the question nobody wants to ask out loud. Did anything we ship actually make a difference? That's the problem Product Driven was written to fix. It's the number one best selling book for tech leaders. Not another agile framework, not another process layer. It's the missing playbook for engineering leaders who are done confusing output with outcomes. And right now you can grab your free digital copy at fullscale IO/product driven. The link is in the show notes. Now has never been a better time to start a startup. And that's what we're talking to Eban Basong today about. He just started a new company and he's the engineering leader from before. Got a lot of experience, been around other startups. He's given us some of his best practices of things he's learned recently using AI and about his journey of being a brand new founder and his brand new company. And you're going to learn more about that today on the Startup Hustle podcast. And we're back for another episode of Startup Hustle. This is your host, Matt Watson, excited to be joined today by Eban Bissong, who is a founder and CEO of a brand new company he just started a couple weeks ago. We're going to talk about that today and uh, all the other cool stuff he's been doing with AI as an engineering leader. Welcome to the show, man.

Speaker B: Hey, thanks for having me. Happy to be here.

Speaker A: You caught me off guard. You just started your own company and uh, you know what? I love the entrepreneurial spirit. Is this your first time as an entrepreneur?

Speaker B: Uh, this is not the first time. I've always been a career founding, um, engineer and I've had a couple of stints as a, uh, as a founder. Uh, this is the one that is fully my idea and I'm owning it fully.

Speaker A: Okay. Whole new world, huh, huh? Now you gotta do everything. You got no business partners, it's all you.

Speaker B: Well, I actually picked up one business partner which is my wife and so she's actually one of the ones that inspired me to work on this. And um, yeah, it's been going well so far.

Speaker A: I feel like working with your partner is like either the best or worst idea ever.

Speaker B: Yeah, for sure. But uh, it's been going pretty well and uh, we um, we definitely communicate really well.

Speaker A: Well, so you've been doing a lot of really cool stuff with AI, I guess. I'd love to hear more about how you've been using AI as an engineering leader at, ah, some of Your previous roles and even what you're doing today.

Speaker B: Yeah, definitely. Yeah. So I was uh, uh, previously working with park, uh, DNA. And um, that came about and when I was uh, I had my mentee that I was mentoring and he was, that he's the CTO of Park DNA. And um, yeah, uh, we just saw that there was a great opportunity to instill AI culture there. And um, that's when I came in and started taking a look at. All right, where are we going with AI? How can we push the boundaries? And um, that led us to a couple of, uh, a couple of great insights.

Speaker A: You know, I think part of the problem with AI when it comes to engineering is we're all trying to figure out like, how do we build software now? Right. Like the whole SDLC has been shaken up and to some degree I feel like as developers, half of our job now is to try and figure out how to get out of the way a little bit. Right. Like, how do we help people that can vibe code things and prototype things? Like how do we help them contribute as much as possible almost directly to the repo while ensuring they don't screw a bunch of things up? Does that make any sense?

Speaker B: Yeah, that, that totally makes sense. And that's, that's something that we, we were um, experimenting with or actually pursuing at Park. DNA was how can we not only just use AI within, um, and having operators just within the engineering team, but having having operators outside of the engineering team as well.

Speaker A: So what did you learn from that? What, like what worked well or didn't work well?

Speaker B: Yeah, I would say that the, with engineers producing code at a high rate, sometimes it's very difficult to do that knowledge share between, let's see, the engineering team and the support staff. The support staff is helping resolve any support bugs. And um, we noticed that when we were pushing a lot more code, we're shipping faster. But when we were working with uh, the support team, the support team would ask a question in our Slack channel and, and because we were heads down, we weren't able to get to it as quickly as we needed to. So what we did was when openclaw came out, that kind of just lit a light bulb for us to say, hey, okay, now we can start building out another teammate per se. So what we did was we uh, integrated openclaw, had to connect, ah, read only to all of our data sources. And so now when a new feature comes out, um, and uh, support has a question about it, then it would have the first. It would actually have a. They would have an early response and until an uh, uh, engineer can follow up.

