Ep. 46 - Give In to the Vibes
Adam Cartwright, VP of Engineering, talks all things AI, including how we’re utilizing it internally and the importance of incorporating it into your role.
Transcript
Corey Ercanbrack (00:06)
Hi, welcome to the Digital Edge, where we take you to the edge of digital transformation. I'm co kore the Chief Product and technology Officer.
Jennifer Janke (00:14)
And I'm Jen Janke, the chief marketing Officer,
Corey Ercanbrack (00:16)
And today we have Adam Cartwright, who's the vice President of Engineering on the call. We're excited that you're here. Why don't you tell us a little bit about yourself, Adam?
Adam Cartwright (00:24)
Yeah, I'm excited to be here. So with ve, I've been here just over three years, and I get up every morning excited to come work with some of the greatest talent in the industry to build really cool cutting edge software for our customers. I've been doing this for a little while. I mean, before I came to Vision, I worked with a global systems integrator for 20 years, just building high performance engineering teams and delivering really cool software solutions for companies all over the globe. And so with Evasion, I've really found my perfect happy place. It's a combination of who you work with and what you do and doing it with a purpose like we do. I love it.
Jennifer Janke (01:12)
Can I ask you a personal question?
Adam Cartwright (01:14)
Yeah.
Jennifer Janke (01:15)
You're on stage and anytime I talk to you, you have tons of energy. Do you just wake up in the morning and jump out of bed with that smile on your face?
Adam Cartwright (01:22)
I kind of do. I get really excited about what we do and how we do it, and I always look forward to coming to work and just being present and being in the work. So yeah,
Jennifer Janke (01:11)
I love it.
Corey Ercanbrack (01:38)
Got to get that energy somewhere.
Jennifer Janke (1:39)
I know, because I'm going to hit the snooze button.
Corey Ercanbrack (01:45)
Well, Adam, one of the things we're really interesting to talk to you about is the topic of ai. And I know maybe some people are sick of hearing about AI and wanting, but frankly, we got to talk about it, right? It's the thing. It's the thing. So tell us a little bit about your take on ai, especially when it comes into engineering.
Adam Cartwright (02:06)
Yeah, so AI is one of those topics where I can't get enough of it because the industry is moving so quickly and advancing, and there's so much that AI is bringing to the world of software engineering. We use AI in our products already. We do a lot of work in our products today with AI to look at data in the documents and make decisions on that data, how to do it different things within our product, your ability to ask questions about the information that's in a form, for example. But for our software engineering folks, there's AI tools and capabilities that are coming out every day that are making a huge difference in the speed of development. And not only just the speed, it's not just about going faster, but it's improving the developer experience to a point where our devs are smiling again when they're using ai, which is really exciting to see.
Corey Ercanbrack (03:15)
How are they using AI, would you say?
Adam Cartwright (03:18)
Yeah, so if you look at the last year or so, the primary use of AI is around a tool called GitHub copilot. And GitHub copilot is used by over 10,000 organizations from all different sizes from Coca-Cola to Airbnb, and it's mainly used as an AI pair programmer. So it kind of helps you as you write code, it finishes your statements, it helps you document your code, and it kind of predicts what you should do next, and it looks at your code and gives you good feedback and it's helpful. So that's the primary use. But what we've seen lately over the past, oh, maybe six months or so, are new tools that really are agentic and help software engineers to build code faster using natural language, using other techniques, using tools to build unit tests really quickly. In our code base, for example, we have thousands of lines of code that have been written over many years. It's a large complex code base, and no one developer understands all of it, but with ai, a developer can go to one portion of the code and just ask ai, explain this code to me, what does this do? And the AI can help the developer know what it does, how it works, what it touches, so that they can do their work much more efficiently and effectively than they could do it before.
Jennifer Janke (04:56)
What do you say to the question of, well, AI is just going to replace everyone's job.
Adam Cartwright (05:00)
Speaker 3 (05:00): Yeah, it is not going to replace your job unless you're not using ai. I guess. I mean, the way I look at it is AI is going to make everybody more effective, so there's more work to do and really give in the software engineering side of things, give our developers the ability to focus more on architecture and problem solving and more complex tasks that AI really isn't suited for. But AI can do all the mundane things and the things that are really time consuming. So really it's not going to take your job, it's going to help you do your job faster so you can do more work, more meaningful work. It's not just more work, but just the stuff that really matters.
