00:00:00:02 - 00:00:02:13 Martin Adam's desperate to disagree. Go on. 00:00:02:14 - 00:00:07:13 Papanii Can we chat about space by cursor? I have a very strong opinion of this. 00:00:07:15 - 00:00:15:02 Adam Here we go. The AI vulnerability and threat. Apocalypse. It is not here. And there is evidence for this. 00:00:15:03 - 00:00:23:02 Martin Tim Cook is retiring. Is this a trend putting builders in charge of companies? Because AI is driving that building mentality? 00:00:23:03 - 00:00:29:20 Papanii Yeah. You can say you're not going to hire any engineers, but you are not going to be able to sail because people are going to burn the hell out. Right. 00:00:29:21 - 00:00:36:16 Martin Where exactly are the seniors for 2030 coming from? And I mentioned the log for Jay, and. 00:00:36:17 - 00:00:44:02 Papanii Everybody was like, what the hell? And he just screwed up everybody. 00:00:44:04 - 00:00:52:07 Adam Hey, everybody, this is ship talk brought to you by harness, where we break down how software delivery is changing in the AI era. I'm Adam Arellano. 00:00:52:08 - 00:00:53:16 Martin And I'm Martin Reynolds. 00:00:53:17 - 00:01:04:12 Adam Let's get into it. So our first story that we want to discuss is that space buys cursor OpenAI notes out and anthropic stays question mark. 00:01:04:18 - 00:01:05:21 Martin So question. 00:01:05:21 - 00:01:26:07 Adam Mark maybe two question marks. That's a strange one. Let me let me get some details out real quick. So SpaceX actually brought cursors parent company any sphere for 60 billion. Chump change. Sure. Cursor is an AI powered code editor built as a fork of VS code that acts as an intelligent pair programmer for software development. As soon as that happened. 00:01:26:08 - 00:01:51:02 Adam OpenAI then cut off cursors access to its models effective November 12th, citing a change of control clause and distrust of Musk's affiliated companies. Anthropic did the opposite. It's keeping clod in cursor and increasing compute support. Anthropic has cut off other partners before, like windsurf and XII staff were cut off via cursor previously, but this time they opted to stay. 00:01:51:04 - 00:01:53:06 Adam What do you think about this? Martin. 00:01:53:08 - 00:02:11:20 Martin So I have a couple of things. I mean, the first one honestly has to be OpenAI was only about 5% of cursors traffic. So, you know, in terms of impact, I think it's going to be small. But, you know, is it really a business decision or is it actually just a political decision? 00:02:11:21 - 00:02:33:11 Adam Well, that makes sense, especially because anthropic pays space 1.25 billion per month for compute, and they're on contract for that until 2029. And so that kind of makes anthropic cursors tenant or like there's a landlord tenant like relationship there. So maybe they shouldn't get Bucky with with cursor. 00:02:33:12 - 00:02:54:01 Martin I know that's that feels like at some point that's going to become a conflict of interest, but I feel like there is going to be quite a big push ultimately in cursor for it to use. EXi not anthropic, not well, it won't be using OpenAI now, but if you sort of mean it'll use X's AI model, not anthropic. 00:02:54:01 - 00:03:00:07 Martin Despite that kind of amount of usage, I think it's like 70% of the usage is is anthropic AI. 00:03:00:08 - 00:03:21:16 Adam So so far. But they've been pushed in other directions in the past where they've been, I think not at space in particular, but at Tesla, they were pushed to use grok exclusively, which is an interesting choice for the safety of motor vehicles to require them to use grok, which has like what, a 54% hallucination rate. That's that's smart. 00:03:21:18 - 00:03:25:07 Martin Which yes, was not. So I know which Teslas to avoid. 00:03:25:09 - 00:03:30:20 Adam It happened this year. So so model year 2026 and prior maybe maybe don't okay. 00:03:30:21 - 00:04:01:10 Martin Essentially $60 billion for a code editor fork. I know I'm belittling what cursor is there, and I don't mean to do that. But you know, essentially 60 billion I I'm not sure they're buying the coding tools so much as they're, you know, they're buying that infrastructure that that powers it, if that makes sense. I feel like it's it's less of a play for cursor, but more of a play for the market that cursor has and the infrastructure that cursor has. 00:04:01:11 - 00:04:06:16 Adam To give us context on this. We brought on our guest. Okay. How you doing? 00:04:06:18 - 00:04:08:10 Papanii I'm doing good, man. How are you? 00:04:08:11 - 00:04:28:01 Adam Happy to have you here. It's good to talk again, but Pontius is the EVP of product engineering at rocket with 20 plus years of experience. Previously, he was the CTO at Venmo and held leadership roles at PayPal. A seasoned engineer, he's known for delivering innovative solutions that enhance customer experiences with a global perspective. From living in Africa, Europe and the US. 00:04:28:01 - 00:04:35:00 Adam He holds degrees in computer science and internet engineering from the University of San Francisco. Hey, Papanii welcome to the show. 00:04:35:00 - 00:04:35:13 Papanii Hey, guys. 00:04:35:13 - 00:04:44:13 Papanii I'm so glad to be here. But before we jump into the numbers, Martin, can we chat about space buying? I have a very strong opinion of this. 00:04:44:14 - 00:04:45:14 Adam Oh, here we go. 00:04:45:19 - 00:04:46:12 Papanii I love it. 00:04:46:12 - 00:04:50:09 Papanii I think that there's a there's a piece that people are not thinking about. 00:04:50:13 - 00:04:51:20 Papanii Oh. 00:04:51:22 - 00:04:54:18 Adam You love the idea of space buying cursors. 00:04:54:18 - 00:04:55:11 Papanii I think. 00:04:55:12 - 00:04:56:09 Adam Tell us more. 00:04:56:15 - 00:04:58:05 Papanii Okay. 00:04:58:07 - 00:05:18:18 Papanii Let's let's try and break this down because the one thing about Elon, he's very strategic. I think he is. And if you think of what he has done and what he's proved is possible, nobody's really done what he's done. But let's think about what SpaceX is. So, Adam, this is going to be a more interactive with you and modding. 