The $15B AI Platform Blueprint for Leaders to Manage Risk and Unlock Creativity with Skip McCormick

Jay Combs: What was the secret? How did you drive innovation at a company or an organization that has centuries of hey, this is the way we've always done things. This is how we do it. How do you bring in new innovation and push them to change?

Skip McCormick: It is surprising when I look back on it. I think probably this pivotal moment for me was

Jay Combs: Today's guest is Skip McCormick. From the Naval Academy to Navy warships, then 17 years in the CIA, Skip McCormick has spent a lifetime at the intersection of national security and technology. Now as CTO and founder of Angularis AI, he leads a powerhouse team delivering enterprise-grade AI governance and model risk management.

Skip McCormick: I have a few sayings. One is if you're not cheating, you don't want to win bad enough. Now, I'm not advocating unethical cheating, right? That's not the point.

It's like if you're not taking advantage of every ethical opportunity that you can.

Jay Combs: How important is traceability and transparency in what you're doing at Angularis?

Skip McCormick: It's essential, but it's not enough. like the AI is looking out for all of us and to do that. You asked about how important is transparency. I think that's the key when it makes a decision.

Can I open up that decision and say how did we get to this decision? I want to be able to do that whenever there's a decision enough till I get to the point where I say I'm confident that this decision is going to be like the other ones I double clicked on and follow.

Jay Combs: Is there something you look back on, hey, that was a mistake, but at least I learned something from it.

Skip McCormick: That's a dangerous question for me because I've been told by a few people who became good friends over time that when they first worked with me, they thought I was

Jay Combs: Welcome to the Good Decisions podcast. The podcast about the pivotal choices shaping careers, technology, and the future of AI. I'm your host, Jay Combs, the VP of marketing at ModelOp. In each episode of this podcast, we're going to sit down with leaders who have faced complex challenges uh such as military officers, turned technologists, data scientists navigating billion-dollar risks, AI pioneers who've had to choose between what's easy, what's right.

We talk so much about AI and the technology and data about making decisions at enterprises. we can sometimes give short shrift or at least I think so to the people behind all that who are ultimately making decisions and what drives them. So that's what's interesting to me and that's what we're going to dive into in this podcast. So the people in this podcast uh we're going to talk about how they make decisions and the moments that got them to be the leaders in AI today.

And so for our very first episode we couldn't have asked for a better guest. Skip McCormick who's the CTO and founder of Angularis AI. He's made a career out of complex situations and good decisions. From joining the CIA after 9/11 to building supercomputer systems to fight terrorism to leading AI risk governance at one of the world's largest and oldest financial institutions.

Uh he's now CTO at Angularis AI and helping enterprises navigate AI with speed, security, and integrity. Uh so let's get into it. Welcome Skip. How are you this morning?

Skip McCormick: Hi Jay. I'm doing great. Thanks a lot. I'm very humbled by your introduction.

Jay Combs: Yeah. Fantastic.

Skip McCormick: I don't feel like I'm that remarkable. I just sort of played the cards as they were dealt to me. Yeah.

Jay Combs: Well, so let's get into that. That's what I kind of want to know because it really is fascinating. Like you started off at the US uh Naval Academy in Annapolis. Uh you served as a surface warfare officer and I think you were uh then you shifted to systems engineering and in consulting.

I mean you developed a client list that I believe included the NSA, but correct me if I'm I'm wrong there.

Skip McCormick: I did support the intelligence community quite a bit. Yeah. Yeah.

Jay Combs: And then obviously, you know, 9/11 was a horrible and transformational event for me for the country and for many people. Um, you went from, I think, private practice there into the CIA where you ended up serving 17 years as an intelligence officer developing data science and supercomputers. Is that right?

Skip McCormick: Right. Yeah. It was a super fun career. I was very fortunate.

Yeah.

Jay Combs: And then you went back to private uh private business and served as the managing director of governance and metrics at the Bank of New York which is uh over 240 years old at this point and you helped innovate the firm's AI practice and led their AI risk and governance for their new artificial intelligence hub. And now you've uh become the CTO and founder of Angularis which is where you created I guess it's enterprise-grade managed services for the governance uh and innovation of AI models.

Skip McCormick: That's right.

Jay Combs: So like you said like how did you get there? I guess that's the question like did you plan on doing both public and private service? Did you think you'd be a career military officer? Like as you look back how did you get there?

Did you Is this surprising to you? Is it not surprising? What kind of shocks you the most about that journey?

Skip McCormick: It is surprising when I look back on it. Um I think probably the pivotal moment for me was I was 12 years old and my dad came home from work. He was a grocery store manager and he was all excited about one of his box boys was going to the military academy. And I didn't know what that was.

My dad was like, "This is like one of the best schools in the country, and I thought you had to like know somebody or be, you know, family friends with a senator or a congressman, but it's all just merit-based." And I was like, you know, when you're 12 years old and your dad hung the moon, I'm like, "Well, that's what I'm going to do." And my dad looked at me said, "Yeah, okay, but the Naval Academy is the one you want." And I was like, "Okay, whatever that is, that's what I want."

So when you're 12. So I went to school the next, you know, next day or two and I went and saw the counselor and looked up the Naval Academy and I wrote them a letter to get ask for an application and uh they sent it to me because I didn't tell them I was 12. I just asked for an application. And so I got this uh you know IBM number two pencil form that you know you fill out and you have to spell your name with you know little dots and everything.

if you're as old as me, you remember those forms. And I filled it out and I just I didn't have much to put on it because I was in grade school. But um I now had the roadmap for what I needed to accomplish in junior high in high school. So I could check off all these things like were you in the student government?

Were you captain of a varsity sports teams? You know, were you an Eagle Scout? All of these things. And then there was forms where like if you ever used um any drugs, it basically said you don't have to complete the form or bother sending it in.

Right? And so when you're 12 going into junior high school, that was like a magic capability. I remember talking to my seventh grade math teacher, Mr. Mosley, and saying, "Mr.

Mosley. Um, I need to take calculus my senior year in high school. And to do that, I have to get into algebra 1 by um, ninth grade, which means I have to get A's in this course this year, right? And he looked at me like, who?

What? Right? But if a kid like that tells a teacher that you're basically giving the teacher permission to bust your butt if you're not if you're not doing well enough, you know, I didn't tell him, give me an A. I told him, make sure that I know if I'm not doing enough to get the A.

And so fast forward, I'm in high school. One of my best friends is like the school drug dealer and uh we're riding the bus and he and he's talking about drugs and I said, "Sounds great, but I can't do that because I want to go to the Naval Academy. If I do that even once, I'm excluded." And he was like, "Okay, that's cool."

So I'm in high school in like a party and somebody's coming by trying to offer me marijuana and the guy who sold him the marijuana comes says, "Don't give him that. He can't do that. He's going to the Naval Academy." Right?

So after, you know, if you share your dreams and your visions with other people, they sort of adopt it too, right? Unless you're a jerk, and I tried not to be a jerk. So once I had that sort of set up, then I kind of had to succeed because all my friends, you know, were counting on me, right? So when I got into the Naval Academy, I felt like I succeeded.

But I remember being there a few months. It's hard the first year as you probably know. I remember one day I went, "Oh no, now I'm going to be in the Navy. What am I going to do in the Navy, right?

My whole goal was just to get there." And I guess I didn't I wasn't 100% sure I'd make it, right? And so now I'm there. Then I have to decide what to do.

And so, of course, everybody says, "I want to be a jet pilot." Well, my eyes weren't good enough to be a jet pilot. So, I ended up um having to think about what to do. And I went surface line and went out and served the fleet out of a ship in Hawaii.