Speaker A: Well, and did you give the support team any ability to try and help write code that way as well? Like they can report an issue or a bug and like it tries to go right at it?

Speaker B: Yeah. So what we did is we've, we uh, added in ah, Jira ticket support. So what I found was interesting is like when a support ticket came up and uh, we called it RTD2. RTD2 was our extra teammate, our droid per se. Uh, we would have conversations within that thread and then from then on um, if there was a bug that was found or needed to be addressed, we started to see that our support team started saying uh, asked if RTD2 can create a bug ticket. So uh, we weren't there until they were shipping code but it was the first step to being able to uh, go ahead and track those bugs and um, ah, start working on them.

Speaker A: What's interesting is we've moved to an era now where it's no longer about the speed of code, it's about the speed of context. It's like how can I get all the information I need to know about what needs to be built, why it needs to be built, who it affects, all that kind of stuff. It's like as developers we really need as much information as possible from people like your support team that you're giving examples of.

Speaker B: Oh yes, oh yes. And, and I believe that what we found is trying to push as much as we can on to, into our Slack and using that as a hub along with any of our other data sources we can uh, we can really get as much context as we need to. Including our note takers as well.

Speaker A: Yeah. So did you guys record a lot of your, your calls internally or with customers or stakeholders? Did you guys record all that kind of stuff?

Speaker B: Always. I think that was the number one benefit. Um, and uh, that was the number one thing that I was really pushing there is that everything needs to be recorded. Uh, anything that's not um, that, that's, that that's not sensitive, uh, needs to be recorded. So one uh, thing that we did was we actually with RTD2 we tagged it had its own note taker so we could um, uh, invite that note taker into the chat and then R2D2 would have the um, we have the ability to pull in any of that context there and help us with organizing meetings and um, any other follow ups as well.

Speaker A: So the sort of bot that you're talking about was that something you Guys built or was that a third party tool or.

Speaker B: Yeah, so we actually leveraged openclaw. So when openclaw came out that's what would lit a light bulb for us. So we just spun that, uh, set that up on a VPS and um, yeah, configured it to our slack and

Speaker A: went off and that had support to join meetings and stuff too.

Speaker B: Say again?

Speaker A: And it could join your meetings like Google Meet or Zoom or something?

Speaker B: Yes. Yeah. So we connected it to an uh, uh, already well known uh, service, uh, Fireflies, and we had to connect it there.

Speaker A: Pretty cool, man. It definitely feels like that's the future of where we're going is like how do we record and collect all that stuff? And um, we're actually internally building a product that's kind of similar to what you just described called Product Wave. And that's actually what it does is it pulls in those same transcripts like you're talking about and analyzes them to help build like a roadmap and have evidence around. What should we be building? Where are we getting all the feedback from all the different places around how to shape the roadmap? It's like we, uh, Darrell, our chief operating officer was just talking about the other day. It's like we used to want our managers to, you know, go to these meetings and they'd have like a few action items or something would come out of the meetings. But it's like now it's like we just want the transcript and we want the conversations to be longer and more in depth so we can mine them with AI for sure.

Speaker B: Yes, that's true. That's what we're seeing on our end as well.

Speaker A: It's like how do I get more information that we can do something with that gives us. It's just incredible. It's like we want more information now.

Speaker B: Yes, we want more information and uh, then figuring out how to do the right things with that information as well.

Speaker A: So I love this. I think it's very smart and I think this is how we capture that kind of context that the developers need. Um, what else did you do at parc? You think that was really good with, with AI for the team?

Speaker B: Yeah. Um, so I would say that when I initially came into the team, everybody was hesitant on using AI, so it was really changing the mindset around it and uh, had a lot of pushback around. Hey, this is something I've been doing for a long time. Um, uh, I don't believe that AI can do it as well. And so um, over time just kind of just really just kind of pushing the horn on that. You started to see things change, things shift and, and then the engineers are start, we're starting to enjoy their jobs. Uh, I enjoy working at Park DNA a lot more because they were able to build a lot more and really focus in, on, on the product and, and actually solving the problem at a faster rate. So, um, yeah, and, and um, I would say that uh, we started to start identifying a lot more use cases that um, that would help make our lives easier. Like having openclaw every single night review where there was um, any potential, like um, any small refactorings that we could do within our code base, like using dry or solid principles that we weren't taking advantage of or taking a look at our datadog to seeing what errors were reoccurring to see if it can go ahead and take a stab at it and come uh, up with a pr.