00:05:18:18 - 00:05:23:04 Papanii But let's think about this. What is space right now. What are they trying to do. 00:05:23:06 - 00:05:29:14 Adam Oh they're trying to get us into the stars. They're trying to make space accessible at a at an economical rate. 00:05:29:15 - 00:05:30:14 Papanii Fantastic. 00:05:30:14 - 00:05:51:12 Papanii And obviously they have their own model, the grok model. And I think you were teasing them about, you know, but you can't make a model better if you don't use it. Right. You've got to use it to make it better. Right. So you have to do something. So they have that. They have that what the cursor do. In fact when cursor first came on the market what did they do. 00:05:51:18 - 00:05:52:23 Papanii They were. 00:05:53:00 - 00:06:25:11 Papanii The IDE for identical essentially. Right. They they I wouldn't say they cracked the nuts, but they technically cracked the code to and everybody was using it. And then they gave you access to be able to use any model with their agents. Right. Okay. So if we believe those things to be true independently, if we believe that unless here's the other piece we all know, and this is where I think Apple has a superpower that people haven't realized. 00:06:25:13 - 00:06:40:21 Papanii We are not all going to be just be using an app for AI. It has to be embedded in the experience. It has to be ever present in your experience. Which is why I believe Apple is actually going to come on top with what they are going to be doing. 00:06:41:00 - 00:06:42:00 Papanii Oh. 00:06:42:01 - 00:06:48:03 Adam Apple is going to come over top of space. Oh, that is a feisty prediction I can't wait. 00:06:48:04 - 00:06:49:03 Papanii No, I'm. 00:06:49:03 - 00:06:53:11 Papanii On top when when it comes to AI in your personal computer. 00:06:53:13 - 00:06:54:08 Papanii I see. 00:06:54:12 - 00:06:55:22 Papanii But they are going to come on. 00:06:56:00 - 00:07:22:03 Martin That is what Apple do right? I think I've mentioned this before like I, I think I went once three years without updating the Apple like my iPhone. And I looked at it and I was like, oh, this isn't you know, it still doesn't seem any different. Then I got my new one. And there is, you know, 10,000 small incremental changes that make the way you use it, the functionality of it incrementally better all the time. 00:07:22:03 - 00:07:43:15 Martin And everybody says that Apple right now doesn't use AI. They maybe they don't use generative AI a lot, but they use AI everywhere in their software. It's built into their home, you know, their whole operating system, the way their chips are built, you know, they're optimized for it. So, you know, I think they can come out on top. 00:07:43:17 - 00:07:46:11 Papanii It's not can Martin they will. 00:07:46:13 - 00:07:48:08 Papanii They oh. 00:07:48:10 - 00:07:51:23 Papanii Look, I am going to bet I'm willing to put away. 00:07:52:00 - 00:07:53:16 Papanii On this. Here we go. Here we go. 00:07:53:17 - 00:08:16:05 Papanii They know what they're doing. Right. If you think about. Think about it. Think about. I don't even work for Apple. I'm a Apple fanboy. But think about it. Think about the way, over the years of time, all the various strategic acquisitions they've done, what data they have, what actions they have. Siri was just the first of many they have. 00:08:16:06 - 00:08:33:13 Papanii They know how you use your phone. Do you know what you say to your phone? They know your Narasimhan. They have all this information about all the apps you use. They are trying to build an ecosystem that is not going to feel as if. Oh, I have to launch an app to use AI. It is going to be there. 00:08:33:15 - 00:08:46:08 Papanii There's nobody that can do that right now. Microsoft can't even do that right now. And so when it comes to the personal computer and how you do that, they are going to that. But anyway, I digress. Let's go back to where we were. 00:08:46:12 - 00:08:56:10 Adam Talking about just real quick, real quick. I just want to say that this entire time I'm still sitting here judging Martin for not updating his phone for three years, I that offensive. 00:08:56:12 - 00:09:01:09 Martin Oh hey, the one I had worked great. 00:09:01:11 - 00:09:36:13 Papanii So we understand that cursor in my opinion, somewhat early on cracked the code on the user interface that humans could use with generative AI, right? With models and all that kind of stuff. So if you know that you are trying to build user interfaces, you are trying to figure out how people, humans interact with models, and those guys have figured it out, why not bring it into your company that is trying to build a system that takes us to the stars or beyond? 00:09:36:15 - 00:10:01:22 Papanii Why not? Then think about how do you integrate that mindset, whatever secret sauce they have in the way they build that, that same mindset into the user interfaces that you're going to build across your fleet of things. Now, I'm not a big fan of the Tesla URIs. The Tesla capabilities are very great, but the UI is not my favorite. 00:10:02:03 - 00:10:22:02 Papanii But imagine that as you start bringing the model infrastructure into your ecosystem, your models are getting better. You have this group of individuals who have figured out how to build user interfaces, and you now blend that all in together. I think it's a brilliant idea. And then you start thinking about the whole concept of physical AI and what you can do with physical AI, right? 00:10:22:04 - 00:10:38:18 Papanii He is not just going to do what he's doing, he is going to go into other aspects and you're going to need interfaces. You're going to need things that are generated. I just I think it's a brilliant thing. I think everybody is just thinking about it from a from what we have now, today, these folks are thinking way ahead. 