And it was great because I was on a supply ship which meant we went to sea all the time and we did our mission every day. And as a junior officer that's the best, right? You're learning how to do your job. If you're on a combatant, you only get to do your mission if there's like fighting.

Otherwise, it's all just practice and pretend. So anyway, that was kind of how it started. After um being in the Navy and uh enjoying that, I finished my obligation plus another year went into the reserves and then went to Washington. My last couple years in the Navy, I actually wrote software.

I had this neat opportunity to go build Lisp code for uh AIS security, information security work at the regional data automation center in Hawaii. You just don't often get a job to do in the military what your schooling training was, right? It was just a free sort of opportunity. I loved it.

I went to Washington uh DC area and looked for a job in computer security and this is like 92. It's all pretty new then. Y and uh you know

Jay Combs: and you your degree at the Naval Academy wasn't computer science, correct?

Skip McCormick: Yeah, it was computer science. I was there in the first year they offered that degree, right? It was uh they didn't even have it before that you could take applied mathematics with a emphasis on computers, but that was the first year they offered computer science.

Jay Combs: How did you select it? Like why? Like if it was that new

Skip McCormick: I didn't select it. I selected English. I was an English major for my first year and uh at the end of my first year, I was filling out the form to pick my courses for my second year. And I wanted to take introduction to computer science because the programming courses that every plebe had to take were fun and they were my easiest grades because I loved it.

And my adviser was like, "No, no, not computer science. You need to take Shakespeare and poetry and you know, he's giving me the English lit courses at the Naval Academy."

Jay Combs: Your adviser said this? Yeah. And um, Commander Shank, he was he was a great he was a great adviser. And I said, "Well, I think knowing computers is going to be important as a naval officer."

And he said, "Well, that's true." And then I said, "You know, all the papers I've submitted, the ones I got A's on were the ones I used the word processor to write." This was an early word processor on a mainframe, right? Microsoft Word didn't even exist yet.

I think uh WordPerfect or WordStar wasn't even out. So, I was writing my papers on the main frame, which would fix my spelling and my handwriting was terrible and all that. And I said, "Look, when I submit my papers from the computer, I get a because I'm I'm a reasonably good writer. I'm just terrible at spelling and handwriting."

He said, "Well, that's true." And I said, "But the ones I write in a blue book, I'm lucky to get a C because it's an it's unreadable and my, you know, my spelling is atrocious, and don't ask me to spell atrocious, right?" And he was like, "Okay, you have a point." So he said, "But you have to take these courses anyway."

So I went and I gave up my summer to take Shakespeare and I think poetry so that I would have the open um electives to take uh introduction to computer science. That's how much I loved it. And that course changed my world. The introduction computer science was taught by the dean of computer science.

Deans like to teach the early courses because they because it's it's a way to meet the new students I guess. And it was the first year of the program, right? Like so it must have been a brand new dean. Yeah.

Yeah. And the funny part was the first quarterly exam um there was one question on the exam. It was some logic question and it had like this much space for the answer and I answered in one line and I was sure I got it wrong. It's like I must not understand this question.

And like the next class I'm in there and he hands out everybody's um exams graded except me. I said, "Oh no, I'm in big trouble now." And he's going through and he gets to that question. He goes, "There's one student in the class who answered this in an interesting way."

And he put my exam up on the overhead projector. Right. Back then they didn't have, you know, Yep. and he says he says he said something like this is the most efficient and most unique answer I've ever seen to this question and then he explained it right and I was like oh wow right and the rest of the class is looking I'm like you suck up you know and I'm like what I didn't know well and so

what kind of question was it was like a short a search or

Skip McCormick: some kind of logic question where you know there's all the space to build truth tables right and I just wrote logical formula And so after class, he handed me my exam and said, "Good job." And he said, "Come see me after classes today. I want to uh um have you help me on my research." And I was like flabbergasted.

And I said, "Uh, professor, that's awesome, but I don't think I can do that. I'm an English major." And he looked at me and he said, "Oh, well, come up after class and we'll change you to computer science. That's not a problem."

Right? It wasn't even like would you or do you want to? It was like, you know, this is the right thing for you. So, I remember going back to Commander Shank and saying, "Commander, will you sign this?

I'm going to change my degree to computer science." And he was like, "Why?" And I and I told him, I said, "The computers make a two-grade difference in everything I do. I think that's where I want to be."

And he and he said, "That makes sense." And he signed it and I, you know, from that point forward, I was, but the funny part was I was a computer science major who had four semesters of German and Shakespeare and poetry. you know, I had a whole year of uh English lit stuff and my computer science peers didn't have any of that. And it's funny with all the ChatGPT stuff happening now, all of that lit stuff turns out to be really relevant.

It's like Steve Jobs did all of that um kerning and font work, right? It's like, what are you doing? You're a computer nerd. It's like, well, it turned out that how it looks and how it reads and writes actually matter.

And so it was for me it's just dumb luck, I think. But uh that was it's funny. I'm glad you asked the question because I sort of forgot all about that. Yeah.

Jay Combs: I don't know if it's dumb a lot. I mean at 12 years old clearly you were very mission driven. And your your your parents weren't in the uh military, correct? Like you were the first to go in.

Skip McCormick: My dad was in the Air Force. That's where he met my mother in England, but he was he was enlisted, did a good job and did his time and then got out, right? And you know, he wasn't a career military or anything. He wasn't um he had nothing against the military.

you know, he has had a positive outlook on especially for a young person isn't sure what they do. I think I think military makes a lot of sense for a lot of people because it's like go do something useful, learn how to be productive and how to sort of be part of a team. All those things can be advantage.

Jay Combs: So yeah, it's I mean at 12 years old, I think it's a very fairly unique or if not really unique to have like a very clear sense of mission and focus. So that's like one interesting thing come out of here. But then also it seems like you were also somewhat uh very logical and data driven like hey if my grades change this much I should be doing this and ahead of time and it's not that you were it doesn't like clearly you're still interested in the liberal arts and uh but also hey I'm really good at this. Yeah you following your passion as well.

It's like, hey, I could go the path and kind of stick with what my superiors are, you know, advising, but hey, I'm really good at this. Let me follow my passion and try something new. Like you said, you're in the first class to do that. That's really risky and crazy, but it from what you said, it sounds like, hey, you really enjoyed it, so you're going to try it.

And you're at the intersection, too, of the liberal arts and technology and the military. So, that crossover is really fascinating as well.

Skip McCormick: I think for me, it's fortunate timing, right? that this new thing happened and um uh and I went for it and uh you know you sort of you make the decision then you just work to make the decision work out right it wasn't all easy some of those courses the computer science courses included um heavy-duty electrical engineering back then and you know uh differential equations and a lot lot of math everybody at the academy studies engineering because you're going to be in the Navy but uh the computer science major they were still figuring out what it should be and so they included a lot of electrical engineering material and some of that stuff was uh mind-bending.

Jay Combs: So somehow I muddled through yeah there's so many aspects of computer science. It's not just programming. There's the hardware. There's the algorithms.

It's such a cool cool I was a computer science major as well at a liberal arts school which was a little odd. It was a very small appreciate Yeah.

Skip McCormick: Appreciate Yeah. where you went and you probably saw that most of the computer science education wasn't just the classroom stuff. It was helping classmates with their programs, right? You'd have these assignments.

You have to build some program that does X. And uh you'd go to the computer room because people didn't have personal computers and you were in a like a room with a bunch of uh terminals attached to the mainframe and you'd sit in there working and a classmate would say my program won't compile or what's wrong. It's not working right. And so you'd get their print out and study it to see how they're solving their, you know, that program.