Speaker A: It's so incredible that, that we can do all that kind of stuff, right? It's like, um, absolutely incredible. So you said the team was hesitant. What, what do you think finally got them over the, the hump there? Like what, what finally got them to kind of change their mind?

Speaker B: Yeah, I would say that it was, uh, having a lot of, uh, we had a lot of deliverables and um, I kind of took the approach of if I was, want to bring change, uh, to the team that I have to kind of show that. So I was setting up a lot of pipelines, uh, sharing a lot of things on our standups at the end of our standups around what's happening and what's new in AI. And so then over time, as I believe the uh, engineers started seeing that they could also be leveraging this power as well. And uh, it kind of started sticking. And now we have, we had more engineers super excited to share about what they've learned and how, what use cases that they've been, um, building to help their own personal workflows.

Speaker A: And I agree with you. I think the only way we can really do it is just show the team. Just show the team like this is what you can do. This is the cool stuff we're able to do and the value it provides. Um, but there's still some people that are still fighting against it. I met a guy the other day at the airport who was like, data centers are ruining the world. I'm never going to use AI. It's never going to happen. It's destroying our planet. And I'm like, all right buddy, I get it.

Speaker B: But yeah. And it was here. Yeah. And for me, to my team, it was always about, you know, it's not perfect, but what can we learn from it?

Speaker A: Here's the truth. Most of what software teams ship doesn't matter to customers. Not because engineers don't care, because the connection to real outcomes got engineered out of the system. Motivation doesn't die loudly. It fades one ignored release at a time. Product Driven is my book and it is the number one best selling book for tech leaders who are ready to fix that. Grab your free digital copy at full scale. The link is in the show notes to improve it.

Speaker B: And uh, my thing is always finding those friction points and solving those friction points, um, as we can to make it work for us and to get the benefit.

Speaker A: Yeah. So at full scale, We've got about 300 software engineers that work for us and across 80 different clients that work for us. And so we see a whole spectrum of it across the 80 clients we still have. I think it's 4, 4 that don't allow AI usage at all. And some of them is, they're worried about intellectual property, like the, you know, copyright of the code, saying, okay, we wrote the code and we can have intellectual property of the code versus if AI writes the code. We, you know, we're worried about that part of it. Like, some people are hung up on that, some are, you know, worried about security or compliance things and whatever. Like, it's, it's all over the board. Like, and I have a friend that works like for a railroad and they, they're not allowed to use AI at all. And he's like, he's like, but I use it all day long on my phone. I'm still using it. I'm not supposed to, but I'm still doing it. I mean, some companies are still dragging their feet at this. And then some of our team members, you know, if they're just a team of one, it's hard for them to learn all the values like, like you described. Like, you know, I had to keep showing the team all the cool stuff that you can do. Some people are in small teams and they don't have somebody like you showing them what they can do. And so they're kind of still struggling to figure it out. It's a huge spectrum, right. And we've got some people that are like cutting edge and experts at this stuff, but it's a big spectrum and we're all learning new stuff every day. Um, it's an exciting time to be a builder is what I would say. It's not A good time to be a coder. The people that were in love with the code, kind of like you described people, like, well, I can't do it as good as I can. And my response is like, I don't even care what the code looks like. I care what the code does. I care what I get from the code.

Speaker B: Exactly.

Speaker A: But for some of these people, that was their craft, it was their identity.

Speaker B: Yeah, yeah, I agree with that. It's about, you know, can I solve this problem, uh, with quality. Right.