00:10:38:20 - 00:11:05:00 Adam Same be a bit cheeky with you though. Like Musk is the example of a builder who turned into a meme, and I think that he's lost his edge on that. His companies like SpaceX, his company like space. I will give SpaceX credit. SpaceX is the one Musk company that has survived his idiocy. It's the one company that has been able to maintain their engineering excellence better than any other company that he's taken over there. 00:11:05:04 - 00:11:06:02 Adam Several of those. 00:11:06:02 - 00:11:07:03 Papanii Habits, probably. 00:11:07:03 - 00:11:21:16 Papanii Right. But you also have to understand that people's public persona, what is on the screen, is very different from what is actually in the background. I don't know Mark personally, but I, I'm sure this is all a facade. 00:11:21:17 - 00:11:34:04 Martin So the Elon thing, I think he's a great salesman. He's a great visionary. I'm not I am less convinced by his delivery. He always over sells but always over sells. 00:11:34:04 - 00:11:36:00 Papanii But I think every single time. 00:11:36:02 - 00:11:36:09 Martin Like. 00:11:36:10 - 00:11:41:21 Adam I give him credit for being like effective. As a result, he had my ass full for like ten years. 00:11:42:02 - 00:11:43:20 Papanii Think about this though. 00:11:43:22 - 00:11:48:20 Papanii Everybody wanted to build an electric car. They said it was impossible for years. 00:11:48:21 - 00:12:14:01 Martin But I think that that might come with him. And and like Tesla is a great example is, you know, that could still be an amazing growing business and it could still be innovating. And it is innovating, but it's falling behind because it doesn't get his attention, because it's not the new shiny thing that he's interested in doing. And I think that is a shame. 00:12:14:03 - 00:12:17:10 Martin I think it's a shame. I genuinely think it's a shame. 00:12:17:11 - 00:12:18:09 Papanii I think. 00:12:18:11 - 00:12:19:02 Adam Was that go ahead. 00:12:19:05 - 00:12:24:19 Papanii But this is this is a this is fine. It's fine. I think I want to hear it. Come on. 00:12:24:20 - 00:12:44:16 Papanii That's something you hire people. It's not like he's run it himself. He's hired people. I think you should actually be the people out there. If you're saying what you're saying. The guy started something and it has changed the way people commute. There are people who haven't paid for God. He was that started and prove the point. That's why we have Rivian's right now. 00:12:44:17 - 00:12:50:04 Papanii That's why all these other companies are doing all these electric cars. It's because he proved that it was possible. Right? 00:12:50:05 - 00:12:50:11 Papanii Right. 00:12:50:11 - 00:13:06:14 Papanii Yes. He's done it. He's he has a team. They figure it out. He's moving on to somewhere else. And for him traveling to the stars thing, that's not my thing. But that's his thing. And he's doing a bang up job. That being said, yes he has. We all have our quirks, but. 00:13:06:15 - 00:13:07:12 Papanii He has a. 00:13:07:14 - 00:13:08:08 Adam To me. 00:13:08:11 - 00:13:11:03 Papanii Yeah, yeah yeah okay. He has a lot. 00:13:11:03 - 00:13:12:22 Papanii Of zeros after it. So everybody. 00:13:13:00 - 00:13:14:02 Papanii You know complains. 00:13:14:02 - 00:13:15:07 Papanii About what he does. 00:13:15:10 - 00:13:17:06 Adam You just called me a poor man's quirk. 00:13:17:07 - 00:13:20:10 Papanii No. I'm just. 00:13:20:12 - 00:13:23:04 Papanii We all are. 00:13:23:06 - 00:13:24:07 Papanii Anyway. 00:13:24:09 - 00:13:28:12 Adam Ellen. Ellen has quirks. I have Teemu quirks. 00:13:28:14 - 00:13:29:22 Papanii You get my point. 00:13:30:00 - 00:13:32:08 Adam I do get your points, sir, I do, I do. 00:13:32:14 - 00:13:33:00 Papanii You take it. 00:13:33:00 - 00:13:47:17 Adam Yeah, I think I think we have beaten the horses enough. All right. It is time for harness. By the numbers. 00:13:47:19 - 00:14:21:16 Martin Harness by the numbers. All right, so 32% of organizations skipped buying software this year. They built it instead using a genetic coding tools. Also, large enterprises jumped from 20%, 27% to 40% agent adoption in a single year. But small firms, they're still sitting at 22% and essentially staying flat. And that gap between small to medium businesses and enterprises is growing. 00:14:21:17 - 00:14:37:00 Martin And my final stat, before we get into our main topic and introduce the guest is Codex just crossed 2 million weekly active users. That's tripling the growth and five times the usage since January alone. 00:14:37:02 - 00:14:49:00 Adam Organizations are clearly using more AI, but the trend still seems to be building versus buying. Yeah. So what do you think about this? The the drop in enterprises buying software and building their own? 00:14:49:02 - 00:15:18:16 Papanii I mean, it was only a matter of time, right? I mean, we've had this conversation and I think a lot of folks have, you know, generally say, oh, SaaS is dead. I think it depends on the type of SaaS. I think they are going to be capabilities that companies and corporations feel they can now build. Because the time of ideation from ideation to implementation has shrunk so significantly orders of magnitude. 00:15:18:16 - 00:15:48:22 Papanii And so they may feel it is cheaper to do some implementation themselves. However, they are going to be certain things like regulation and policies that people forget about that is not easy to do those things. And that's where some of the SA's capabilities come in. And I think if you think about some of the complicated SaaS providers, I'll take Sierra AI, for example, where we are dealing with various types of capabilities. 