And then you could say, "Oh, here's here's what's wrong. You're using the wrong index or you need a different loop or stuff like that." when you're helping other people debug their programs, that's when you learn a ton of different ways to approach the same problem, right? And so, you know, that was looking back on it, the other students helped me learn computer science is really how that worked.

You know, the usually by then I was working on making my com my program, you know, neater looking. It already worked and then I was I like to do everything in as few loops as possible just because it was a puzzle.

Jay Combs: So do you think that collaboration is changing now like with just AI vibe programming and different things like are people relying less on that collaboration to learn from each other especially with a lot of remote work happening or is that still a big part of what's going on?

Skip McCormick: You know I wonder I hadn't thought about that but I'm not sure if development works like it used to where you write something and you're stuck and you ask a friend. I think a lot of uh programming now is you go to Grok or perplexity or something and you have it write your first version right and then and then you tweak that and uh I don't know many programmers today who don't start with a first version that was written that you know written by AI that's the normal way to begin and uh why wouldn't you start that way right it's there's there's a whole bunch of sort of mundane stuff that needs to be um um done you know I don't know some of these some of these products, you know, like I've mentioned, I use Grok a lot, Perplexity a lot, you know, everybody uses ChatGPT and, you know, and the meta ones, but um, you know, they give you a pretty good starting point. When I'm done, what I wrote doesn't really resemble what I started with, but it's a lot easier than starting with a blank page.

Jay Combs: Yeah, it definitely is. Yeah, it's an interesting topic. I think applies to writing as well. I know there's some studies for writing.

I use it for outlines like what are the my most common queries are give me the top seven you know bullet points with one paragraph on X right and uh with some of these tools like uh Grok it'll give you that all reference to where it came from you know if you just use that you're in you're in for trouble because it makes up things but it's it's a really great starting point I think that's probably different you know so I don't know if people go to their friends anymore for help debugging. That's kind of sad. I hadn't thought about that. I haven't gone to a friend for help for a long time and I don't know how professors teach this anymore.

Yeah. How does a professor teach programming these days? You know, it's like Yeah, that's a challenge. So, at this point, I think every paper submitted the professor has to run it by ChatGPT to see if ChatGPT wrote it.

You know, there's there's a Yeah, I was listening to some study. Yeah, there's a I mean, ton of questions that I think folks are trying to figure out. I did uh hear one study on writing and they um you know, ask you know uh I think they tested um the quality of papers written by uh by AI. They had a bunch of students use AI to write a paper and they had a bunch of other ones just do it on their own.

And I think as a whole the quality of the generative AI papers were better quality like just overall in terms of uh writing style and everything better outline better structure but um the ones in my human they had higher peaks like the ones that were really good were really good and more creative and more diversity of ideas whereas ones with the AI had more similarity of ideas and concepts. So you might get the better structure and overall it's a little better but like the uniqueness and like pushing the ideas still comes from uh from the human side at least from this you know one study.

Skip McCormick: I agree. I remember probably five years ago, maybe four or five years ago, an interview with Elton John about the music industry and he was ranting about um now with all the modern technology, anybody can make any music they want, right? And it was and it made him crazy, right? And I was like, well, that's kind of arrogant.

And his point wasn't that he was like, I spent my whole life mastering this field and now um people using uh computers can make something um in a couple minutes that's um mediocre and good, right? It's not a it's not a creative genius thing that changes the world, but it's it's mediocre and good. And it frustrated him. You know, I don't know how you feel about Elton John, but some of his music, you know, was like, "Wow, you know, you never heard anything like that before because it came out of his head."

And now now you can tell AI to make a song in the structure like Yellow Brick Road about cats. Yeah. Riding on rainbows. And you know, a minute later, you have a pretty good production of a stupid song about cats on rainbows that sounds like Yeah.

the Yellow Brick Road album is amazing and so unique. But uh so it's wild but I don't think GPT would ever make a Bohemian Rhapsody. You know what I mean? That broke every norm.

You know what I'm saying? Yeah. Exactly. Those peaks are Yeah.

That original truly original content I think is still so challenging and uh the most interesting. But I think Freddie Mercury would use uh AI tools because he was he was into technology big time. Yeah. You know what I mean?

He would use them. I mean going back just like the word processors and I know you've used that analogy on word processors related to AI governance in the past but yeah it's a tool. How do you use it to push the boundaries do something original not just something mediocre? I think is like the true power of what all this uh generative AI and other technologies can do.

That's the right way to use it. Not just to be mediocre. Yeah. Yeah.

I'm thinking of a lot of other things. In the 90s, I bought a software product called Band in a Box. Have you ever heard of that? I'm a musician and I was learning to play bass guitar and uh you know, there's a lot of things bass guitars do that isn't obvious, but Band in a Box would add a bass and a guitar and a piano and drums and you just give it chords and tell it what style and then it would it would, you know, make music.

And then could look at the music and see what the bass player was playing. And I remember looking at that and learning from that and that's like, oh, this is how you do bass for a samba or this is how you do bass for rumba or for, you know, hard rock or whatever. And you could see it and it was, yeah, band was giving you the standard basic mundane, you know, uh, boring stuff, but that was what I needed to learn. And so, yeah.

So even at that point I we wouldn't have called that AI but you know it was it was computer-generated music from chords and uh

Jay Combs: no it I mean just I mean the concept of tab and then tool like now I go on my like also like similar thing like I've used tool like you know YouTube or yes the tab programs that translate music into tab automatically like it's incredible right back in the day you just had to like put on the record or the tape or whatever the CD and rewind it and figure it out and transcribe it yourself and now it's it's done so quickly. Were you playing bass in the Navy or were you learning back then? And is that part of uh your life back then?

Skip McCormick: I was in the church's worship team at Pearl Harbor Station. My wife is a singer and I would like to spend time with my wife so I joined with her just so we could do stuff together. And uh we found a drummer and she would join but she would only join if there was a bass player because she wasn't that confident. We couldn't find a bass player.

So, so I went to the guitar store and I bought this. I didn't have a lot of money and I bought a bass guitar that was taken apart in a five-gallon bucket and I took it home and put it together and soldered it together and I said, "How hard is it to play roots and fifths?" And that was when I started playing bass and it was just so we could have a drummer. But, you know, I did just like you said, you're listening and trying to figure out and I could play a few things, right?

But ended up being pretty fun. You know, I'm not a great bassist now, but I can play, you know, whatever I need to. Yeah. So, that's so cool.

When people ask me, "How should I learn guitar?" I tell them, "Learning bass guitar is easier because the strings are really fat and they won't hurt your fingers as much." Yeah. Yeah.

You can start with four of them. It's easy. Yes, that too. And you only play one at a time.

Yeah. Fair enough. Um, well, yeah. Back to Hawaii.

Jay Combs: when as a surface warfare officer like were you using your computer science degree in that program? What does a surface warfare officer do?

Skip McCormick: Was the Navy was uh obviously there was computers used in the Navy in uh weapon systems and things like that. The Navy's been using computers uh you know physical computers since you know before World War II.

Jay Combs: I mean it's really incredible. If you look at targeting system was differential hardware computers that would like aim the gun that based upon you know range and bearing to the other ship and all this stuff so that you could put a put around where the ne other ship is going to be with both ships in motion right it was complicated things and the computers were like gears and cogs and stuff if you know if that kind of stuff fascinates you it's a really worthwhile thing and there's a bunch of YouTube videos you can watch about it so computers were not new to the Navy but personal computers was kind of a new thing and I remember our Captain Egan was kind of forward thinking and he uh used some of the ship's resources to buy a few laptops for the officers. So these were all, you know, early Zenith laptops. They were, you know, LCD, you know, black and white displays, you know, and they probably had like 500K of memory, you know, it's just it was very early.

