Episode 1: Modern Alchemy, a Country of Geniuses, and the Day We Pulled the Wrong Plug

Greg: So, we're here. First episode, episode zero if you will, of Modern Alchemy. And why do we call it Modern Alchemy? You said sounds good in the chat. Because we take two other podcasts, we put them together and make a third.

Zane: I'm just trying to think what alchemy is.

Greg: We've talked about what AI is becoming and what technology has become, and it's just like modern freaking alchemy. It just freaking works. There's a lot of stuff that goes on in the background, so many things perceived as so easy that it's magic, which is the perception of alchemy. Enterprise IT connect flow, a little play on words, and the merch coming, kids.

Zane: You got to wonder, is it really gold? Is it based in science, and how much of it is science, how much of it is magic?

Greg: And that's why we're here. It's a foundation built on trust. So, Zane, I've always been curious, what is your entire IT journey? We both work at Pure Storage.

Zane: I had a relatively long career in IT. But before that, I really started in printing. Ran a press in high school. I was going to be a graphic artist. Sat down my first class in college to do that, realized I don't want to draw for four hours. Still life, not my jam. I really like the mechanics of things. So I pivoted to audio engineering. Actually went to audio engineering school, did some live sound. Through that I fell in love with electronics, got back into college, studied physics so I could be an electrical engineer, and ended up with decision science, in the information assurance department, which I later got to teach at. Did IT for 12 years in various forms up to management and director, did adjunct teaching, and combined the two and found this life of sales engineering. It's a life of curiosity. Ooh, what's that about? Could I go do that?

Greg: That is the one prevailing trait, the curiosity. Every person that's not just in it for the job, they're continuously curious. I'll tell you the greyhound statement. When I was at an MSP, there was an architect firm we were taking care of. Him and I are trying to figure something out, roaming profiles probably. And the head architect goes, you guys are like bloodhounds. Bloodhounds don't smell better than any other dogs. They're just tenacious and won't give up till they find it. I took that as a compliment, but I kind of think you just said, you're not smarter than other people, two fools just won't stop until you fix the problem.

Zane: You're just too dumb to quit.

Greg: And I was like, that is what somebody in this role does. Willing to bang their head against it until they solve that problem.

Zane: So how about you? I know you have a larger career than Pure.

Greg: Many years ago, before I got in it, I was doing product management and store design for Walmart. A very cool job for someone in their early 20s, travel across the country, help remodel stores. I'm running Cat 5 across stores. The real work is done in the middle of the night. 3:30 in the morning, 75 feet in the air on a scissor lift running Cat 5, going, huh, this kind of sucks. So I went back to school, got a couple degrees, dabbled in VB.NET. Grew up in the Hampton Roads area of Virginia, you can sling a cat and hit 15 military bases. So naturally I became a contractor. Did that for years until a co-worker said, have you ever heard of this company called EMC? That turned into being an implementation engineer, a hands-on person, learning the best way you can, by doing stuff. The DMX, Clariion, which became the VNX, RecoverPoint, for years in the federal space. My coverage patch was the planet.

Greg: A good place for someone in their late 20s to see the world. But let's be honest, this is not the exotic travel people think. Oh man, you went to Los Angeles. Yeah, I was in a data center for four days, I hated myself and ate bean burritos out of a microwave. Then I was in the partner space, enterprise architect, and found my way back to the mothership, which had been absorbed by Dell, got really heavy into hyperconverged. VxRail, VSAN, SimpliVity, Nutanix, you name it, I probably got certified in it. Then I became the software defined specialist, the global field CTO for software defined solutions at Dell, which brought me here to Pure. From running Cat 5 on a scissor lift 70 feet in the air at Walmart to here, it's an interesting bumpy path. There's a lot to be learned in running Cat 5.

Zane: When I taught adjunct, one of the things we did was make Cat 5 cables. Realize why you buy it. It's the ditch digging of our day. Builds character, by golly.

Greg: So one of the things we want to get to in this podcast is interesting topics. You shared an article from Dario Amodei, CEO of Anthropic, about a couple of things as we approach the singularity of AI. What is the honest take of where we are? It was fascinating, we're at a fork in the road. Is humanity mature enough to deal with the power? Uncle Ben, great power comes great responsibility, except we really freaking mean it this time. There's no coming back from this. It's either T-800s in your face or everyone's life is improved. Are we really evolved enough to understand, hey, we should do this, but if we want to, there's a commitment. Can we actually come through on the commitment?

