Musk’s Macrohard Under Scrutiny | AI Closes the Coding Gap | Identity Crisis in the Cloud

Hello everybody, and welcome to the Techstrong gang for Friday the 13th. What a great day it's gonna be. I mean, we've been afraid of the number 13 for as long as we can remember, and probably goes back to Roman and Norse mythology, and all kinds of fun stuff.

But here we are on Friday the 13th talking about Elon Musk. How much better does it get than that? So we are now at least having a conversation about this thing called Macrohard, and we have our guests today.

John Schwartz is of course joining us from the West Coast. Guy Currier is here. Tracy Reagan, good to see Tracy as always, and finally, Jack Pollard's here.

And this conversation's gonna start off with, well, just what the heck is this thing that Elon Musk is talking about? It's-- he's, he's named it Macrohard, which is kind of a, a dig at Microsoft, I guess, and it's supposed to be a software engineering tool that will enable you to build just about anything that a packaged application gets. At least that's what I got out of it.

But John, you wrote the story on this. What's going on? Oh, G*d, I sh- I drew the short straw on this one I think.

Um, so far for me, far be it for me to try to get inside the head of, of Elon Musk, but essentially given the series of moves that have been happening, and I, I know we've all kind of followed it, but it's hard to, it's hard to kind of understand what he's doing. But it seems like he's integrating and moving around different parts and concepts of his companies amid some lawsuits from, Tesla shareholders. So to just kind of backtrack a little bit, Musk previously argued that the two companies, that, that, that are involved, Tesla and xAI, were separate.

But this new project, which he's nicknamed Macrohard, it's also called Digital Optimus, the idea is to position xAI as the brain for Tesla's hardware. At least that's what I think. Um, there's a, uh...

It's interesting the context of what's going on because I think about late 2024, he insisted that there was no need to license anything from xAI. He actually argued that Tesla's, AI needs were fundamentally different than xAI's large language models. So, the context that I wanna bring up is that in August of 2025, xAI had moved to trademark the Macrohard name.

That's-- was proceed, that preceded, a two billion dollar investment that Tesla made in xAI's Series E funding round, and in February, SpaceX acquired xAI in an all-stock deal. So i-i-it-- there's a series of these moves. I mean, they're, they're ma-me-meshing the two organizations in terms of tasks together.

What's intriguing to me is that there's a Tesla shareholders lawsuit, fiduciary duty lawsuit in Delaware Court, and it alleges that Musk diverted talent, proprietary focus, and critical, GPU shipments from NVIDIA, from Tesla to xAI, which is a private entity. So I think in a sense he's kind of playing around with this Frankenstein monster, and he's-- I guess what he's trying to do ultimately through all these moves is emulate the functions of an entire software companies by using AI to autonomously navigate computers. I think that's what he's trying to do.

I'm gonna open it up to you because I know you all have your theories about him and, his madness, so to speak. Oh, Guy, what did you read into this? Um, I don't know how much I read into it, more like confirmed my priors, which is that, the guy's flailing to get back on top over Larry.

I mean, I don't know. Um, I-- it, it was classic Musk to me, which is a whole lot of gobbledygook, whiz bang sounding stuff so that, Tesla looks better 'cause that's where, where the majority of his wealth tie, is tied up, and that's normally the focus. Uh, xAI looks better.

Um, Grok itself has come under a lot of questioning from those who follow this sort of stuff about how it was trained and, and, you know, what-- how, how, how it operates, how, you know, reliable, let's say, its, its outputs are. There's lots of people who've had good experiences with Grok, so, you know, I'm, I'm, I'm not putting a, put-putting a stake in that fight. Um, but yeah, the-- I mean, just reading John's story, John, I'm sure you clarified it as much as you could, and it still is really hard to track.

Yeah, yeah. You're, you're right. I mean, it's- Yeah ...

Guy, you've used a perfect word. Sorry to interrupt. Used a perfect word to describe what he's doing, this gobbledygook, because there are a series of moves that I think are more about flailing and trying to prop up or appease the Tesla shareholders at the same time, to appease them and also to prop up xAI because it seems as if he's just-- he's looking for a solution through all these weird pie in the sky concepts that he's throwing at us, like he's throwing spaghetti against a wall.