And um and you got the computer and you got like a one of those thermal printers with it so you could print out. And I remember using that computer to automate a lot of my junior officer duties, right? Inventories for supplies. you had to write all these reports.

You had to write fitness reports for your the division and stuff like that. So, I used the computer uh increasingly a lot, right? It really helped me u get things going. I remember helping some of the other officers uh with it.

They're like, "How did you do that?" And I would show them and you know, bring them a floppy disc with my WordStar, you know, uh template for a fitness report kind of thing on it. And I think that's why the captain recommended me for the next job at the at the Naval Regional Data Automation Center, which isn't a typical next step for surface line officer. But um he saw that uh that was a that was sort of where I could do the most and I'm glad he did.

He did me a big favor. Yeah, it seemed like you had some really good mentors and leaders kind of helping you, not necessarily pushing you down this path, but opening those doors, giving, you know, bringing you a laptop. How did you see technology different than your classmates at that time that who maybe went more traditional military routes or became, you know, career officers?

Skip McCormick: I guess I saw it as an advantage, right? It's like I can do this job faster and better with this computer than I could without it, right? And so I'm always uh I have a few sayings. One is if you're not cheating, you don't want to win bad enough, right?

It's like there's Right. That's a good point. Yeah. Yeah.

No, I'm not advocating unethical cheating, right? That's that's not the point. It's like it's like there's if you're not taking advantage of every uh ethical opportunity that you can, right? And so there was that and uh and it was kind of an early stage too.

They were putting um they were putting Novell networks on ships, right? They didn't have, you know, um networks and the ships had to do it with fiber optic. And this is like 1991, right? And fiber optic was really expensive, but ships um can't do a bunch of add a bunch of wiring because it um it could give the ship away in an EMCON sense if the enemy ship is scanning the spectrum to find you, you know, and you have a bunch of uh you know, Ethernet uh activity going on, you know, that could give your ship away when you're trying to hide.

So, they did everything optical back then. and it was I don't know if you remember all the Novell stuff but you know that was kind of the beginning of local area networking and so I had a chance to sort of be in the center of that. But I often wondered I when I got to NARDAC they offered me a chance to work on the computer security team or the networking team and I chose the security one because there was a chance to do some programming and uh I often thought maybe I should have chosen the networking one because that was a more immediately marketable skill right there wasn't a lot of call for Lisp programmers but uh um to this day I still write code in my head using Lisp even if I'm writing in Java or C or Python I think in Lisp because that's how that's how I learned. So, it's not you like all those parentheses like parenthesy parenthesy parenthesy.

And I love recursion. It's like the best of all uh most challenges puzzle-wise. But um I think it worked out because I still had to learn all the networking stuff, right? That was that came anyway.

Very cool. So, but you mentioned something about um I was fortunate to have leaders and I think that was a that really was a blessing and it's something I like to emphasize going forward is um I've told people I work with they say you should never say this. They say I'm a pretty good leader but a really bad manager rather than like don't ever say that right? It's like no it's it's actually self-awareness and by what I mean is I'm not like a bad manager in terms that I don't care.

It's like I care so much about the people work for me that I don't want to be their manager because I won't do as good a job as they need me to do. So I try very hard to identify like a chief of staff on my team whenever I can to say look you're going to manage the team and I'm going to lead it and lead you. But you have to actually let them manage. You actually have to give them the budget and let them make the decisions otherwise it's no fun for them.

But uh I've always viewed uh management is very important for the people being managed. It has to be good. But I think leadership is more important and it's you know here this trait thing about leadership's more important than management. No, they're both really important.

And so playing to my strength, leading is uh there was a psych professor at the academy who said leadership and manipulation are identical except for the motive. Right? And I've thought long long and hard about that. In fact, I remember telling my son that and he said, "Not true, Dad."

Because if you feel manipulated, it still feels like manipulation, right? So, you have leadership is you're basically leadership is getting people to do the things they should for everyone's benefit, including their own, where manipulation is getting them to do the thing that helps you. And I think that's success, inspiration versus manipulation there. and ModelOp has a very good leader.

A couple of them, you know, your your founder and uh Stuart and Pete, uh those are things I like about them the most is they spend a lot of cycles on the leadership part and you're a beneficiary of that, I'm assuming. So, I am. I definitely am. Yeah.

And then, so yeah, how would you classify your leadership style? I mean, first of all, I love the self-awareness. Like I think that's so critical in when you have leaders or managers that aren't self-aware or can't reflect on what they're doing, how they're influencing others. Like that's that's not a good situation.

But it Yeah. But it seems like you've had leaders that were always self-aware or not always, but like from the examples you gave like trying to un like they bent their like hey you know maybe their position was hey English was way better to do but you know what this makes sense. like let me reflect on this and guide this person the right way. Not manipulate them but lead them.

Right. How would you describe your leadership style? Well, um for me I don't like to be micromanaged. Right.

And I think that's common with the more ADHD-ish kind of uh engineers is like tell me what the puzzle is, tell me what the problem is and tell me what the resources are to solve it. Give me an impossible deadline. That's that's okay, right? But then then let me try to solve it and be there.

So, I can I can bounce ideas off off of you and I can, you know, bring you some like challenges like um you know, I need more twine here or how do I do this, but don't um tell me to do it the way you would do it or don't make me do it the way you would do it. Tell me the way you would do it. That's good input, but don't expect me to do it exactly your way. And uh maybe that's why I like the um AI tools.

You know, you go to Grok and ask it for help. It you don't say tell me how to do this. you're tell you're saying tell me a way to do this right and then you see that and you go okay and then you work you know you sort of riff on that um that's the kind of that's the kind of leadership I respond best to and that's the kind of leadership I try to uh apply it doesn't work for everybody though I've had employees who needed a lot more um direct instruction and uh and sometimes they're very good but I'm not the right leader for them so if I could pick somebody on the team for them to report to who will tell them now do A now do B now do C um sometimes that's great but if they're reporting to me and they're expecting me to tell them what the next thing is um both of us end up frustrated

Jay Combs: and yeah and sometimes you don't know what the next thing is supposed to be even like a bigger context right like so you were so after you left yeah after you left the Navy you were in private uh I think you were a defense contractor for a while and then

Skip McCormick: ended up at which is a which is a system engineering company. That's where I learned to be a professional systems engineer. I was sort of you know seat of the pants engineer up to that point. But at MITRE, you know, it's proper system engineering with documentation and plans and structure and I had to I had to learn all those things to survive there.

But it's funny, I remember in the interview, I was interviewing with Ron Zahavi, who is a well-known author and a technologist and um he's interviewing me for this job and at the end of the interview, everything's good and he asked me that the sort of question that you always have to ask like, "Do you have any questions for me, young man?" You know, and most times in interviews, you're just like, "Oh, no, not really." But I actually did. I said, "What are the most three most important things I need to do to get to where you are?"

Right? That was my question. And it was kind of like, "How do I get your job?" Right?

And he sat back and he actually gave me the answer. He said, "At MITRE, you need to um be the program manager for a well-known program, you need to run a conference, and you need to uh write a book, right? So, if you do those three things, you'll succeed here." And I was like, "Okay."

So when I got the job, I that became sort of my next mission. And so I was hired for a project to run that was came with the job. And then um I said I need to write a book. So I was like in a hunt for what could what book could I write?