00:15:49:02 - 00:16:07:07 Papanii You're not going to roll up your own voice, AI or text AI and stuff like that because of all the things that below the ocean, right? And so it really depends. So I take it as a grain of salt, understand your business, understanding what you need and what you can build yourself and what your engineers are capable of doing. 00:16:07:09 - 00:16:16:17 Papanii At the end of the day, if there's a capability that you can pick off the shelf so that you can focus on your core business, do that, focus on your core business. But that's just my opinion. 00:16:16:23 - 00:16:17:06 Papanii So. 00:16:17:07 - 00:16:35:23 Martin So I. Have a couple of questions on the on the back of that actually. So, you know, and I'm rolling this up with I did a panel about a year or so ago, and one of the people on that panel was saying, you know, one of the things that AI's enabled, which is, you know, their POVs used to be three months and now they're two weeks. 00:16:35:23 - 00:17:00:06 Martin If they're proving out a concept for building something was way, way faster. But you know, which of the sort of software categories that you think are at risk and which are the things that you think are going to be, you know, they're they're those good, stable capabilities that are just too hard. So which are the where is the risk if you like, if you're a SaaS company today and you're thinking, oh, I might be in trouble, you know, what are those areas? 00:17:00:06 - 00:17:03:20 Martin Where do you think people are building those homegrown systems? 00:17:03:22 - 00:17:04:11 Papanii Yeah. 00:17:04:13 - 00:17:08:11 Adam Come on. We want to hear. 00:17:08:13 - 00:17:10:15 Papanii I'm not going to do that. 00:17:10:17 - 00:17:35:04 Papanii Well, look, I think you have to understand what your moat is and what you are really good at. If it's something that somebody can use a weekend or a week or a month to build, you're in trouble. If it's something that basically you have built that scales to millions of users, you have figured out how to do across various geographic plains. 00:17:35:04 - 00:17:56:00 Papanii You've understood how to deal with, you know, internationalization. You've figured out all the security capabilities and vulnerabilities out there. You've understood how to deal with, you know, all the various data policies like GDPR, CcpA and all that kind of things. And you should be fine. But if not, that's just my opinion. Or that your host. 00:17:56:00 - 00:18:07:13 Adam You're the moat. I like the word. I like the word moat. It's like, how hard is it going to be for somebody to cross that moat? Right. And if it's if it's full of FedRAMP and international law, that's a big moat, big alligators. 00:18:07:17 - 00:18:08:09 Papanii And the other thing. 00:18:08:14 - 00:18:10:08 Adam About speed, then it's not. 00:18:10:09 - 00:18:34:16 Papanii Correct because speed and and and time to build was a reason why a lot of folks didn't do things. Yeah, that was the reason. Right? Getting the domain knowledge right. Getting the right people who understand it. Right now, technically intelligence is a commodity and it's cheap. Now. That's what models have given us. And so I don't need an expert. 00:18:34:16 - 00:18:40:00 Papanii I have an expert. I have a PhD level expert with me right now, all day, every day. 00:18:40:00 - 00:18:48:15 Martin I think the getting it spun up is faster now, but I still think the long term challenges haven't necessarily gone away. Most of them are still valid. 00:18:48:16 - 00:18:48:22 Papanii No. 00:18:48:22 - 00:19:16:08 Papanii I would agree to an extent, Martin. And the reason being is listen, I think right now the senior, the principal engineers, the staff engineers, the architects, they are in their golden years right now because they understand how to build systems. They have built it, they have made the mistakes. And so when they are dealing with AI and using AI, they know the right ways to write their spec. 00:19:16:10 - 00:19:42:14 Papanii They the personas to use when the defining the aspects, they understand the right evaluations that you need to create to do these things. It is not that hard to do. It's I'm not saying it's easy. It's work that needs to be done. But once you shift left and you have all your your specifications done, I do not believe that it's going to be that hard to replicate a lot of capabilities, because we've done it before. 00:19:42:16 - 00:20:02:09 Papanii Adam and I, when a paper with you and I was, we had to deal with scale. When it came to Venmo, we had to deal with security. We had to deal with all these things. We know these things. And so I think that architects, that senior engineers who have a lot of tenure, who actually really dive into AI, they're going to be the superpowers, they are going to be the ones. 00:20:02:10 - 00:20:36:07 Martin So if like if you're saying that seniors are the superpower, like one of the articles I came across this week from the World Economic Forum was, you know, essentially saying that it's like a slow moving problem in engineering talent that US entry level software engineering postings dropped 67% between 2023 and 2024. So kind of insane, honestly. So, you know, if you're hiring has been declining essentially since like 2022, and it takes 5 to 9 years to build up a new senior. 00:20:36:11 - 00:20:40:20 Martin Like where exactly are the seniors for 2030 coming from? 00:20:40:21 - 00:20:41:03 Papanii I have. 00:20:41:03 - 00:20:44:01 Papanii So many thoughts on this. Is that really for me is that is. 00:20:44:01 - 00:20:44:22 Papanii That is that. 00:20:45:00 - 00:20:45:20 Adam Is why we see this. 00:20:46:00 - 00:20:46:18 Martin You take that. 00:20:46:19 - 00:20:47:11 Adam Go for it. 00:20:47:12 - 00:20:47:23 Papanii All right. 00:20:48:00 - 00:21:17:13 Papanii So let's think about. This because this is multifaceted right. The data that you are just sharing Martin, is based on history that has happened before the explosion really of AI right now, which also means and this is this is this goes to another conversation that you and I had when you really think about education and how people learn, it's going to also evolve, right? 