I don't think he's onto a new idea. I think what he's trying to do is get in front of a parade, and the parade kinda looks something like this. You don't need necessarily to buy packaged software from a Microsoft or an SAP because the AI is gonna build the applications you need to run your company for you, and you can just go ahead and kind of, you know, move on without them and save all those licensing fees and maybe that helps Tesla.

But also under the heading of just because I might be able to do something doesn't mean I should, Tracy. So at the end of the day, I mean, do organizations all wanna go in and rebuild every piece of packaged application software to save on a licensing fee and then have to support that? Is that realistic or how do you read this thing?

I'm sorry. When I read it... " I honestly thought he's trolling Microsoft and his, you know, it, it...

Oh, come on, Microhard. Macro- no, it's Macrohard. Macrohard.

Macro- Macrohard ... Macrohard, I don't want to take it into the trash, but it started there when he named it that. And I feel like he's just having a little hissy fit because he's not in the game where he should be.

Ho- honestly, that's how I feel... I... When I, when I read the article, I was like, this is his way of throwing a f- a little hissy fit.

Um, I don't think that developers are gonna run to him, to, to, for what he's offering. Uh, but, but I- ... I, I swear I thought it was he's just trolling.

He's trolling- Yes ... and this will go nowhere. You know, I wouldn't be surprised if he said he's also gonna build the data center on Mars, right?

No. Well, he said that- Yeah ... he had data centers in space already.

But I think you guys- Yeah, the moon I think ... are missing something else, which is this is a repeat of his last play. He's done this before.

You're just not thinking about it the same way, which is he created SpaceX and said, "I wanted to go to the moon and Mars," and then he said, "The way to do this is to get volume, and- Mm-hmm ... " So I launch satellites- Mm-hmm ... in space, and the s- the, the volume Starlink satellites was what drove SpaceX volume and the ability to, to do launch cheap.

And now if he needs, to develop hardware to drive his AI, and he's getting Tesla doing custom chips, which I didn't realize until I read John's article. If Tesla's building its own AI inferencing and training chips, then he needs customers for that, and the perfect customer for that is xAI, right? So this is the same play, create something and then create a demand for it to drive volume.

The big winner here, of course, is the chip manufacturers, right? Hopefully Intel will get a piece of this, but this is really going to, you know, TSMC is gonna be the big fab who's gonna get to build more chips and, you know, I think that s- that's an interesting side look at what's going on. Can I, can I bring up- I'm sorry, what was that?

Really it's, it's, it's depending on how far this goes, because I- I think 90% of the strategy and statements and everything that... I mean, remember Musk was gonna buy OpenAI or sue them. Remember all that whole s- I mean, that whole story.

Yes. That's what I'm saying. He's having a hissy fit.

This is all de- this is all a sense of deja vu. We've been through this before. I'm glad you mentioned that, Guy, and also Jack, that if...

And since you mentioned, but you know one thing I was gonna mention that, kind of channeling Alan, that the fallback that Musk has, as crazy as his ideas are or as, as calculated as they may be or immature as they may be, he has a willing customer or participant in the government. Can I, can I- That's a customer. Can I, can I just add one thing, which is i- if you're watching this, you're either like sort of from an investment or where's the market going standpoint, I think we've kind of covered that.

But you also might be like a, you know, a, a developer, an implementer, an IT user, whatever it is, and there really is a serious element to this to my mind, which is the, the death of SaaS, the death of software, if you like, the death of software development. In other words, agents just... That, that's the strategic play here, supposedly anyway, is, is replacing your enterprise software suite with agents that just build stuff and then destroy it as you go.

And, that's a longer conversation to have, but if you're looking at the future, that's, that's a potential er- way for this all to go, at least temporarily. Mm-hmm. I think that's a possible future.

I just... The part that I don't buy, it's like, so am I really gonna create a, a general ledger, every time I wanna go create some sort of task or- Yeah, I, I feel the same way, Mike. Yeah.

Right? Or am I gonna- Yeah ... just invoke that as an API that the agents go invoke, but it's a back-end headless service, but it's still provided by SAP or whoever?

Well, and, and the oth- the other part of that equation is no matter how much Elon or the, you know, the Linux community or everybody else wants to pound their chests, the world still runs on Microsoft and Android and iOS, and it's gonna be very hard as Microsoft learns to shift people off of Android- Mm-hmm ... and iOS, and it's gonna be very, very hard to shift people off of Windows on the desktop and Windows servers and the Windows environment. You know, the Microsoft environment is so entrenched.