And I was at a or object world conference or uh I think object world conference in San Francisco. And I went I sat in one of the sessions by a guy named um oh I'll think of his name in a minute. Akroyd, Michael Akroyd, and he had this concept of um AntiPatterns, right? I didn't think of AntiPatterns.

That's where I got the idea. Design patterns was a big thing. And he was sort of a sort of a counter culture guy, right? And so he said design patterns are about here's the pattern for solving certain kinds of problems.

And he said these are patterns where I see people applying things and the consequences are worse than the benefits. Right? and he and he outlined maybe three um what he called anti patterns in that session. I was like this is brilliant.

Right? So I went up to him after the session and I said I love this. Are you going to write a book about it? He said well a lot of people said I should but I don't have time to write a book.

And I said I'll write the book with you right. I'll do all the work. I love this and I and I'll help you write the book and I just need you to collaborate with me to make sure I'm I'm not getting it wrong. And he agreed.

And so I went out on the floor where all of the vendors were later and there was a floor by Wiley, the book publisher and there was there was a lady there who ended up I didn't know but she was one of the senior editors and she was trying to build their computer science uh book uh sort of library and I went up and I told her that I said I um I just got Michael Akroyd to agree to collaborate on a book. we're going to write the AntiPatterns book. And she said, "We want to publish that book." And I was like, "Really?"

She said, "Yeah." And so, you know, how often does that happen where you want to write a book and before you even do an outline, you have a publisher that wants to publish it. And so, so I was like, "Okay, cool." So, I started an outline and uh and I emailed Michael and I never got a reply from him.

And there was another PhD at MITRE, good friend of mine, Tom Mowbray, who I talked to. I said, "You know Michael?" He said, "Yeah." And I said, "He's not replying to my emails."

And so he said, he said, "I'll send him a note." So I gave him the note. And he sent it. Still no reply.

And I said, "What do we do, Tom?" He said, "Well, we don't need him. Why don't we collaborate on the book?" I was like, "You're already a famous author."

He said, "No, this is a really great idea. We should write the book." And so we ended up with four of us and we wrote that book. We acknowledged Michael.

We, you know, we never claimed to invent the idea. We just, you know, but uh that book ended up being very popular. Not just because of the topic, but because I have a wry sense of humor. So, I did all the comics and made everything I could interesting to read by making it funny because there's a lot of stories in there about failure, which who doesn't like reading about it was like a early Dilbert.

In fact, we licensed two Dilbert comics in that book and this is before Dilbert was even that big of a deal. So, the book was surprisingly successful and uh that's how that got started and you know it was right. Um the project I was running ended up I ran a conference I wrote a book and next thing you know I got promoted at MITRE and you know it's another thing just follow the find the formula and follow it. Yeah.

And so like AntiPatterns is about right recurring failures or patterns of failure in software development. It's the things that seem like the right answer but actually it's it's not the right answer. You have to refactor it that if you do this you're going to end up worse than when you started. And the refactored solution is what makes an anti-pattern.

It's like we're not just complaining. It's like change your change how you're doing it to get to the good outcome. And that's that's the refactored solution. That's an important part of the anti-pattern.

Jay Combs: So with refactoring a solution or an idea that maybe like you said the bad consequences outweigh the benefits. Now that you're in data science, in AI, are there lessons there from the failures in software development or refactoring software that apply now with AI or is it is it exponentially more dangerous now with AI or what are some of the lessons that you can share from your current work?

Skip McCormick: I don't know if I would say exponential, but it might be. There's a bunch of new AntiPatterns that are obvious, you know, when you look at it. And I've I've um I'm too lazy to write that book right now, but I think I think there's a book there like, you know, AI AntiPatterns, right? Here's what you think you should do, but here's what will happen.

I don't know if we're ready to write the book because I don't know all the refactored solutions yet, right? Part of it is not just saying this is leads to a bad result, but it's only helpful if you can steer them towards a better result by refactoring, but you I do see it a lot. And it's funny when I got to um Bank of New York Mellon, I remember um going to some of the engineering meetings and discussions and in the meetings they were talking about AntiPatterns, right? They're like we need to avoid this anti-pattern that anti-pattern.

And I remember going, "Wow, this is so cool." They are talking about anti patterns as well, but I don't think all the people worked there knew I was one of the co-authors of the original book. And so I had books on my shelf, including that book. And one of the other sort a real Java champion, a high-end developer, really skilled, came by and he's talking at my desk, and he pointed to my book and he said, "Oh, that's my favorite book."

And I thought he was just kissing me up, right? He's like, "Oh, he's telling me he loves my book." And I was I was like, "Really?" You know, "What do you like about it?"

And he and he's saying, "It's just so practical." Whatever it was he said. And I said, "Well, thanks." And he said, "What do you mean, thanks?"

And then he looked again at the authors. He went, "Wait a minute. That's you?" And I said, "Well, yeah."

And he was like, "What?" Right. So, he didn't know that I was one of the co-authors. and he was being honest that book and I found a lot of software developers really uh got a lot out of the AntiPatterns book especially the first one because it sort of spoke to them right they're like yeah that's what I think of doing and then they read the anti pattern they go oh I need to do a little bit differently so

Jay Combs: yep and the anti anti-patterns.com is still up and uh your name on there is Hays McCormick yes the third and the picture of you on there is pretty spectacular lot younger. Yeah. Anybody listening, I suggest you go check it out. Mr.

Pointy Finger. Yes. Speaking of, how did you go from Hays to Skip? That's a good question.

My mother gave me that nickname as an infant in the crib when my grandmother started calling me brother. My I was in this little sailor suit. I don't remember it. I was only a few months old.

But uh my grandmother started calling me brother. My mother said she had this flash that oh if he's called brother and he gets a sibling, he's going to become Bubba. and I don't want a son named Bubba. So she looked at me, I had a sailor suit on and she said, "His name is Skipper."

And so that was when I got the name. Had nothing to do with Hayes. It was just a defensive maneuver to prevent me from being Bubba. You definitely had a destiny for the ocean lifestyle in the Navy for sure.

But you thinking about like refactoring, you've refactored your careers. You know uh Naval Academy

Skip McCormick: several times. Yeah. private business, CIA, back to uh private business.

Jay Combs: What drove you want to go into the CIA after 9/11? And then after 17 years there, what made you decide, hey, it's time to leave. It's time to go back to private business.

Skip McCormick: Well, CIA was another huge opportunity. I was at MITRE. I supported the agency as a contractor and then when I left MITRE and went out on my own as a consultant, they kept me as a consultant and so I worked as a consultant and CIA was one of my clients and um you know I think I was helping them transition from C to Java and doing some other things like that and uh then 9/11 happened and um some friends of mine and I we sat down and we put together a proposal for some things that agency could do to help to help, you know, with that war on terror. And I put together a slide deck and met with u my COTR, that's the government speak for contracting officers technical representative, I think.

And I gave him the slide deck and I said, "Here's something I think we could do." And he looked and he said, "You can't do that as a contract. You'd have to be an actual intelligence officer." And I was like, "Oh."

So I said, "Okay." And he's like, "Okay, what do you mean?" I said, "Well, I'm this is important. We're at war.

I'm I'm willing to do that. And he said, "Intelligence officers don't make as much money as contractors, right? You'll be taking a big cut in pay." And I was like, "Okay, you know, I can handle that."

You know, who didn't want to have a chance to help, right, when that happened? You know, you're I don't know how you remember how you felt, but you're really angry and you want to do something to protect, you know, your loved ones and to sort of get back at the attackers. And that was what I had in mind. And so he said, "If you're serious, you know, we I'm hiring a lot right now."