Jennifer Janke (05:50)
So it will only take your job if you're not using it,
Adam Cartwright (05:54)
If you're not using ai. Yeah, those are the,
Jennifer Janke (05:56)
That's when you should be worried.
Adam Cartwright (05:58)
That's the key. And we all need to be learning how to leverage AI just to do everyday tasks better. And we call that becoming AI native, and it's a topic in my own life. I'm looking at things that I do every day, whether it be helping my wife with the grocery shopping or trying to figure out our finances. I'm looking at using AI in every task that I do to be as AI native as possible, and it's amazing how it can help in everything, but particularly for us at Visions, it's making a difference in our software engineering team right now, we're riding stuff and doing things in ways that we've never been able to do before. We've got a couple of guys that are using AI to make them so efficient that they're in competition with entire teams to see who can deliver more points or who can execute quicker and more swiftly.
Corey Ercanbrack (07:05)
That's cool. That's fun. One of the trends I've been hearing about lately is this thing called vibing. What is that?
Adam Cartwright (07:11)
Yeah, this new term that has been very popular out there called Vibe coding, and it was coined by one of the founders of OpenAI, and he really said, there's this new way of coding where you just give into the vibes and you tell the computer what you want it to do using natural language. You either type it or you just speak it and you let it make all the decisions, and you just explain what the output or the outcome you want is, and you give into the vibe and let the AI make the decisions on how to build the code, write the code, and you get to a point where the AI has written so much code that you don't even understand it because you can't go over all the code. And when there are errors, you ask the AI to fix those errors. So it's really a new unique way of developing. It's become kind of controversial in some ways. Some old school devs say, Hey, you can't have code that you don't fully understand, but in the new world, the feeling is your AI can explain it to you. So there's not a reason not to give into the vibe and let AI just build the code for you. I think though, for complex code bases, you have to be careful. You still have to have the human in the loop. You have to have a skilled engineer that can understand and know when AI is telling you things that may not be true if it's hallucinating or vibe coding is great for building smaller projects or personal scripts and things like that, but you still want to be really careful that you make sure it knows that you know what it's doing. Right. Yeah.
Jennifer Janke (09:09)
Why do I feel like you should play Bob Marley in the background where
Adam Cartwright (09:13)
You do that vibe, Cody? It's funny because I've seen some of our engineers change their status in the vibe, right?
Corey Ercanbrack (09:22)
In the vibe. Yep. I love it.
Jennifer Janke (09:24)
That’s awesome.
Corey Ercanbrack (09:25)
So it sounds like it's a big kind of one of these hype market things a little bit it's coming to, but there's some reality to it. We're still, where would you say we're really at on that?
Adam Cartwright (09:37)
Well, we're there, we have tools today that developers can use to do a combination of vibe coding and then doing a lot of personal coding as well, where they're totally in control, and you can do a mixed match of it when you're setting up a new code repository, for example, you can use Vibe coding to complete the structure for you, complete all the templates, set things up in a way for you so that you can be much more efficient and effective, and even finding bugs. Now, here's the thing, AI is great at coding, but it isn't really great at fixing bugs, but it's really good at finding bugs, which is kind of interesting. You think if it can find it, it could fix it, but it's really good at finding bugs. And so using these different coding tools and techniques, you can find the bug, you can help fix it, you can tell it how to fix it, and then you can be on your way to being doing more things. So it's more than hype right now, but it's one that's growing in its popularity to a point where there is a lot of discussion
Corey Ercanbrack (10:50)
Still. So you feel like there's people out there who's building products a hundred percent behind vibe coding and then putting those out in production.
Adam Cartwright (11:00)
So the CEO of GitHub said recently that he expects 80% of all code to be written by ai, and then the CEO of Anthropic came out and he kind of increased that, and he said, I think it's 90% right? And so it's hard to say where we're going to land, but in reality, yeah, there's going to be in the future as we move forward, a lot of code written by ai.