The switching cost is so high, peop- that's gonna be really hard to say, you should throw all that out for this unaudited AI agent that's g- gonna go and do things willy-nilly for you. You know, that, I think that, that whole thing is just... There's a lot of hype in there that's, you know...

I mean, we could spend hours on that one alone, Guy. And, and there's a- Yeah, that's what I'm saying. Yeah ...

there's a cultural aspect of it, of, of this too because I think what... You know, I didn't dig into it too far, but it seems like what he's saying is AI systems should be managing development teams, and that's not gonna happen anytime soon. No.

Developers are fiercely independent. They still like to script for God's sake. So, yeah.

It may be Macrohard to get this done. You know, I'm glad, I'm glad that he made this announcement when he did. If he had made it closer to April 1st, I would've d- I would've clearly not done the story or I would've had second thoughts about writing it.

Right. So the other side of this coin though, and let's, you know, be candid about it, how much of this is just like some effort to maybe tank a few stocks and short some things with an announcement, and that's part of the hissy fit? John, what do you think?

Yeah, that's pretty much what he does. I mean, he's going back to the OpenAI acquisition and then the consequential lawsuits, is he's does h- often makes announcements to damage competitors Or to tweak 'em or to g- engage in some sort of a, a sophomoric social media debate. Um, I think it's just attention mongering, and I actually think...

I, I can't remember who said it earlier, but he's kind of on, he's kind of on the outside looking in a bit, and I think there's a lot of anxiety and nervous, I don't know, jealousy on his part. All right. I think we leave it there.

I think you have it exactly right. It's attention mongering, and we'll see what happens from there. But we're gonna hear more of this kind of noise from the system, so pay attention to what these folks are saying, 'cause a lot of them have ulterior motives, shall we say.

All right. Let's move and shift topics. We've been talking about AI and App Dev on this show regularly, almost at least every third show, and there's a new report out talking about how, it looks like at least that the laggards are benefiting more from AI in terms of improvements to software engineering practices than some of the more advanced companies when it comes to software engineering.

There's a story about this on DevOps, and based on a study that an outfit called, PlanDeck did. And on the one hand, that may not be all that surprising, but a lot of folks, Tracy, have been saying that it's the most experienced developers who know how to build software that benefit the most from AI, and yet maybe the laggards are benefiting more because, well, they're just coming from so far behind that, you know, they're using AI just to catch up and close the gap. What do you think?

So yeah, this is a interesting kind of discussion that we should start having about how AI is, is changing, the way we develop software and who really does benefit from it, and I think both sides do. You know, if you're already... If you have a high-performing, team already, a development team that's really able to get software updates out the door, you know, i- in, in agile time, you know, doing several releases a week, AI can only help you.

It's true. Um, i- if we are looking at new developers coming onto the market that are not as high-performing, AI can only help them. So i- if we look at, y- you know, if we take a look at each individual o- on a team, l- let's say take me compared to some of the developers that we have, or Steve Taylor, I benefit a lot because I do, I use a lot more of AI to do my coding, and then I ask them to go check it.

And they're high-performing, so they can actually look at it quicker and determine if this is gonna work, and we can start, and, you know, we can build off of that. So AI has benefits for both sides of the, of the coin, and neither one of them is bad, right? Because we don't have that...

I mean, there, there are plenty of great programmers out there, but we need a lot more. So I think what AI is doing is kind of pulling people up, and allowing the people who are doing really well to do just everything faster. So, you know, I, I used to not be a fan of AI in the beginning because it was, it was breaking stuff so often.

But over the last, I would say the last year, it has changed a lot. Um, Anthropic has really pushed the envelope when it comes to really good tools for w- writing software. Claude has been great.

We depend on it more and more every day. And I can take, I can even take my, you know, a markdown language or HTML, the stuff I'm doing for the Linux Foundation, and run it through something like Claude and fix problems that I've, would've taken hours to fix because I'm not a high-performing software developer. So, it's both.

It's both, and it's not bad on either side, right? It... And if these high-performing teams, are bringing in people who are new, they're just gonna become high-performing even faster.

Mm-hmm. They're gonna learn faster. So I don't- I'm not sure...

I'm not entirely sure that the gains for the high-performing teams are that much greater. And I, and I ask this question because... So every DevOps team I ever met has some sort of bottleneck or something that prevents them from going faster, and it's just built into their workflow, and it's something they learn to deal with and live with.