And I said, "I'm serious." So we left from that lunch meeting and went right downstairs to the HR office and I applied right there. And uh you know, sort of that spur of the moment thing. That was I want to say late September, early November.

It's pretty close after 9/11. And then you know there's a lot of vetting that goes on. So, I didn't start until April, but uh you know, I got hired and I still April 2002. Yeah.

I still remember Yeah. I still remember sitting in to in the hiring session at headquarters where you take the oath and they give you your blue badge. And it was such a proud moment. It was just awesome.

I was like, "Wow." Because if you work there, you get an appreciation for the amazing caliber of skill there that some of the most talented people I've ever known work there and they get they get very little recognition um don't want it um don't want to be uh known um you know because they're causing harm to bad people who might try to harm them back, right? So there's it's a very secret business for good reason. Um I was lucky that somebody needs to be um overt who can talk to industry and um you know you know make that bridge happen.

And so um you know there's there's a risk doing that but it was good. And so you know for the next 17 years I did we called it big data analytics. We were building the first supercomputers that the agency designed for text right not for you know numerical you know processing and um combining data from systems that couldn't be combined before because they were already maxed out and couldn't get any bigger and uh you know this system put things together that we couldn't see before. It's like when when you're fighting an enemy that doesn't believe you can see them and you can see them, it's it's the best advantage there is, right?

It's like, how did you know of that, you know? And when we could we could put people on the map and say, "Here's where they're going to be and then the real heroes who put their life on the line, you know, go there and meet them and get them." When they call you back and say, "Yeah, that's where they were. We got them."

That's the best feeling in the world, right? There's a lot of there's a lot of angst in it while you're waiting. But um that was awesome. And then then I got to follow the software we built into the counterterrorism center itself and I felt like I was part of the most elite team ever.

It was a wonderful place to work and uh immense responsibility. You feel like you can never do enough.

Jay Combs: So it's high consequence, right? You're dealing with big data analytics. High consequence something goes right or something goes wrong. And I imagine you're dealing with irreversible consequences, intelligent, irreversible consequences, but you're deal you're you're trying to put together probably imperfect data, ambiguous information.

And I know in the tech world where you know consumer tech or things where people are saying, hey, move quickly, break things, right? And there's validity to that. It's hey, let's let's not wait for perfection. Let's keep moving, right?

But how do you manage especially in high consequence environments like health care, financial services, the intelligence community? How do you balance managing uncertainty and managing perfection? Trying to move quickly but making sure that you're giving the right difference to those high consequences. How do you manage those two different things or balance them?

Skip McCormick: It's a tough one and I had some excellent mentors. There's a gentleman named Jeff Jonas who's the CEO of a company called Senzing at the time. He had a company out of Las Vegas that was helping casinos catch cheaters and um In-Q-Tel invested in him and we leveraged his technology. But his huge contribution was teaching me about entity disambiguation and how can you figure out who's who in a world where they're trying to fool you and they're trying not to be disambiguated and the data has mistakes in it.

In fact that I just was talking about this morning. If your data hygiene takes out all the errors before you put it into your correlation system, you might be removing the signal, right? The old garbage in garbage out. Sometimes the garbage is the signal and it in the cases where people are committing fraud, they have patterns for their fraud.

They'll change threes to sixes or fives to, you know, it's like things they can do that if they're caught, they can explain it away as oh I'm dyslexic or oh that I didn't write that clearly because they, you know, but if you fix those, if you go this zip code doesn't go with this um city and state and you just fix it the classic data hygiene, you lose that signal. And so Jeff taught you never change the source data, you annotate it. And that led to something we call dynamic metadata where you take the data as is. You preserve it as-is, and then when you do your hygiene, you add more data.

You say, I think the zip code is this, right? And you know when you thought that and why, but you still have the original one, right? These are really simple examples of more complicated things. So that ended up being very important, but um we needed supercomputers to deal with the amount of information.

And at some point we had so much metadata that um we didn't have room for all the data, right? And so um you know you'd have like I don't know identifiers like phone numbers or something. And every phone number had a bunch of data about where it came from, what your uh responsibilities are for protecting it, um what its sensitivity is, what its uh you know confidence is, all this other metadata that was essential for making some kind of a conclusion. And we would add that to it because you can never separate the metadata from the data and then be confident in your answers.

And everything you answer, you have to be able to walk the dog backwards to why you thought that. And so we said we need to get the metadata out of the supercomputer, but we need to preserve it. So we created a concept we called dynamic metadata where we put one identifier on every data object and that identifier went into a separate system a big graph where we could add any annotations that were needed including ongoing confidence or relevance or corrections or all kinds of things. That concept I brought with me to the into banking and it helped a lot.

And I remember talking to my boss at BNY Mellon, Joe Sakowski, about their metadata solutions and he said, "We got metadata." I said, "No, you have static metadata, which is essential, but you don't have dynamic metadata." He's like, "What in the world?" And I told him, I said, "Look, at the agency, we would steal data from some other country, right?

That's what they do." And then we would bring it in and you're looking at it and you'd say, "These identifiers are supposed to be unique, but they're not. We can't go back to that country's leader and say, "Your unique identifiers aren't unique. Can you please fix your data so we can steal it again?"

You just have to deal with it, right? But the funny thing is sometimes when you find identifiers that aren't unique and a system creates them uniquely, that's a big signal. It's like, "Oh, somebody added this identifier probably for some intelligence purpose, right? Here's a new document that has the same number as the old document, but it's modified in a certain way, right?

that has intelligence val may have intelligence value and if you right so you have to add that information then have the ability for other experts when they see it to add more information and so when you have some final result you can unwrap it all the way back to the source elements that whole concept came from Jeff Jonas right and Jeff Jeff I think did that because he was doing business for casinos and they didn't they weren't very forgiving if he got it wrong so he had to be able to explain to them yeah his answer here's why we kicked this guy from your casino. He isn't just your brother-in-law, he's also cheating, right? It's like what? Right.

Stuff like that. So, these were important things that I think we're taking them forward at my company with an application we call agentic metadata, which is the same baseline. We can add more metadata dynamically to any to anything. And now we want to actually have the AI be able to add metadata as it goes.

So, we want the AI to be able to ask itself questions like here's my conclusion. Um, how do I test if it's a hallucination, right? Can you build into the AI this the scientific method to say what would disprove this and can I find evidence that would validate the disproof and so that I know it's an hallucination. What I'm trying to do is take the hallucination problem and turn it upside down.

There's a Dr. Philip Sean I worked with who said it's it's a chance to make artificial creativity. If you can take hallucinations and say actually we want to we want to double down and have the AI create things that no one would have created and then test them scientifically right maybe it can it you know and there's some things like you know protein folding and stuff like that where we're seeing those kinds of things but can we see it in other areas so all of these things just give you a foundation to work from and back then it wasn't called data science it was just called big data and we were just nerds you know trying to help the real you know operators. Yeah.

Jay Combs: And it's I mean the common thread from that story to me seems like twofold. One transparency into when you're making changes annotate them so that you can go you can trace everything back. You're not getting like this modified data set that's going to give you something totally different. It's how did you get here?

What did you because you're basically documenting your assumptions as you go and being able to prove like this changed here. If somebody audited you or looked back at it, why did you do this? Well, there here are the reasons. We can go all the way back to the original data set.

Yeah.

Skip McCormick: When we made this, there was uncertainty and this is our assumptions and they say, but that was wrong. So, we know that now, right? But back then, this was what we thought was true. And another reasonable person at that point would have agreed.