Corey Ercanbrack (11:28)
Well, speaking of ai, last week you and I had an opportunity to head down to Vegas and spend a couple days at a conference called Human X. Tell us a little bit about your experience there and what,
Adam Cartwright (11:38)
Yeah, that was an interesting conference. So this is a new AI conference, human X, and what I really liked about it, it was smaller and the way that they had it created is they had leaders from different companies, particularly AI companies, some of these frontier model companies like Philanthropic and their leaders were on stage being interviewed in a way where you could learn new things about the market, about the technology, and just about where things are going in ai. And then we also had an opportunity to have more intimate q and a sessions with these leaders throughout the week. I thought it was pretty exciting, but I did take away a couple of things. Like the biggest thing that I noticed is everybody's talking about AI agents, the ability to have autonomous agents that are doing the work for you, but barely anybody's figured out how to do it. There's a lot of just still trying to understand how to make all of this stuff fit together. And so I think the opportunity is huge for us, for vision and for really the industry to discover and innovate new ways of leveraging AI to do some of these things that we've only been able to dream up, but now we have the capability to fulfill on those dreams.
Corey Ercanbrack (13:07)
I would say also, just to add on to, I had very similar takeaway, Adam, as I was listening to it in multiple panel discussions, and these are with the CPO of OpenAI. Was there, I mean, these are big companies. Anybody who was anybody of AI was kind of in this conference and they would do panel interviews, but I would say probably eight of the 10 panel interviews I would go to, somebody would ask the question of the audience and say, Hey, will you raise your hand for those of you who have AI running in production today? And it was super rare to see anybody ever raise their hand from that perspective. It was really quiet. So there's a lot of noise on the marketing sound like marketing is definitely outpacing on this one. On reality in my opinion. But there's some real key things that are important. One is that the foundation stuff like the infrastructure, the hardware stuff, the GPU that's really making a lot of hate. People are starting to figure out how to get back from the shortage of GPOs and how to shrink the models down to only run on the GPUs that's in a serious place. I think that the build out of the base models are really starting to spin. Again, people are getting optimized, but there was so many people who would stand on this stage and say beyond that at the application and at the product level, we don't know how to do it yet. We are still trying to figure this whole thing out. So there's a lot of excitement there, but there's also a little bit of caution on what you're hearing. Even when people talk about genic ai, what does that really mean? There's a lot of discussion about what genic AI really is too.
Adam Cartwright (14:50)
Well, and we keep seeing quick innovation in the models, these new reasoning models that think this is a big step forward because the results that we can get out of thinking models is much more improved than what we had before. So in the world of coding, being able to use a thinking model or a reasoning model to build the plan of what you want to code, you talk back and forth and have a conversation with the AI model about what you want to build, and then you switch to one of these other models like Claude Sonnet and put it in act mode where you're now actually using the agent to build the code. So you used reasoning to think about it and build the plan, and then you use a different model to actually create the code. These techniques are starting to get formalized, and it's amazing the results that we're starting to see from it.
Jennifer Janke (15:52)
There's so many ideas going through my head about this. We could again, have five hours of this conversation. We wouldn't even scratch the surface, but I think we're almost at time. So last question
Corey Ercanbrack (16:02)
One question before the final question I'd like to ask. As we think about internal, Adam, I know you've been working with chance a little bit on how are we going to get some AI tools to the employees here, right? Tell us a little bit about where we're at on that.
Adam Cartwright (16:17)
Yeah. We've been working to identify the different tools that are in the marketplace today that would make the biggest or be the biggest benefit for our employees. One of the things we recognize is not everybody has the same need. Different departments need different things, and there's different tools that are better tailored for different needs. So we have a handful of five or six tools that we're going through a process right now to evaluate, to look at the security and the privacy of these tools to figure out what the best practices for usage is. And then we're going to target certain individuals and departments to leverage these tools to help us learn and better disseminate that knowledge out across the company and then make the tools accessible for people in the company who will use them and be able to make really good use of those capabilities. So that's an ongoing process that we have right now. We know a lot of people are using AI already using personal accounts or some of the free accounts that are out there, and it's just really important that everybody remembers to use 'em in a safe way. I think people know you don't put company information like IP or customer names or other things in these public AI models. We just got to be really careful with that. And then of course, there's the privacy settings that you need to make on those tools that we were talking about before where you go in and make it in private mode or check it that it won't share your data, things like that.
Corey Ercanbrack (18:01)
And usually those settings are found underneath your profile. So if you haven't done that already and you're using ai, please take a moment and go do that with that ai. And if you have questions about it, reach out to one of the IT guys. They can help out for sure,
Adam Cartwright (18:14)
Which is really exciting to see. And we have more AI in our company today than you think because there's a lot of tools that we use that have AI in the background. It might not be right in front of your face, but a lot of these tools are using AI in new ways because every company is finding usage of AI tools or just the capabilities that AI can bring. Okay.