00:21:17:14 - 00:21:37:03 Papanii So the fact that it took us nine years to do the way study to get to where we are right now, I am convinced that it's going to shrink drastically to 1 or 2 years, because all their experiences and things that we had to go through to learn those things is now going to be much more visible to us. 00:21:37:04 - 00:21:56:20 Papanii Right. And step on, number one, the way we learn and educate ourselves is also going to rapidly change, right? Because you'll be able to get more information a lot faster. Just remember, before even the internet, we used to use that we had to buy books. And then when the internet came, it sped up how fast we learned certain things. 00:21:56:20 - 00:22:18:02 Papanii This is going to do the same thing. So I don't think it's going to take that long for a junior engineer to become a senior engineer. However, there is going to be an upper bound where there is going to be very specific things that, hey, you just have to get your feet wet. You have to go through your trials to become that expert because the barrier for entry has lowered. 00:22:18:03 - 00:22:44:02 Papanii Right. But the barrier or the the bar for excellence has also extended because average is now risen up. Right? Technically because of all those capabilities. So that's number one. Number two is a lot of companies are assuming that their business domain is not going to grow. So if you're going to have the same domain and not grow then yeah, that makes perfect sense. 00:22:44:02 - 00:23:10:19 Papanii But just like Moore's Law, what happened? Whenever we bought a new computer, we didn't do the same problems. We solved new problems. We saw more complicated. We loaded bigger spreadsheets, you know, all that kind of stuff. So it's natural that the problem space is going to expand. Right. And so yeah, you can say you're not going to hire any engineers, but you're not going to be able to scale because we all know that the what do you call it, the vampire phenomenon of AI is real. 00:23:10:20 - 00:23:13:00 Papanii People are going to burn the hell out, right? 00:23:13:02 - 00:23:13:17 Papanii Right. 00:23:13:21 - 00:23:34:10 Papanii And so you have to bring in those other engineers so that your senior engineers can coach them and teach them how to do certain things, to be able to understand those dynamics. But here's another thing. The junior engineers are going to be much more efficient in using these new digital tools much faster than the old guys. Right? And girls, right. 00:23:34:11 - 00:23:52:23 Papanii Because of the end of the day, they say you can't train an old org new tricks, right? We are so used to building new things the way we use. They're going to try different things much faster than us. It's just like your kids, right? When you have your kids, for some reason, they're able to understand how to use the phones and things much faster than adults. 00:23:52:23 - 00:24:02:09 Papanii It's the same thing. And so you're doing your self service. If you don't hire new talents to bring new blood and new ways of thinking into the ecosystem. And that is. 00:24:02:09 - 00:24:03:15 Papanii The key thing. 00:24:03:17 - 00:24:25:00 Martin I mean, before Adam jumps in, I will say that like one of the interesting things I pulled from that article is that it is kind of bucking that trend. So they actually tripled their junior intake in 2026, but they have also restructured the role. So kind of less on the routine coding and more on interpreting requirements and validating AI outputs. 00:24:25:02 - 00:24:37:15 Martin So like from my point of view, that's you know, I think that's a great play. I think that puts them in a really strong position going forwards. And it's certainly the way I would bet. 00:24:37:16 - 00:24:38:16 Papanii Yeah, absolutely. 00:24:38:16 - 00:25:09:20 Papanii And I think the even the, the concept of engineering is going to change. Folks like the and I wrote an article about this in my blog post. Like you have the concept of a solution builder. You're not going to have like, oh, I'm a software engineer and this is all I do. You're going to start thinking about solutions to problems, which means kind of like end to end a little bit, but you should be able to span across multiple disciplines because again, you have a PhD level person by your side, no person persona by your side. 00:25:09:20 - 00:25:16:13 Papanii Whenever you need in any topic that you need. There is absolutely no reason why you shouldn't be able to do much more interesting things. 00:25:16:13 - 00:25:41:08 Martin I'm loving that evolution thing. You know, I I'm not the only person who says this, but if you work in building software, it is always change. It is constant change. This is just the next wave of change. And if you want to stay relevant and, and, and, you know, go forwards as a, you know, in building software solutions, you have to constantly adapt and evolve. 00:25:41:09 - 00:25:45:02 Martin It's just it's part of why the role is actually so exciting to do. 00:25:45:03 - 00:26:07:01 Adam So based on that principle, Martin, and what you just said basically was change is constant. Those who are able to change as well are the ones that will survive. I think it's worth looking at the institutions surrounding this that are least capable of change. And the first one that comes to mind for me is universities and graduate universities. 00:26:07:01 - 00:26:36:18 Adam So the university, the bachelor programs and the master's programs take years to change. Curricula take years to get on a new material. Do you know what educational institution doesn't? Community colleges. And so I think what we'll see is those institutions starting to be less and less relevant. I personally have in the past taken kids with associates, not kids, adults with associates, degrees over bachelor's degrees from Ivy leagues, universities because they're capable of thinking differently and capable of thinking around things. 