Speaker A: So, yeah, one of the things we talk a lot, uh, about from a marketing perspective and operations perspective is just good enough. Like what's good enough? Like I can use AI to create blogs, I can use AI to create images that go on our blog and stuff like that. And some of them are mind blowing good. Just like, I don't think a human being could have did this. This good. Um, and some of this stuff, it's just like, ah, but it's good enough. We're able to automate it. And good enough is good enough. Right. Like same stuff with. Same thing with a lot of the things you probably saw with openclaw. Right. It's not necessarily perfect, but it's good enough. Right?

Speaker B: It's good enough. Yeah. And the thing about openclaw, it's definitely opened up my mind to more possibilities of a lot of things that you can, that it can do to help you with the proactivity. A lot of the watchers that you had to write before, um, that can just through English, um, can be set up.

Speaker A: Well, I think we're going to see a lot more systems that are doing stuff like openclaw is doing or Claude's agents and skills that it can do cowork and all these things. But you still need a lot of human in the loop stuff. It's like a. I went and did all these things, but now I need to review what it did. Yeah, right. I need to get like final approval to it and then some things we eventually let it loose. Okay, I trust you to go do this thing. But, um, I feel like we're going to have more and more systems that some of them are just for that reason. It's like I had to build this system just so I could kind of manage what the AI, uh, does and, and I can review its work and have workflow around what it does. That makes sense. Yeah.

Speaker B: Um, and that's something I've been, uh, exploring lately as well, is how can I. For tasks that we need to validate, how can I Enable the AI or the agent to be able to give me artifacts to make that validation a lot easier. So one thing that comes to mind is if I have a new feature on the front end, um, having that uh, after defining all the different um, acceptance criteria, having it spin up a, a, an application, uh, seeding that data and then actually using like playwright to actually create a gif or some type of video and attach that to the PR so I can review that as well.

Speaker A: Yeah. That's so cool man. Well, so you, so you decided to start your own company a couple weeks ago?

Speaker B: Yes, I did, I did.

Speaker A: Are you crazy or just an entrepreneur?

Speaker B: A little bit of both.

Speaker A: So entrepreneurs are all a little crazy.

Speaker B: Yeah, yeah. So um, yeah, this, this actually came about with um, um, my wife. She, she teaches or she was teaching psychology in terms, in terms of how um, your ability to think about an outcome reflects about how you get to the actual goal. Right. And so um, and so in that I saw that. All right, well with, with linguistics and how you talk, how we what the words that we say dictate our outcomes, how can that relate to AI? Um, and in that I saw, well, hey, you know, if I think bigger, I can essentially build anything I want using Hermes, Open Claw, everything else. And I saw that, hey, this is actually a great time to be, be building, to be being able to start uh, a new venture. Um, and that's kind of where I started heading into building out my new application.

Speaker A: So tell us more about your application and why that problem.

Speaker B: Yeah, so I created synv, which um, is the AI Voice Journal. Um, so you talk to it. Ah, it reflects your thinking back and then it helps you reach your own decisions. And so where that came from was back in December I was, I found out about Whisper Flow and I was like, oh, I can talk to my agents and that's just how I usually process information. And so I, I actually had a session with my cloud code and I said hey, why don't you build out this, the system for me to be able to journal. And over time I was able to get insights on my thinking patterns, on things that were uh, blocking me to those goals or things that I was, I was, I was doing very well. So I started sharing this, the system out to sharing the system with uh, some of my non technical friends are my, my less AI native friends I would like to say. And they saw that, that it was really, really, it was super cool. So at then, at that point I shared it with my, my wife and she started using it and at that time, um, at that time I was just pushing up bugs and she would text me and saying, hey, like, you know, send these down. Um, so, yeah, so then I just really kind of decided to just go all in on this. Just, uh, because, um, yeah, this is the big opportunity here to build and um, uh, the opportunity.

Speaker A: Well, so as an entrepreneur, I feel like one of the hardest things is validating your product and figuring out, okay, how am I going to sell this thing, what am I going to charge for it, how do I find customers? Have you figured all that part of it out yet?

