Inspired with Alexa von Tobel
00:26:57 · Dan Lorenc, Alexa von Tobel
[00:26:57] S0: I... You know, everyone has a lot of different opinions, so it's hard to come up with one that everyone disagrees with. But, yeah, I think I think the models are already good enough to do the work of most engineering teams, and I think not everyone has realized that that piece yet. Some have. So I don't know if this is something that everyone disagrees with me on, but I posted a while ago, like, cloud opus 4.6. I don't know if I would call it AGI, but it also kind of is at the same time. Like, I don't know that we can tell anymore. We don't need to wait for another wave of models out there to really put these things in their production and get work done safely and, ship correctly. A year ago, that wasn't true. And I think a lot of people, if they're not actually trying this and at the forefront every day, every week, every month, you can kind of get, fatigued. And, you know, six months ago, it wasn't that good. You might have some brilliant moments and then a bunch of terrible ones. But in the last three months, things have changed dramatically, and I don't think we actually need them to get any better to do everything we we need to be doing in the software industry. And they are gonna get better, and that's the crazy part. [00:27:57] S1: I'm gonna just play that back. Effectively, what you're saying is, like, we are near AGI already, and, actually, you can't even necessarily fathom things that we are currently not capable of doing. [00:28:07] S0: Yeah. For, in the software space, the models are... Yeah. They're there.
00:28:07 · Dan Lorenc, Alexa von Tobel
[00:28:07] S0: Yeah. For, in the software space, the models are... Yeah. They're there. [00:28:11] S1: That's incredible. That's a... That's like a a real moment just to to to soak that in. Yeah. I mean, that... Dan, that's really well said. [00:28:23] S0: Win five point o or whatever comes out, I'm sure I'll use it, but I don't know that I I need it. [00:28:28] S1: Let's shift a little bit to security because there is actually I sit here as someone running a venture fund, and just every day worrying about the exponential number of ways that security is getting worse. And actually, so that was one of the questions I wanted to ask you, which is, what do most companies fundamentally misunderstand about software security today? So if you were talking to all your customers today, and we've had many CSOs say to us, this agentic software is happening, and it's happening in our in our orbit right now, and we actually don't have our arms around it. Just give a sense of, like, what what worries you that you would wanna say to everybody.
00:28:28 · Alexa von Tobel, Dan Lorenc
[00:28:28] S1: Let's shift a little bit to security because there is actually I sit here as someone running a venture fund, and just every day worrying about the exponential number of ways that security is getting worse. And actually, so that was one of the questions I wanted to ask you, which is, what do most companies fundamentally misunderstand about software security today? So if you were talking to all your customers today, and we've had many CSOs say to us, this agentic software is happening, and it's happening in our in our orbit right now, and we actually don't have our arms around it. Just give a sense of, like, what what worries you that you would wanna say to everybody. [00:29:13] S0: Yeah. There's probably a couple different angles to take this one from, but I think you you touched on it right there at the end. This is the first, you know, massive technical change movement like this where security teams are not allowed to say no. Everyone understands how important this is, and they need to find a path to yes. And if they're too slow, it's happening anyway, without them being in the loop. They're not allowed to say no to this, even if it is dramatically adding risk, dramatically, you know, putting a company at risk of violating trusts, getting hacked, all of these things. And they have to figure it out quickly. And on this exponential slope that we're on, if a security team is, say, three months behind what engineering is doing, that three months is getting, you know, much, much, much bigger every month that passes. [00:29:57] S1: It's, like, three years.
00:29:57 · Alexa von Tobel, Dan Lorenc
[00:29:57] S1: It's, like, three years. [00:29:58] S0: Yeah. Three months today is, like, three, yeah, three years a year from now. And it's it's it's a very scary challenge for them. The attack surface is increasing. The more code gets written. People that don't know how to write code are writing code inside of companies. And they have to figure this out quickly. Otherwise, bad things are gonna happen. But what I think here is, like, I don't actually worry about it inside of software as much. A lot of people are very concerned about, you know, vibe coding and things like this inside of inside of software teams. The software industry is already pretty good at all of this. Right? You can take an intern off the street at most companies and have them shipping production code. The models are already better than interns. Right? If if we have guardrails, we have code review, we have tests, we have all this kind of process in place. The part where I think we're always want other parts to start writing their own... To do their own things, and they don't have all of that culture and all of that process baked in. If you have a finance team that builds some new app to, you know, do payroll next week, you don't have unit tests or code review for payroll, and that's the area where I think there's a lot of risk in the organizations aren't really well equipped to deal with it today.
00:31:02 · Alexa von Tobel, Dan Lorenc
[00:31:02] S1: Dan, important... What you basically just said is CSOs can't afford to say no, because each month is actually, you know, ex... It's three months. It's A year is three years of their organization being behind on what's possible. So it's basically a race for relevance inside these companies. Obviously, there's different types of companies, right? And I sold LearnVest to Northwestern Mutual, a big regulated company. So in health care, in financial services, places where they're regulated businesses for a reason because the gaps of what could go wrong actually are severe and have real consequences. Do you feel like there... If if you're, like, a big financial service firm, is the safer thing to wait and see and actually just... Because breaches can be so dangerous. How do you think a little bit about the moral hazard decision of CSOs right now in different types of companies? [00:32:07] S0: Yeah. I... I'm a little surprised that you don't see that a real change there in adoption between, you know, the most heavily regulated companies out there and startups. [00:32:17] S1: Wow. [00:32:18] S0: Yeah. It's... They all get it. You know, the CIOs of the biggest banks there are, you know, telling their teams to get on board here, because they know if they're not one of their competitors will. You know, there's a... One of the largest, oldest banks in The US has agents inside of their org chart as employees. [00:32:35] S1: Yeah.
00:32:35 · Alexa von Tobel, Dan Lorenc
[00:32:35] S1: Yeah. [00:32:35] S0: They're they're really leaning in here. Most shifts, yeah, you see the laggards, you know, the basic kind of crossing the chasm curve. That's all compressed. Even the the people that are going slower, like, two months behind, not six years behind like they would normally be with a technology rollout. And, you know, they're depending on everyone else to figure out to do it safely, but they have to be able to ingest that very quickly to that knowledge and those practices. You know, we had a customer conference a month or so ago, you know, I asked everyone, like, raise your hand if you're using AI. Everyone in the room, their hand goes up. Raise your hand if your security team is happy about it. You know? About a quarter of the hands stayed up. Raise your hand if your security teams are using AI. Even fewer hands went up. Like, that's the scary part of all of this. If the security teams can't get on board as quickly, that gap, those that one month turns into ten years.