And from what I've seen so far, AI doesn't really resolve those issues for them. I mean, it's, they're not really coding issues. They're process issues and things that prevent them from, shipping more software faster.

And I think if we're gonna take on those bottlenecks, we need to maybe get to the next wave of AI. But Tracy, feel free to tell me if I'm crazy. You're crazy a little.

Uh, a little. I mean, there, we, our, our DevOps pipelines are definitely long in the tooth, and they need, AI. In fact, I, was given the honor of being the chair of the CDCon 2026, and I had us do a full h- a full half day of nothing but, discussion around bringing AI into the pipeline.

Uh, CD events, things like CD events, and ways to improve and get, i- improve overall workflow execution. Uh, look at platform engineering. I know MCP servers are always a bad topic here.

Jack will always g- get me for that, but using MCP servers. But the, the real, uh... So for example, our team was looking at how, how many, and I think I brought this up last week, how many, MCP servers does it take to replace 1,000 GitHub or GitHub actions?

Less than 10. So these high-performing teams are already doing that. They're starting to do that.

MCP servers now are everywhere, and, and tools like Jenkins are starting to be replaced by MCP. It's just that, that's just the way, the way it's going, and the high-performing teams are getting there first. But you're right, Mike.

Not all of the t- teams are working in a, such an agile environment that they can do that. Enterprises are less likely to jump into that. But even those processes are getting easier to fix with AI.

Okay. So that's the point. I wanna, I wanna jump in here with a little bit.

My inner skeptic was screaming when I read this article, right? And, and, you know, I, I'm, I'm no longer immersed in the developer world. I was a developer, but that was, you know, very many...

That was a couple decades ago. Uh, but the metrics that they're using to define high-performing, low-performing teams scream to me as, as a little bit contrived, and I feel a little bit that this is sort of, conclusions with a study to back it, preconceived conclusions with a study to back it. In that, you know, Tracy's been around and I've been around and, you know, Guy, I'm sure you've had the same experience a little bit that we sort of, you know...

To, to paraphrase the Supreme Court, I don't have a definition for it, but I know what it... I know a high-performing team when I see it. And, you know, the...

For instance, one of their metrics was the, the number of hours it takes to merge pull requests. Well, you know, a lot of that is context dependent, and, and AI can help that, accelerate that if you've got a very complex environment. And if you have a very simple environment, then you don't need AI to help you, and your pull, pull, your, your pull requests merge very simply because it's just a simple thing you're working on.

That doesn't necessarily mean you're high-performing, it just means you're working in a very simple environment. Uh, you know, and so it's... I think AI lifts the boat for everybody, and it definitely is a tool that can be used to accelerate your development and to improve your development, and it can also hurt you just like any other tool if you don't know how you're using it and you don't use it correctly.

Mm-hmm. Jack, I thought you were gonna jump out of your chair when Tracy said that they're using MCP to replace all their, uh- Oh, oh, no, no, no, no. I, I want Tracy to use MCP.

I just- ... I'm, I'm raising the red flag, which I'll raise again on our next topic, of know the security issues when you do so, right? Know...

Understand the tool and know what it is, right? If you're internally doing MCP stuff, great, go for it. If you expose that to the outside world, you better know, understand how to secure your environment really well before you do that, right?

And it's because it's not necessarily designed with security in mind. It is designed to do exactly what Tracy wants it to do, which is to enable you to accelerate and to automate what is a very complex process of software development where you might have 1,000 different actions you take on the course of designing and developing a complex software product. And if you can take that down from 1,000 steps that people have to walk through manually or run this script, then run this script, and automate that, and automate that, and you can feed that to the AI, AI, and it becomes a PhD project.

Push here, dummy. Then we have a, a very interesting thing to look at. Mm-hmm.

Yeah, it's interesting- Well, this is all kind of wild. Go ahead. 'Cause...

So my first thought, remember the, the, the, the theme of this was that the, the underperforming teams, improved the most. So my first thought was, it does pay to be lazy after all. Just wait long enough and- ...

someone will come up with something to, you know. 5, that, that the productivity benefits, s- you know, w- which... Jack, your point is really well taken.

Like, s- all these metrics seem to come out, come, come down to productivity. I, I've always thought that was secondary. Um, and I th- and I continue to think so.