In fact, they did. Here's their annotations.

Jay Combs: With your work with Angularis, how important is this transparency and traceability, especially and I know essentially important. Yeah. And especially I know. No, go please.

Yeah. No, no. I'll let you finish the question. No, I was just going to kind of tie it to kind of the regulatory uh environment at the moment.

Obviously a federal level it's very uh interesting at the moment but at the at the state level um you have very different states putting forward uh pretty um comprehensive regulation around AI. You have the Colorado Consumer Act, the purple state, California just passed a bunch last year including AB 2013 which is about generative AI transparency around training data and things like that. And then Texas probably has the most comprehensive at the moment and uh Governor Abbott just signed it four days ago, the Texas Responsible AI governance act which is somewhat modeled on the EU AI act. So yeah, I'm curious that Angularis transparency traceability and now with varying states of different, you know, political um places on the spectrum, they're all focused on this.

So yeah, how important is traceability and transparency in what you're doing at Angularis?

Skip McCormick: It's essential, but it's not enough. We're pushing now on a sort of a double click into that. There's a team we're working with that uh may invest in us and they've said what's more meaningful, right? Our Angularis sort of story is about we can help you make your model uh risk management and governance um reliable, dependable, help you pass your audits well.

Um make sure your models are continually performing to spec. And we've added metrics for difficult to measure things beyond just performance. You know, precision, recall, Brier score. There's metrics like is it is it explainable?

Is it transparent? How can you measure those things? That's that's very hard by the way. And so we're trying to create a um not for-profit group we call excellence in AI.

Exai. And uh I don't know Elon Musk might have EXAI. We won't be able to keep that. But um anything with an X in it.

Anything with an X. But we need there needs to be somebody that's not us that defines what ethical means so that we can build measurements that measure it because an ethic is just just a metric to the good. But it's the question is to who's good. And this is where the double click came in is if you're just help helping the bank be ethical, comply with rules, pass their audit.

That's a good thing. It saves them money, helps them deliver AI, makes them more productive, all all kind of good things there. But who's looking out for the consumer? Who's looking out for the client?

Well, a lot of these laws, especially in banking laws, say you have to provide notice. The client has to agree. You know, there's all these things you have to sign. They have to know the API.

they have to know all these different things. But the truth of the matter is most normal people just aren't that aware of the financial details. You know, you're taking out a loan and um I don't know if you've done this lately, but you sign a lot of things that even if you sat to read them, you're not really sure. There's a certain amount element of trust.

You say, "I've been working with this with this team. The bankers are good. Um you know, they're they're audited, all these sort of things. They're not going to have me sign something that hurts me.

There's a trust there. But inside, you're saying, I hope I'm not signing away something that's going to be important to me later, right? And so, why not have the AI also look out for the consumer's interest, too? The AI can look at the problem from both perspectives, right?

From three really. The financial institution's perspective to say, is this legal, ethical, compliant, all that kind of stuff. the auditing group's uh perspective of um can we validate that they're doing everything right? So if we give them credentials uh we're not responsible for failure.

But how about the consumer says this thing I'm agreeing to is actually the best fiduciary option for me. Right? The AI has said I'm getting the best deal. Right?

AI can make sure everybody's doing things the way they should in an ethical way and to everybody's benefit. Uh even if you have somebody who's who maybe puts profit ahead of other concerns, the AI won't let that get out of control, right? There's there's that's kind of where we're trying to go now. So Angularis started as a way to provide a managed service that I wish had been available to me at the bank when I was doing this.

If there had been a company like Angularis, it would have been a lot easier for me to do my job. I would have just hired them and said, "Uh, build our model inventory, build our metrics, track it, show me, you know, track our whole life cycle, all those things." Um, you wouldn't have had to build our own. We could have just done that.

Um, didn't exist, right? When I left the bank said, you know what, that should exist. Let's make that. I think we can sell it.

But now we're looking at saying that's great and it should be a profitable business, but we can actually do something that's not just helpful but meaningful. And that's that's where we're thinking right now. And uh it's a little bit fuzzy, but I like the idea of can we instead of being afraid of the AI, can we make the AI our tool to help us be less afraid of each other? Like the AI is looking out for all of us.

And um to do that, you asked about how important is transparency, explainability. I think that's the key, right? Is when it makes a decision, can I open up that decision and see how did we get to this decision? I want to be able to do that whenever there's a decision enough till I get to the point where I say I'm confident that this decision is going to be like the other ones I double clicked on and follow up.

Does that make sense?

Jay Combs: Yeah, I think so. I mean, you've always been very uh forward-looking and but even when you were at BNY Mellon, you won a platinum innovation award and I know you talked about, hey, I wish we had, you know, kind of what I'm doing now, which is probably one of the reasons you kind of started Angularis, but BNY Mellon, it's about 240 year old institution.

Skip McCormick: Started by Alexander Hamilton.

Jay Combs: That's right. And so what was the secret or how did you drive innovation there um at a company or an organization that has centuries of hey this is the way we've always done things. This is how we do it. How do you bring in new innovation and push them to change?

Skip McCormick: I think part of it is it's a hard thing right and um there was a book I read years ago by I'll think of his name in a minute but it's called dealing with Darwin. I can get you the author's name, but he has this graph in there that I've used many times where he shows if you're trying to optimize your IT investment for the best value, right? You get the most for the least amount of money.

You opt the curve has a like a peak over here like in um value to money. You get the most out of it. If you're trying to optimize for the best innovation, it's a different peak, right? It's a different place because you're to innovate, you have to take risk.

You have to do things that don't pan out, right? And so companies want to be innovative, but they also want to be efficient and they have this sort of these two driving forces. And um um the author put those on the same graph and it showed that the two barely intersect, right? this is peak efficiency and this is peak innovation.

But there's a little area here and too many organizations go we're going to focus on the optimal for both which is really suboptimal for both right you're not very efficient and you're not very innovative and that's doom right so what you have to do is have sort of have two different programs one that's the goal is to innovate and there's things you have to do there to keep it from killing you from using up all your you know resources and then you have to have another group that's about innovation And the team that's working on this needs to allow both of those two things to optimize themselves. But the goal of the innovation side needs to be to get things into the into the optimization side as soon as soon as they can. That was the point of his book. And I remember showing that diagram many times when I was in that fight and they say, "This isn't how we should do it.

You know, what you're proposing isn't the way this is done." And I was like, "Yeah, but this is a better way to do it. We just don't know that yet, right? And so you it would it would allow you to carve some resources out to do it.

And at a bank like BNY Mellon where um there's literally trillions of dollars a day being transacted, uh the efficiency and the reliability and the sort of perfect operations is just it's not debatable. It just has to be perfect, right? And they're very good at that. Any of the major banks have had to perfect that.

I showed up and I'm working in the AI side of things. Um to get sort of credibility and get some resources and be allowed to innovate, you sort of have to help with the this side for a while to where they say, "Okay, we believe that you understand our mission and that you're contributing to it and you're not just sort of a mad scientist over here doing something that doesn't matter, right?" And then you wait for the opportunity where you see a problem they're wrestling with over here and you and you go I can think of a way to solve that problem and then you go solve the problem and come back say does this help right if you get a couple of those little wins pretty soon they start coming to you for solutions right and there's that and then the other thing is dealing with the regulators the regulatory constraint on especially on large banks is a is a large burden for good reason, right? And there's there's a lot of um fear is a word you could use.