Speaker B: Yeah, getting there. So, uh, with my wife's help and, uh, we actually put together a beta group that ran over, uh, a month. And, and so we had about 25 people, uh, half, uh, in the. In US and in Thailand. And um, they. And then we, after that we got that feedback and recorded all of it via our note taker and really saw that, okay, this is something that we should move forward with.

Speaker A: So you've been building this for a little while?

Speaker B: Yes. So started around February, Start February with the alpha product, but we ran the actual beta in May.

Speaker A: So what did you learn from the feedback?

Speaker B: So we learned that, um, people really want to, um, people. Some people need a space to be able to reflect. Right. And um, uh, especially if they, they don't. If they don't actively pay for any type of, I guess any type of counseling or if they, um, they don't have anybody to talk to at that point in time. Um, but we also learned a lot of things just around how they, how they use the app versus how I was using it. Right. And um. And so I'm a person that I love to be on my laptop, but not. But a lot of people like to be on mobile. So like, the biggest feedback I got was like, can I, can we. Can I have a mobile version of this? And I would use it more? So that kind of. That shaped my, that shaped my thinking around. Okay. You know, initially I was my initial customer, but now, you know, talking to other, uh, other, uh, users of the product, I. That kind of can shape, uh, how this product is being built.

Speaker A: Well, what's next, what's. What's next for your company? Are you going to bootstrap this thing? You think you need to raise some capital? Like what, what does the rest of the year look like?

Speaker B: Yeah, right now, um, we're taking in all that feedback and uh, iterating on the new product. And then right now it's just sharing it with others. We're Planning on, uh, just purely bootstrapping at this point, but, uh, really, just really taking this opportunity to figure out what are some of the automations that we can build to make this more, uh, or less a smaller company. Right. Um, and seeing what limits can we push to really be truly AI native. Um, and building out this whole company end to end, from engineering all the way to marketing.

Speaker A: You know, it's so cool. All the things you can do. I've went way down the rabbit hole the last three weeks with Claude Cowork and rebuilt our entire website for full scale. Um, deleted a thousand blog posts that weren't performing well, rewrote 200 and something other blog posts. Um, I spent $8,000 on AI tokens in the last. That's how, that's how much work I got done. Um, but it's absolutely incredible. And it started by, you know, I'd been using Cowork for a little while and collected stuff from my book and from LinkedIn and I would use, you know, link AI to help with LinkedIn content strategy and stuff like that. But I decided I'm like, you know what? I'm gonna pull in all my old podcasts I've done, um, all the newsletters I'd done. I just collected every piece of information I could find and built up my own little library basically. So now if I go and I want to write content for our website, it can reference like, oh, this is what you said four years ago on a podcast or whatever, and all these things, and it's just absolutely incredible. It's just a, um, wild time to be alive, man.

Speaker B: It's absolutely incredible superpower at our fingertips.

Speaker A: The amount of productivity you can get is really, is really high. But you got to understand the strategy and you got to have like the passion and ownership of whatever the problem is you're trying to solve. Because you spend all day just asking other people for permission. You're not going to speed up at all.

Speaker B: Yes, that's fair.

Speaker A: And that's really our biggest problem right now.

Speaker B: Yeah.

Speaker A: Is people having ownership to go execute.

Speaker B: Yeah, I agree. Um, yeah. M making it, uh, your, your problem that you really want to solve. For sure.

Speaker A: Yeah. Yep. Well, thank you so much for being on the show today. Um, if anybody wants to learn more about your product or is it open for beta testers or anything like that that you want to send them to.

Speaker B: Yes. Go to Senvi AI so S E N V I AI all right.

Speaker A: Well, Eman, thank you so much for being on the show today, man.

Speaker B: Thank you. So much, Matt. Appreciate you.

Speaker A: Before you go, if anything in today's episode resonated, I wrote a book about it. It's called Product Driven. It's the number one best selling book for tech leaders who want their teams to stop shipping and start owning outcomes. It's free. Grab your copy at fullscale IO productdriven and click the link in the show notes.