Agentic makes it a little bit different, but the, the, the two obvious benefits to me were, number one, it helps you get off the blocks as a, as a coder or creator, whatever you are. Like, you know, you don't really, suffer from, from, you know, block as much if you can just ask your friendly AI companion to produce garbage that you can then react to. And then also quality, because...

And I think, Tracy, that's what you're, that's what you're touching on, is that, if you know what you're doing, if you're already high-performing, now you have, an, an assistant which may or may... They're getting better, but may or may not be good. But, it, it gives you more, like, let's say, review cycles.

It gives you a template against which or, or from which to, you know, to, start a little bit ahead. So, so I feel like, like you say, Jack, the metrics, that's what's really wild to me is the metrics here continue to be c- consistently off. Yep.

They seem- Really ... to be designed to sell AI or to sell AI products and not really designed to, to get at the root of what we really want from software development, which is consistency, repeatability, quality, that sort of thing. And I think AI is a terrific help with those.

Agentic AI, because at its base it's nondeterministic, I think engenders some inefficiencies at the same time as it, it, it helps more directly with productivity. But now we're seeing more and more deterministic models or deterministic ways with, with agentic. So, so I- That may change my thinking.

I, I agree with you for the most part, but I think product... I- in this case, productivity is a very important thing when we look at these types of tools. It's just how we measure that productivity is a little bit wonky sometimes.

I look at this, like, when I started in software de- and hardware development, I was writing in Assembly. Now, I wasn't writing in binary, right, but I was writing Assembly code, and the amount you could do in Assembly code was very minimal on a day-to-day basis. When you moved from Assembly to a higher level language, like C, your productivity jumped by, you know, more than one order of magnitude.

It was a huge ability to do more and get more done on a day-to-day basis, or get your project done. You could get the project done quicker with less people when you were working in a higher level language. And then we moved from things like C to Pearl and Python, and all the other fourth generation, fifth generation languages, and we have SQL for database stuff, so you're not doing database the, the way we used to do it in the old days.

But we still don't really have a good metric for that. But like I said earlier, we know that improvement when we see it. We can, you know, or we can put our nose to it and say, "That's a good...

" I would put- I agree. Yeah ... I would, I would also put forward this, there's this other metric that matters maybe more than all the other metrics, and it was illustrated in a article that appeared in The Times this week where they went out and interviewed 100 developers, or at least the fellow said he talked to 100 developers, and they all were experiencing more joy than ever, right, building software.

And basically, at the end of the day, the metric that maybe is most important of all is that joy. You know, Tracy, are developers just enjoying their job more because they're basically now all architects rather than coders? They're enjoying their job more, that's for c- certain.

Um, you know, when you get a... " It can be really, really frustrating. Now everybody has a teammate that can look at it for them.

The, the other thing that I'm seeing in a change in, in what I would call... A- and metrics are needed. I don't...

You know, in this article I really was struggling to understand what a high or a low-performing, you know, team was. Uh, if we look at door metrics, I guess we could determine that based on how many releases they do a week, that's why I use that term. But the difference I'm starting to see in how the developers who've been around and coding for a while, the how they're interacting with AI is pr- very different from junior people.

And what they do differently is they interact with AI like AI is a, a human. The way they write their prompts, the way that they, string together prompts is very different than somebody who's just started coding. It's almost as if the hi- the, the developers who, who are...

People who are really strong developers can explain to AI what they want better than people who are not strong developers can explain to AI what they want. So really we're s- I think that we're seeing a, a, a change in how we think about code and how we generate code. And I, I've just seen people get better at their prompts and having a discussion like they're talking to a junior programmer and telling them a task that they want them to go off and do, and they make them smaller tasks.

So they're getting more code like snippets back as opposed to a junior programmer who tries to get it all done at once. It's a very interesting way, it, it, you know, to, to watch people interact with AI in c- in, in generating code or doing anything in the software development life cycle. One talks to the AI as if it's a junior programmer, the other one talks to AI like it's a senior programmer.

You could- And that's a difference. You know, there was, there was a, there was a story in this block about the Frankenstein monster story, I think Alan wrote it. And I just wanted to point something out that kind of applies to the corporate world.

So within that story, Alan writes that AI's writing vastly more code, vastly faster at a cost structure that no human workforce can match. And you're starting to see CEOs acknowledge that during layoffs. " And that was one of his rationales for, for the, these, these job cuts.

And it kind of echoes something that Benioff said about around agents and sales teams. So I just kind of s- wanna see this through the prism of the corporate corporations, they're thinking in these terms too. Yeah.