00:26:36:18 - 00:26:49:01 Adam So I think looking at what particular enterprise is capable of rapid change, and even what institutions surrounding all this are capable of rapid change, will be where it shakes out that some survive and some don't. 00:26:49:06 - 00:27:06:07 Martin But is this really a challenge to basically saying, how relevant is that? You know, that four years of school that I just spent a huge amount of money on, you know, in the future, is it actually going to help me get a new job or is actually a different approach going to do that? 00:27:06:08 - 00:27:08:12 Adam I think that's already not true. 00:27:08:14 - 00:27:10:07 Papanii I this is this is. 00:27:10:07 - 00:27:10:17 Papanii A great. 00:27:10:17 - 00:27:13:05 Papanii Topic because it's. 00:27:13:05 - 00:27:17:09 Adam Almost it's almost as if we've had this conversation over a dinner table like seven times. 00:27:17:09 - 00:27:19:00 Papanii But that's. 00:27:19:01 - 00:27:36:21 Papanii Us to go crazy. Listen, I think there is there are two there are two competing ideas that we we keep on trying to merge into one. And I think we are doing a disservice there. There is the the desire to find a career and a job like career and a job are two different things, by the way. But right. 00:27:36:22 - 00:28:03:13 Papanii And then the desire to grow and learn how to adapt as a human being in society. What universities give you isn't just about I'm going to set you up for a job. It is more about learning how to be with different cultures, learning how to critically think, learning how to adapt in situations that you won't. You are not going to get at home. 00:28:03:13 - 00:28:23:21 Papanii It is the first time a lot of teenagers leave home and be in a different environment. And here's the other part. It is ludicrous to me that I'm at the age of 16 or 17, that I come and says, I'm going to define the career and subject for the rest of my life, where I'm literally lived only a quarter or less of my life. 00:28:23:21 - 00:28:25:00 Papanii It's ridiculous. 00:28:25:02 - 00:28:25:07 Papanii Yeah. 00:28:25:07 - 00:28:36:19 Papanii So universities to me give you that ability to learn how to learn, learn how to think, learn how to be critical at what you do. 00:28:37:00 - 00:28:38:11 Martin I actually. 00:28:38:11 - 00:28:40:10 Papanii Think that is there. 00:28:40:11 - 00:28:41:13 Papanii That's my challenge there. 00:28:41:14 - 00:29:03:07 Martin I mean, that is the superpower of universities, though. And maybe they need to change what they're doing and really focus on the fact that you're learning how to learn, because that's what actually sets you up for a career. Not here's a bunch of facts, but here's a bunch of skills that allow you to go forward. Adams definitely desperate to disagree. 00:29:03:08 - 00:29:05:21 Martin Go on. 00:29:05:23 - 00:29:29:10 Adam I agree with the with the lofty ideals, but Bonnie, I really do. I would say that having taught in a university the involvement of parents and how much parents influence who gets into what and how much they influence what teachers can and can't do, I think removes some of that and makes it cheaper. But I agree with you that there's very few other places in the world where you can get that. 00:29:29:10 - 00:29:32:00 Adam I just think that it takes longer than it used to. 00:29:32:03 - 00:29:56:14 Martin So because we're talking about jobs, I wanted to bring up the new Apple CEO. Tim Cook is retiring, so they brought in John Turner. Now, what is interesting, they have an actual builder who's running the company. You know, he he was responsible for iPhone hardware. He's been responsible for the like the M-series chips. He was responsible for building the Vision Pro stuff. 00:29:56:18 - 00:30:19:16 Martin And now what you have is an actual builder kind of sitting in charge of Apple, which, you know, has grown under Tim Cook from 100 and something billion value to 400 and something billion value. And what he's come in with is, hey, I'm a hardware guy who knows how to build hardware, and he's got a big AI problem to solve. 00:30:19:18 - 00:30:34:13 Martin That is apples biggest challenge. So, you know, do we think it's a good fit? Do we think of is is this a trend putting builders in charge of companies because AI is is driving that building mentality. Do you think it's going to work. 00:30:34:15 - 00:30:36:00 Speaker 5 This this is. 00:30:36:03 - 00:30:40:09 Adam Amazing to me. And I love the fact that we have a builder in charge. 00:30:40:10 - 00:31:06:04 Papanii I'm very excited. I think it's the best thing that will happen to Apple. I think look, I think personally Apple is very strategic in the things that they do. They look and understand the customer better than anybody, but they never the first to do it. They see what other people do, they look at their core competencies and they just do it better. 00:31:06:06 - 00:31:30:10 Papanii And having an engineer lead this organization is going to be phenomenal. Because he's not just an engineer. He understands the business better than a lot of other people actually taking care of him. Credit for. And he's under the tutelage of all those leaders. So he's not coming in fresh. He's been tutored, he's worked with some of the greatest minds, is going to be an epic. 00:31:30:11 - 00:31:36:00 Papanii I'm just super excited about it. And here's the thing. Do you realize that the stock didn't change that much? 00:31:36:02 - 00:31:38:03 Papanii Yeah. That is proof. 00:31:38:05 - 00:31:40:22 Papanii That there's confidence in this person. 00:31:40:23 - 00:32:03:04 Martin I like that as a that is a good point because normally there's a let's just say a wobble on a change of leadership at a company like that. So that is super interesting. But do you think it's going to be a trend. Do you think we're going to see more of that. So you know, maybe less of the because quite often it's the sales business people that come into those CEO roles. 