Jennifer Janke (18:40)
Now the last question.
Corey Ercanbrack (18:41)
Yeah, Jen, you're on.
Jennifer Janke (18:40)
What are you currently geeking out about?
Adam Cartwright (18:45)
So that list could be really long, but I'll be completely honest. What I'm geeking out about right now is family. So my kids are starting to have kids, and it is so fun because I'm now grandpa and to have babies and toddlers running around in my life that that's what I'm geeking out about all my free time. I'm finding so much joy in just being with those kids, and I'm excited to teach them how to use AI as they get older.
Jennifer Janke (19:20)
How many grandkids do you have?
Adam Cartwright (19:21)
I have four. And the fifth one's on the way.
Jennifer Janke (19:25)
Oh, congratulations.
Adam Cartwright (19:25)
Yeah, it's amazing. They just multiply.
Jennifer Janke (19:31)
It's funny how that happened. It's a whole other podcast topic right there.
Corey Ercanbrack (19:34)
Well, thanks so much, Adam, for being on and giving us your insights into ai. I really appreciate it.
Adam Cartwright (19:41)
Awesome. Thanks for letting me be here.
Jennifer Janke (19:46)
Stay tuned for the post-show wrap up.
Corey Ercanbrack (19:50)
So Jen, yet another great interview with Adam AI is a topic that, man, it feels like it's, I dunno about you, but I don't go through a day, honestly, don't go through an hour anymore without talking about AI in my lifetime. I dunno if that's like you or not.
Jennifer Janke (20:07)
Yeah. And I think you can have two approaches where you hear the word AI and you kind roll your eyes and be like, not again. Or you lean into it and start to figure out how to enrich your life with it.
Corey Ercanbrack (20:17)
Yeah, I agree. So what was your takeaway?
Jennifer Janke (20:19)
Well, as he was talking, he was talking about engineering and dev and how they're using AI to create code. And I was like, wait a second, I can do that too. Not ever at that level, nor do I ever want to be at Adam's level. But just the other day, working on a side project of building a website, I couldn't, the theme I had and the components already built weren't there. I wanted something unique and different and I just went over to my little Chas GBT and gave it a couple prompts and boom, there was the code that I needed to create that custom section. So it's fun for someone at my skillset and level to be able to do that. I wouldn't have been able to do that without ai. And then the other thing I was thinking is we talk about it a lot at work, but we also talk about how you use it in your personal life. And I hate mill. I love meal planning. I hate to eat the same thing every day though. My husband could eat the same thing every day, but I hate meal planning and I hate writing the grocery list that's dawning off the table in two minutes now versus me going through recipes and creating the list. So even something like that just gives me more time to spend with my family. It's not always to do different work. Sometimes it's just to enjoy life more.
Corey Ercanbrack (21:30)
Yeah, I agree. And I think a lot about our conversation about the anxiety of people today thinking AI's going to come take my job and my life's going to change. And I think that's real. I think there are people who are concerned about that. And I agree with Adam when he's like, no, AI's not going to take your job. It's going to be somebody who's using AI who might actually do that. And the thing that I've realized lately is that when we think about ai, more so than ever, if you really want to utilize ai, you have to stop and think about what's my process? What's the way I go about doing that? Because if you're going to tell a computer to do that or an intelligent machine to do that, you got to be able to describe that, right? And so I think that's forcing everybody to stop and look at what are all the things I do and how do I do them, and how might I be able to use ai? Speaker 1 (22:32): And I think that's putting everybody in a very unique position to really grow and to really expand. I'm excited about the future. I think we're going to see some incredible stuff happen. And it's because of these basic things like pausing. I mean, if you think of our product and our solution, like automate right now, when we go talk with customers about automate, one of the biggest hangs up is people don't understand the process that they have in place. They haven't documented, they don't understand it. And then it takes forever to get it documented and understood. And that's just something I think that's going to continue to flush out, and it's going to be something that's going to be really exciting. So that's what gets me really excited about some of the movements I'm seeing with AI and yeah.
Jennifer Janke (23:13)
Yeah, just think how far we came in the last couple months. Yeah. Who knows even what, by the end of this year, what will happen.
Corey Ercanbrack (23:20)
Yeah, it's incredible. It's a good time to be here for sure. This is Jen and Corey jumping off the digital edge.