Well, when you're building that, that, you know, that monster, you better hope that the person you ask to go get its brain doesn't pick Abby Normal. Oh, yeah, no, you don't... Yeah.

It's all short-term- And that's what junior program- ... strategy. And the junior program- Sure.

Yeah ... programmers will a- you know, they'll go get me the brain and they ask for Abby Normal, right? Yeah.

I, I gotta- Reference ... I gotta move this conversation along, but I will share a conversation I had with somebody a long time ago who's in IT, because IT people are different from the rest of us. " So maybe that's why the developers are really digging the AI and talking to the machines.

But w- let me go to this next topic here 'cause it kind of speaks to the next issue that comes up with all of this, which is security. And Jack has an article on Security Boulevard talking about a report from Google highlighting the fact that, well, we now have all these identities, and we couldn't manage the identities we currently have, some of which are already non-human, most of which are human. But now we're gonna have all these AI agents, and I'm frankly not even sure if they're non-human, or human, or some hybrid, or some new thing altogether.

But Jack, how are we gonna manage all this and secure this? Because it seems like it's a recipe for havoc. Well, it, it is a recipe for havoc, and this was, my article discussed, both the Google, threat data that they released which said basically 83% of attacks were, originated through an identity compromise of one form or another.

And that backs up what we've had for 20, almost 20 years of Verizon data that says the same thing, that identity is the first point of access for, most attacks that, that, that people are doing. And what we're seeing now is that, organizations typically have anywhere from, 40 to one to 100 to one, for every human identity they have 40 to 100 non-human identities, whether it's a machine, an AI agent to service an API endpoint. It's just this huge explosion on identities.

And we-Did a kick the can down the road with human identities. We did a second kick the can down the road with API credentials, and now we're doing it again with, agent AI, and a- agentic identities, where we're not putting in the proper controls at day zero to secure and authenticate and authorize these things. And the proof of the pudding today came out with an attack that just was announced.

Uh, we heard about it late last week, but the details were announced this morning that, Stryker, which is a medical device company, they were attacked, and, somebody got access to an administrative credential for their Microsoft Intune, which is a device management platform. Now, being very good corporate citizens, what they said is they wanted all of their endpoints, all their laptops, all their mobile phones, all of the medical device units that they have to be managed by Intune. So should an attacker get access to that particular device, they could remotely wipe it and protect the ...

prevent the attacker from getting access to it. It's now unclear how the attacker got access to the admin credentials, but once they have admin credentials, they could do the same thing. And so instead of just wiping one device, they have destroyed, wiped every single device in a twenty-five billion dollar company with fifty-six thousand employees, and they are right now in disaster recovery mode for something that they never even gamed out what would happen if everything disappeared at once.

And I feel very badly for them, but I also wanna ... And I've been talking about this for years and years and years, that if we don't do basic cybersecurity hygiene, and we don't do authentication and authorization for every single identity in your environment, then y- you know, these things are gonna happen. And I had hoped that the, casino attacks in Las Vegas two years ago would've been the, the canary in the coal mine that people woke up and said, "Oh, we've gotta fix this," and nobody has done that.

Now we're gonna see whether Stryker survives as a company and can recover from this. Maybe that will get people to understand how critical this issue is. Every time somebody tells me that, you know, this incident is gonna wake everybody up, and yet right after that- It doesn't ...

everybody rolls over and hits the snooze button. Another one. You know, Guy, are we ...

Is, is this the moment for identity or what? Uh, yeah, definitely. I think that ...

I think Jack's perspective is perfect. Jack, please don't put a red star on the day. It will happen again maybe.

I think his perspective is perfect. I think that thinking of, um ... I, I guess, y- you know, correct me if I'm wrong, Jack, but, you know, thinking about hardening your systems and then allowing access as opposed to allowing access and then hardening from there.

That is something, I mean, every, every security team, SOC and CISO and so forth know, but in, in practice is more often violated because of how, h- how much demand there is on the business side to keep things rolling and moving. Right? Um, I think that, that ...

What, what occurs to me, Mike, is that, you know, bad actors can put together bad agents as rapidly, if not more rapidly, than everybody else can. Um, and, you know, the, the, the ... That's both good and bad.

I think that their failure rate, might dissuade a fair number of them. Um, they're not automatically experts in how to do this sort of thing, and in fact, I don't think anybody's an expert in how to ... Well, not anybody.