00:32:03:08 - 00:32:15:21 Martin And now do you think you're going to see more of those like the, you know, the co CIO type people coming in into that role instead? Do you think it's going to be a trend for those for for organizations? 00:32:15:21 - 00:32:16:07 Papanii I don't. 00:32:16:07 - 00:32:45:16 Papanii Know whether it will be a trend, but I think depending on how you want your business to evolve in the age of AI, you see AI, what AI actually does right now. It helps you bring your imagination to life. And people who are creative, people who build stuff are the ones who think about the crazy things. Sometimes I'm not grouping everybody in one, but these strategic advisors and business folks, they just focusing on the dollars, the revenue, the margins. 00:32:45:17 - 00:33:05:05 Papanii Right. What will make us more money? Right. And so, you know, I'm going to do a selfish plug here. It's like our current CEO Vernon Krishna rocket was a product guy. He's hands on product and he's pushing all of us to think outside the box. And he comes up with ideas. He comes to product reviews and stuff like that. 00:33:05:05 - 00:33:22:00 Adam And I think it won't be a trend. Martin I think it'll be a leading indicator of companies that will take off, because there will be some people who will bring an engineer in or bring a builder into the to the to the top seat. That will effectively disrupt whatever area there. 00:33:22:00 - 00:33:25:18 Papanii Then even Marc Benioff, isn't he like hands on technical? 00:33:26:00 - 00:33:27:00 Papanii You sure? 00:33:27:01 - 00:33:51:07 Adam Yes, I'm very sure. Parker, his co-founder. So technical that we were in incidents at Salesforce and we had to call Parker and nobody else could figure out exactly what was going on. We call Parker the dude that wrote the original OG code, and we were like, hey, this is the problem. And he's like, did you try that? 00:33:51:08 - 00:33:53:02 Papanii And, you know, I think, Larry, I listen, sorry. 00:33:53:02 - 00:33:55:07 Papanii I was thinking about Oracle, Larry Ellison, he is technical. 00:33:55:08 - 00:33:55:22 Papanii Oh yeah. 00:33:55:23 - 00:33:57:16 Papanii Larry Ellison yes yes yes, yes. 00:33:57:16 - 00:33:58:02 Papanii Sorry. 00:33:58:03 - 00:34:09:19 Papanii But my point is there are clear examples where builders, regardless of whether they're technical or not, builders have led. Apple is one. Oracle is one, you know. 00:34:09:21 - 00:34:31:10 Adam Right. Last one last topic. This is from an article in in in the Hacker News. So basically ransomware groups in particular are optimizing for repeatable low effort playbooks rather than sophisticated new attacks. The majority of bad actors out there are not trying to use frontier models in live attacks. They're not even approaching them. They don't want frontier models. 00:34:31:10 - 00:34:50:04 Adam They don't want the bleeding edge of AI to help them hack into these systems. They're actually using repeatable, tried and true methods that still work because they're economical. What they're actually doing is they're instead of using AI in the live attacks, they're using AI to build attacks. Only when they get stopped on the old attacks that don't work. 00:34:50:08 - 00:35:10:12 Adam For example, Quick Fix, which is a fake capture prompt that tricks users into pasting and running malicious commands, still accounts for 47% of initial access methods tracked by Microsoft. So that's old news. Living off the land tactics using something that's already built in the system. No custom malware showed up in 84% of high severity incidents that Bitdefender analyzed. 00:35:10:12 - 00:35:43:00 Adam Across 700,000 cases, ransomware made up 48% of all breaches, which is up from 44%. So if we extrapolate that and look at it. So even as ransomware attacks rose, the medium ransomware payment fell to about 140 grand per, and still 70% of victims refused to pay it. All those builder kits that we were talking about, the quick fix builder kits, those things cost from 200 to $1500 a month on the dark web to get into those $140,000 ransoms. 00:35:43:00 - 00:35:44:01 Adam What are your thoughts on that? 00:35:44:02 - 00:35:44:11 Papanii Well, I. 00:35:44:11 - 00:36:06:11 Papanii Would say is this I think maybe the data's you know, I'm not going to argue with the data, but I'll say from a practical standpoint, yes, there are things and basic attacks that basically we all do right get we get fooled for. But what I would not agree with is that they are not using these models to find those loopholes. 00:36:06:12 - 00:36:32:12 Papanii A couple of months ago, there was this whole miasma attack. And the way that we're able to figure that out, they did use frontier models to figure out those loose ends. Now, the attack itself wasn't because they used it, but how they figured out how to to do that attack was based on that. If you think about what was it mythos right from anthropic that showed all the various vulnerabilities that were in open source software and libraries we all use. 00:36:32:12 - 00:37:02:08 Papanii So yes, the the, the attackers probably won't use some sophisticated script to do whatever they need to do, but they will use the the model to find out what are the vulnerabilities. And the final piece is you just have to do it once to a big institution. And yes, maybe the count wise may not be that Grand says yes because it's all these things, but you do it to one institution that has brand implications, customer data that can have a significant punch. 