Zane: So it's from his essays site, you can see all of his essays. The one I shared, and you had a good point, is the adolescence of technology, but Machines of Loving Grace was a better one. He has such an almost boyish outlook. He was a biologist, so he really looked at AI as, oh the intricacies and the leaps and bounds and discovery we could be at if we had these tools that were very helpful to humans. Not, let's build a tool that replaces humans, but is at the disposal of humans to do discovery, take away the minutia, the error correction, like in biology and drug trials. But with that becomes great responsibility. If you're just putting everything into the AI and letting it make the decisions, that's where this dark turn can take. I really like his outlook of, AI should be for humans. There's going to be a lot of jobs that humans should still verify, understand what the output is, interpret it. We're still humans, so humans should still be involved.

Greg: He uses a phrase, powerful AIs. Not talking about AGI, but powerful AIs. He summarizes this as, and I quote, a country of geniuses in a data center. That's just the capability. Altruistically, that sounds fascinating. We could solve a lot of problems, from a human genome perspective, disease mitigation, quality of life from a global impact. Will folks actually use this power to advance themselves further, or will they use it as an altruistic way to bring humanity up? That's a much deeper philosophical question. That's Star Trek versus Terminator, living with the AI and using the tools, the utopia, as opposed to the tools block out the sun by setting off nuclear weapons so we'll behave. Do you want to be able to say Earl Grey, hot, and get tea out of a magic machine, or do you want to be dipped into lava and called Earl by a robot?

Zane: There's a lot of flaws in humanity once the machines figure it out. Do you build fallibility into AI so that it also believes it's like one of us? Or is it empathy it needs? Or do we have to set a new set of rules? How do you impose ethics, but most importantly how do you impose morality?

Greg: It's a very fine line. We ourselves are flawed. We have different moral codes, different standards of ethics. This is the kind of culture that AI has inhabited. Innovation comes at any cost, which can be a little scary. You ring that bell that can't be unrung. But what I also noticed, I had breakfast with someone today, we were talking about our kids. I have a 21-year-old son. Our kids don't really trust AI. And it's fascinating. While we're talking about optimization, I use Claude to make a front-end app for me in like four sentences, which still blows my mind. Twenty years ago that would have taken me a week. But it's interesting, the generational gap. Is our next generation the John Connors to fight the Skynet?

Zane: You would assume the kids love the technology. I have not seen one sign that says, oh it's great. Is my view skewed because we're in the industry? They were raised with the internet. I wasn't. They do it in a very different way. Or will they kill it? Or just have it woven into their life so much you won't even see it anymore.

Greg: If you look at what's going on, the thought leaders in the space are very diverse in their thoughts. Folks on the hardware side of AI, manufacturing GPUs, have a very different stance. But you look at the software side. We're recording this in February 26, what OpenAI just announced. Could this be the zero event of looking at what OpenAI has taken from a software driven approach, what Claude has evolved to, and what Anthropic's leader is saying, hey, we'll put this out here, however this is not just buyer beware, this is, holy crap, did we understand we just gave all the nukes to chimps? It gave me a cold shower kind of moment of, wow, we really are at that point that quickly.

Zane: You could build in all the regulation, but it's moving so fast you're not going to be able to keep up. Do we want to do this, and are we doing it for the right reasons? The hardware manufacturers have a different angle. Anthropic takes a different take, they try to minimize their footprint on data centers, they spend less than the others. The others are maybe coming from a VC background, trying to get evaluation. So there isn't just one overarching AI. Different approaches, different companies, different models.

Greg: Think about how much of a blip Deep Seek was. About a year ago, everyone lost their mind. It was Chinese secrets, there's no way they could have done it. They did it through older hardware and optimizing software. They just did better math. It impacted our market here in the US for like three weeks. But then, back to pushing more models out every single week.

Zane: What are the things you can do if you want to go into full MREs in the bunker mode when it comes to your digital footprint? Guess what, you're already archived, dude. We already got you. You're in the system.

Greg: I know they're measuring me on every moment. It's not what do I have to hide, it's how am I being commoditized. I've shut down some things because I don't want to be measured there. I'm taking the stance of fully adopting and using AI, just exploring it. I talk to it like it's a human. I ask what it can do for me. I'm not going full parasocial, not going to break up my marriage. But it's interesting, this changing model of how software licensing is going to go. If I can interact with one tool that interacts with all my tools, I don't need individual accounts with each one.