There's very few folks who really know how to, how to put these agents together, run them through a life cycle and so forth. I ... Like, the whole OpenClaw experience is an example of that.

Talk about blowback. Um, in this case, blowback on bad actors. On the other hand, as some people say, they may be clowns or clownish, but if they now have a flamethrower, they can do a fair amount of damage.

So I, I think that, that, that the, the question is the real question, which is, is this finally gonna wake people up to the importance of cybersecurity finally? No. It might.

I don't think so. Yeah, I, I ... Yeah.

I'm usually the one saying that, Tracy. I don't. I, I wanna say that.

I wanna say that, but ... Go ahead. It's just a non-differentiating, you know, feature, and it, it ...

Because of that, we, as, you know, companies that are providing digital services don't focus on it. They focus on what people are asking for. So until the, the, you know, every mom and pop starts to ma- or every person who ...

every buyer at a enterprise demands some cy- some sort of cyber report or understanding of what the company's doing around cybersecurity, it's n- it continues to be non-differentiating. So, y- you know, coming from a person who's developing software, we worry about cyber 'cause we are a cyber tool. But at the same time, we also are chasing, you know, we're, we're, we're chasing the next feature, trying to meet what the develop- what the community's asking for, and that's the differentiating feature.

So for many of us, security is way down on the bottom. Uh, it's something that we know we need to do- Yeah ... but maybe we don't have the resources to handle it.

And there are so much ... It's such a giant elephant to deal with. There are so many sides of it.

Mm-hmm. And the, you know, the s- the Stryker one kind of bothered me because it made me reali- It ... Well, I've always known this, but I don- I don't believe that the Iranians just hacked into Stryker recently.

I think that they have identified many, I would consider ... You know, and Stryker I would consider a critical infrastructure considering it's in every ambulance, right? Yep.

Um, th- I ... How many of those are, are, are cyber attacks that they have planned for the future? And how many of them are coming from, actually coming from Russia and are being given to Iran?

Mm. I, I ... It, it just tells us that there is a, a new battlefield, right?

It ... There's a new battlefield in critical infrastructure. And didn't we ourselves bring down, the power grib- grid in Venezuela?

Yes. Right? Mm-hmm.

So we, we used cyber as a weapon as well. So weaponizing software through the cyber channels, right, is a new thing, a new and important thing to start discussing. And I don't think that the current administration is that serious about cyber, to be honest.

Yeah. I- Um, I mean, their recent re- their recent report was pretty f- you know, was pretty flimsy. It gave some high-level important, suggestions on what we should be thinking about doing that everybody's already been talking about.

Um, and we have new technology like agentic AI that is starting to surface, but there are probably many, many vulnerabilities sitting out there waiting to be exploited. Uh, and that was the creepy part about this, because I just don't believe that they hacked it that f- recently. But- But- But here's, but here's the thing, St- Tracy, just one quick thing, is that, you know, we've been talking about strong authentication for, for decades.

We've been talking about MFA for decades. We've been talking about zero trust. Zero trust is a concept that's almost 15 years old now, right?

Mm-hmm. And zero trust is mandated by the federal government. And y- you can do all the zero trust you want, but if you give your password to your admin account up on the internet somewhere and somebody gets access to it, zero trust doesn't help you.

There's a lot of things that we ... There's a lot of technology that we have if we practice good cybersecurity hygiene. There's a lot of different ways to do authentication.

There's a lot of different ways to do this. The, the military has this concept of dual controls, right? You can't fire a nuclear weapon without two people agreeing, right?

If you have the equivalent to bar a minus RF star, right, you can destroy your entire environment with one push of the button. Shouldn't there be dual controls on this, right? And if- I can tell you recently- You know ...

I've ... We've, we've started working with the US government, and it's surprising how far behind they are, even in DevOps pipelines. Yes.

Much less adding all of that that you just described to a deployment process or to any of the software itself. Mm-hmm. Dave?

I'm sorry to say that this issue will have its moment in the sun for about the next six months, and then what will happen is what happens with every crime-ridden story that we ever do, especially in the media circles, because it's ... John, it's just like, you know, if you work in a big city and suddenly there's a lot of robberies, well, now you start not covering anything that didn't involve maybe more than 10 grand. And so the same thing's gonna, like, play out here.