00:37:02:09 - 00:37:03:13 Papanii Let's put it that way. 00:37:03:17 - 00:37:23:07 Adam Maybe it comes down to like hitting like critical infrastructure or hitting infrastructure of the internet that isn't visible to most people. But like the SolarWinds hack, for example. Correct. Almost nobody knew what SolarWinds was or where it even existed. Most, you know, CTOs didn't have a good idea of what SolarWinds was doing for him until it got busted, until they got popped for it. 00:37:23:08 - 00:37:41:04 Papanii Look, we know that some of our engineers, we are lazy. We we know certain things. Like there was a probability this is going to happen. Ship it. Right. Stuff like that. People have done those things. And when you're fighting with your product counterparts based on timelines and stuff. Oh well, we'll come back to that and fix it. Let's get it out, you know, and then it goes on your backlog and you forget about it. 00:37:41:04 - 00:37:43:20 Papanii Bad practices that we all know we have to fight. 00:37:43:21 - 00:38:14:03 Martin I love that example of, you know, engineers are lazy. I have an example of that. It happens in all languages. But I think things like node and stuff that are very dependency heavy, I've we've gone in and done audits of code in, you know, in the past and gone in and found that they're importing like 100,000 line open source dependency to use a five line function, just like the laziness of that is almost beyond belief to me. 00:38:14:04 - 00:38:17:15 Martin It's like you're a noted us up and you wanted this. 00:38:17:15 - 00:38:35:15 Papanii But that was the whole point of dependency manager. That's why we use Maven, right? It's like, yeah, I want this function. Give it to me so I can move on. Right. And you know, it's funny. Was it three years ago we had the log for J issue. And look at what it did to all industry companies like love for J. 00:38:35:20 - 00:38:40:10 Papanii Like everybody was like what the hell. And it just screwed up everybody. 00:38:40:12 - 00:38:41:15 Martin Everything I. 00:38:41:17 - 00:38:44:11 Adam Still out there. 00:38:44:13 - 00:38:55:05 Martin I was doing a lunch and learn thing for security recently, and I mentioned the log for J and the people in the room. It was like like a. 00:38:55:05 - 00:38:55:19 Speaker 5 Visible. 00:38:55:20 - 00:38:57:22 Adam Physical reactions. Yeah. 00:38:58:00 - 00:39:05:00 Martin A genuine physical reaction like, oh God, please do not remind me of that thing. I never want to deal with it again. 00:39:05:02 - 00:39:11:00 Papanii I remember we had to do that stuff, man. The whole company, everything was on pause. 00:39:11:01 - 00:39:11:10 Martin Yeah. 00:39:11:11 - 00:39:19:16 Adam That's like seeing. Seeing an old girlfriend in the supermarket. You're like, oh, come on. 00:39:19:18 - 00:39:24:05 Martin Adam, I have to say. 00:39:24:07 - 00:39:36:05 Adam So great to have you. There is one thing that we do with every guest that we have on. We ask them the same question. What's the one thing teams are getting wrong about AI and software delivery right now? 00:39:36:07 - 00:40:02:22 Papanii They're getting wrong, I don't think I don't think people are getting things wrong. I think that they're going on a journey of discovery because this is a new way of approaching problem solving. Why do I say this? Intelligence has never been a commodity. Intelligence has been such a thing that it is the technical hardware that has constantly been changing. 00:40:02:22 - 00:40:24:09 Papanii If you think about, you know, when, you know, the combustion engine came, you know, like the horse chariots and were affected. Right? But this is the first time where intelligence is now completely different. This is the evolution of intelligence. And so some people are like, well, I used an agent in 2022 and a code of scraps. I'm not going to use it anymore. 00:40:24:10 - 00:40:45:08 Papanii And that's no longer the case. And so going through that transition of trying to pull or bring people along the journey, it takes time. So there's the evolution that has to happen. And I think there's a there's a tie in between what the business wants, what the individuals and the company wants and what reality is and what AI is capable of doing. 00:40:45:12 - 00:40:58:07 Papanii Implementation is the cheapest aspect right now is no longer the. So how do we make sure what processes we have are relevant and ensure that we have the right data so that the implementation is actually done right? 00:40:58:08 - 00:41:03:15 Adam Well, thank you so much. Great to talk to you, great insights and look forward to seeing you soon. 00:41:03:16 - 00:41:04:17 Speaker 6 Thank you for having me. 00:41:04:18 - 00:41:06:12 Martin Yeah. Thank you so much. 00:41:06:14 - 00:41:12:00 Adam All right. Great to have you on. Martin, what is your takeaway from what we talked about today? 00:41:12:00 - 00:41:40:15 Martin I love that he actually had a vision, if you like, or an understanding of a vision of what's going on in space and and cursor and exi and all those things. Maybe I'm a little bit more skeptical than him. Just a little. But I love that, you know, he had that openness division, which which I think is, you know, it's exciting to hear, even if I don't fully agree. 00:41:40:17 - 00:42:01:11 Adam So for me, the one thing that landed that I'm going to take away from this conversation is the value of university time in expanding people's knowledge beyond just getting a job or a career, but actually teaching them how to live life. I forget about that part, and I focus a lot of times on how do you get the job? 00:42:01:11 - 00:42:06:17 Adam And I forget to focus on how do you become a more full human that can make better choices? 00:42:06:18 - 00:42:09:05 Martin Yeah. Learning how to adult. 00:42:09:07 - 00:42:14:17 Adam I have not found that out yet. So if anybody has a guidebook. 00:42:14:19 - 00:42:28:22 Martin Well, okay, that ship took it bought to you by harness. If Adam's point of view got into your skin, it regularly gets undermined. Subscribe. Tell us why in the comments. And until next time, stop talking. Start shipping.