Zane: There's a phrase I've used, the tool is not the skill. AI is a tool. It's not the skill that's going to be replaced. And there are some signs we might not need to be too terrified of AI in its current state. A good buddy of mine, a big Tesla fan, has a fleet of Teslas. Him and his wife came over, and they went to leave, all about, hey, full self-driving, man. He takes his hands off the steering wheel and Mr. Full Self Driving starts to back out and almost turns into my house. Full self-driving couldn't get out of my driveway.

Greg: Did the Tesla know that maybe you were a hater? I think I saw someone try to get it to pick him up in a parking lot in Arizona. It would go in and out of the driving spot. We walked there twice by the time it didn't come to us. It will do exactly what you ask it to do. It is not natural understanding intelligence. I saw a funny LinkedIn post, an AI influencer talking about how he optimized his path around the Grand Canyon, saved a lot of time. And the next post is, I asked AI if I should catch an Uber to work or run to catch up to the train. Oh, you should absolutely run and try to catch the train at the next stop, you'll get there much faster. So you're assuming this man can run as fast as the Flash.

Zane: It'll give you exactly what you want, but it needs context. It doesn't understand parallelism to human experience.

Greg: As long as AI never understands the human experience, I don't think you have anything to worry about. Don't build in the human empathy. It just doesn't need to know the human experience. Keep it artificial.

Zane: I was in a Waymo in San Francisco, it was great. But it came up on a road, I should be in this lane. It didn't see that there was a truck with the gate down doing an unloading. He's going to cross the intersection, I have to change lanes, everybody knows that. Didn't have the context.

Greg: Here's one thing where me and Mr. Musk are in lock step. Data centers in space. If you can get these donkeys to stop putting data centers by my house just because I have a water source near me. Let's launch them, buddy. Utilize Starlink. You have to get the macro view to understand the micro move.

Greg: So every episode, we talk about your quote unquote welcome to it moment. It could be a really cool thing, or a resume generating event. Everyone has one. If you don't have one, you're lying or you got fired for it.

Zane: It ties in nicely. When I did a transition into that MSP, first day on the job, we were onboarding this healthcare client, all hands on deck. Everybody's moving around the data center. It was their main SAN. He's like, okay, we're going to flip the plugs over. So you do that one now, and he pointed at it. I remember questioning it, like, I don't think that's right. But I just blindly listened to the output, and I yanked the thing out and we just dropped their whole data center. I think they were running Epic. It was bad. Nobody knew how to get back into it. We got them all back up. But the funny part was, I get to work, they set up my station, and the screen background was a picture of the back of the SAN power supply with a circle around the one I was supposed to pull. That was my screen. I lived with it for a week. I knew he was wrong and I didn't question him. It was the stupidest thing because it was a really simple task.

Greg: I'll share an edited version for my own sake. Big background in federal, doing implementation work, many years ago. Allegedly a building that may be a geometric shape in the Washington DC area, doing a SAN migration. Who doesn't love a good old-fashioned brocade director migration. The most important thing is knowing which fabric we're doing first, especially with legacy applications that are single-pathed, some really old COBOL applications that don't like losing their path to their data. I'm about 10 months into the job. Pretty smart guy, I know my left from right. We said, hey, we're doing fabric A first. I'm in the data center working with someone remotely over Skype or MSN Messenger. And I had this flash, like Spaceballs, they're getting ready to go to Ludicrous Speed. Are you sure? Ludicrous speed, go. The last message I got was, ready, good to go on fabric A. I go to fabric A. I'm in a data center, you don't hear the AOL instant messenger alerts. Oops. Meet fabric B. The A and the B are not anywhere near each other on a keyboard. And these are directors, not tiny little one-U's. Ludicrous speed, go. Never ever talk about this again. Once everything got their paths back, everything's fine. But communication is the most crucial part of any migration. It is not planning, it is not execution, it is communication, ladies and gentlemen. Check your A's and B's, kids.

Zane: Trust but verify. The old phrase.

Greg: It's timeless. Well, I think we've gone a while here. It's always interesting to talk to Zane. I'd welcome feedback, what you'd like to talk about. See what their take on the world is. And we'll give our infamous sign off. Kids, don't forget to drink your Ovaltine. Thanks everybody.