We're only gonna cover the Strikers and maybe Venezuelan grids, but all those common- Gonna have to ... Yeah. The incidents are gonna have to be bigger and bigger ...

all those common, everyday things we're gonna ignore. I mean, isn't it ... Do you think it'll play out any different?

No. No. We're gonna wa- we want something bigger and splashier because we get bored, you know, with the re- repetition of something.

And maybe that's almost, intentional on the part of the, the bad actors, is keep it below, keep it a certain threshold and, and see what you can do. Uh, i- yeah, it's just gonna continue. I mean, there's ...

" And it's like we're still in a, in a slumber. Right. It's the same thing, Jack, that we see with- Well, what's, what's the heck, am I the optimist on this, call?

We ... I mean, on this pod? Because, I think that, that agentic specifically- Mm ...

allows the volume of attempts to once again go up at the same time as their quality. Not that they would be undetectable or what have you. I ...

That's, that's what I'm sort of looking ahead to. " I don't think Stryker's gonna make any difference. Jack, huge company.

Tracy, like you say, all over the place. Iran... It's got all the elements that y- we would think would make people to sit up and to take notice, but it's not going to.

That I agree with. But these things could start happening once a week or something. I mean, they, they, they're mo- They, they ...

It could reach a point of actually getting some salience in the general public, and I think that's where it become, it, it gets, finally gets proper attention from enterprises, governments, and so forth. M- maybe, but, you know, at the same time- But, yeah. " So, Jack, aren't we gonna look at this and say at some point people gotta take responsibility for their cybersecurity, and if they get mugged, they get mugged, and America's gonna shrug.

Uh, I ... Right now my heart goes out to the CSO and the teams at Stryker who are ... You know, they probably haven't slept in four days, and they're not gonna sleep for another two weeks, if they can do anything about this.

At some point somewhere there's gonna be a post-mortem that finds the, the, the smoking gun of how the attackers got in. And yeah, I think, a lot of people are gonna look at it and say, "You had it coming because your cybersecurity hygiene, as good as it was, was really poor on some of the critical aspects that we've just ... " And, you know, I, I wish, I hope you, that you guys are all wrong, that, that we will wake up on this one.

But that's a cross my fingers wish that I know isn't coming true. It's not Christmas right now. Yeah, it'll- well, the problem is with, you know, the analogy of getting your wallet pickpocketed is that it only impacts you.

Yeah. Let's hope that, you know, all these services that were relying on Striker are the ones that start pushing back. " It has to be a differentiate- differentiating feature, right?

It, it, it- It must... It, it has to be. And Tracy, you're right.

I mean, when, when... Years and years ago, when Target got hit with a very simple attack, right? That hurt them.

Their... They, they lost market. For a good two years, people didn't trust them.

Right. Yeah. Now, now my problem is, you know, let's say Target got hacked tomorrow.

" Right? Yeah. It's becoming the new battlefield.

That's what I'm saying. Uh-huh. Right?

And I, I, I'm curious how many more of these we're gonna see in the next few months as, you know, they're, they release what they've already done, what, what hacks they've already done. Right. Do we have a- It's gonna- Do we have a grid, an infrastructure grid that, you know, electrical grid that's about to be t- taken down?

Maybe. But people- It's possible. I mean, I'm not- That'll, that'll get people's, that'll get people's attention, right?

That would get people's attention. Your infrastructure. That, that will.

Yeah. That will resonate. But- Yeah.

To Tracy's point, it doesn't change until the citizenry decides that they're fed up with petty crime and all this other stuff, and then they start voting people out, and then voting in people who will take care of it. And that's basically what's gonna have to happen here, right? We're gonna have to have s- people screaming about it, because if it just continues well as is, you know, hardly...

Heck, we hardly cover any of these low-level ransomware attacks anymore because there's just so many of them, and, you know, the media's not gonna, like, keep covering something that no one's re- else is screaming about. So we need some help on this one, folks. So if it bothers you, say so, and write your congressman, and start pounding on people about it, 'cause if it doesn't come from you, it isn't gonna happen.

I would also just like to thank our guests for everybody for being on the show. My Fridays are always the same. I'm sure they are for everybody else.

" But I'm gonna let you all get back to finish up your work, and we all know that everybody's rushing to get home on Friday. And I wanna thank you all for spending some time with us. I wanna thank our guests.

And please stay tuned for the rest of the rerun of the Techstrong TV lineup. We'll see you guys Monday.