We don't like to talk about being afraid of the regulators, but they have a lot of power, right? And so when you go into those those audits, you really want to pass, right? You want to come out of it with no findings, which doesn't happen often because the auditors are rewarded by finding problems, right? And so they're super motivated to find them.

and uh um all of that context is sort of the same thing when you um can set up a partnership with the auditors just like setting up a partnership with the optimizers where the auditors see that you care about the thing that they care about as much as they care about it right and you can say my motivations for caring about this goes beyond I just want to pass your audit my motivation is because the thing you're auditing for is important and I want the same outcome you're after the reason for your regulations is important just as important to me. If you can establish that trust with the auditors, you actually can do much better. It's like it's like in the audit there's nothing where we're where it's open kimono. Here's here's this honest situation.

Here's where we where we're having a problem and things aren't as good as we want. Here's where we're doing something uh novel. Here's where we're doing something risky and here's our mitigations to manage the risk. If you can establish that relationship, um, you get a lot more latitude and you know you've established it when those auditors call you back later to ask question about somebody else they're auditing where they're saying we're we're looking at somebody else.

We can't tell you who, but um they're telling us this, is that reasonable? When you get that kind of uh relationship going, then you have a partnership going and I think that's the most productive situation. And so, uh, Bridget, who is the CIO at the bank at the time, now she's at Wells Fargo, when she recognized me for that award, I think it wasn't as much about the actual innovations I accomplish. I think it was about the partnerships that were fostered that caused things that caused things to um to improve and we did some we did some cool things that nobody had done.

We, you know, we had a lot of fun, right? and work hard. But um but that recognition was about that. And it's an important thing to do to recognize that because your innovators sometimes are viewed as disruptors and can be a pain in the butt.

I know I was to some people because I'm very impatient. An award like that gives you a little bit of protection, right? when you show up and they say, "I'm actually here to help you and I've helped others and they look, they gave me this fancy award and I have this cool little plaque, you know, whatever." It's it adds to your credibility.

So, takes a while to get there. It took took a few years of sort of um taking on impossible challenges and making enough headway to make an impact. Does that make sense? It's a long long answer.

I mean it's so I mean it's more of a leadership and process award about how you went about doing that and doing it the right way and obviously innovation like you're saying it's risky. It's not a linear progress. It's all over the place. You know we talk about oh you need to be disruptive.

It's like no disruption is disruptive. You know it's like revolution is revolting right? It's like no uh those can be consequences but you can mitigate those right? you can you can do it in a in a in a way that's less um expensive and doesn't uh use up your personal capital, you know, fighting.

Jay Combs: Well, tell me about like you said, you maybe you were a pain in the butt to some people, but tell me about a uh a career decision that looked like maybe a mistake at the time, but it turned out to be formative or you learned from it. Is there something you look back on, hey, that was a mistake, but at least I learned something from it.

Skip McCormick: Well, I've been told by a few people who became good friends over time that when they first worked with me, they thought I was a bully. And that was that was hard because I don't want to be a bully. You know, I don't like bullies. And I remember when they told me, I was like, "What?

Oh, no. You know what? You know, why did you think I was a bully?" And um it turns out it was when I was trying to get people to see things my way.

Sometimes I didn't leave a lot of room for them to rebut it. And um that can be hard to do, especially when when you're in a hurry and it's like, "Okay, we've looked at this every way. I think we're done debating. Let's just do this."

And um um finding the right time to sort of take over the whiteboard marker and say, "Here's how we're going to do it." can be hard. And there's some people who aren't where you are yet who when you cut them off, they feel bullied, right? And I think this goes back to leadership versus manipulation.

If they feel manipulated, it wasn't very good leadership. And uh there's been a few times where I've done things like that. And looking back on it, I'm I'm kind of ashamed of the way I handled myself. you know, it's it's like uh I wish I could go back and do it better, but I have to remind myself that I was a stupider, younger, dumber person then and um would do it differently today.

But uh those are the failures and the risks and most of the failures that I regret are things that um discouraged somebody else. That's what I hate to do. So yeah, the interpersonal stuff is challenging and I think like good leaders give space for debate for dissent to do it um in the right way and like always trying to get to the right answer and I think when you're very mission-driven person sometimes you can get that tunnel vision and be like this is we got to go here and it can be hard to kind of look around a little bit. uh but the self-awareness to be able to reflect on that and say hey how can I be a better leader and you know merge these things together because not everybody has the same personality and I think that interpersonal stuff and you know I worry about that in the age of AI and I bet kind of why I asked you that question early on in the meeting of how do you collaborate and talk to folks because those soft skills are still so critical and will be so critical I don't think that's ever going to change

the soft skills are the hardest skills and uh You know, I read a lot. I'm constantly, in fact, one of my interview questions that I ask people, how many books are you reading right now? Because I'm looking for people who have like 10 or 15 books open that they're reading at a time because they have things they need to know and there's things they're interested in and uh um and it's kind of an ADHD problem that could be a huge advantage if you if you apply it. Um, since I read a lot, um, I can recommend a lot of books.

And so when I'm talking to somebody and I'm trying to get them to think a particular way or think about a solution with an open mind, a lot of times I'll say, you know what, um, I got this idea from this book, this chapter in this book. Uh, uh, there's a great book, uh, Digital Transformation by Tom Siebel, right? It's it's a wonderful book. um he has does a great job of explaining the whole transformation, but there's like two chapters in there that are solid gold.

There's like the 10 things to for the C CEO to know and uh the architecture the AI architecture I think I bought if Tom's a bestseller is because I bought so many of his books and given him right and I'll say I have to yeah yeah read this chapter and then and then let's talk about the decision. So that's a that's a technique I've used a lot and I think that came from AntiPatterns because some of the feedback from AntiPatterns. Um one young man came to me he said my company was doing something really stupid and I could see it was stupid but my boss wasn't having it right.

Didn't want to hear about it. So I went in one day with your book and I said um I think we're doing this anti pattern right? And he said what? and he said and it since it was in a book written by somebody else it had just enough of credibility where he could read it and um he read it and he saw the where it would head right he said oh this I don't want that outcome and um it fixed the problem they've made some changes and so a book by somebody other than you can be very powerful tool to help somebody because otherwise there's this personal thing of like it's my brain versus your brain and when you say this isn't even my idea I got it from Siebel's book right here, right?

Um that helps a lot. It softens the sort of interpersonal battle. It can just take it off the table. It's like this isn't about me versus you.

This is about us trying to find the right answer. And uh I'm pretty confident I'm on the right answer and here's why. But I could be wrong. See what you know there's that.

Jay Combs: And that's I think Stu Bailey did that Idiot's Guide to model operations. Yeah.

Skip McCormick: Model operation which is Yeah. From Wiley. Yep. Yeah.

that book was so powerful to me because I was trying to get the bank to make some pretty big changes to how they did model risk management because we needed to make it go faster. You know, they were dealing with statistical models that they could have, you know, six guys with, you know, PhDs with more degrees than a thermometer and give them six months to study it and make sure it's all perfect. And we're saying, well, if you get get this model approved, we're going to retrain and it's going to be different in a week. And they're like, "We can't review a model every week.

It's like we need to change the way we do things." And it was like it's like, "Skip, you just don't know how banking works." And I was like, "That may be true, but I can see that this isn't but with Stu's book, I could say, you know, you're probably right. Um, take a look at I think it's chapter nine in Stu's book and see what you think."

And it would it would be amazing. They would come back and they say, "Okay, do what you're going to do. It makes sense, right?" And so, yeah, you know, that's that's one of the

Jay Combs: Stu's huge meaningful contributions to this whole space.

Skip McCormick: I'll be forever grateful for that book. And it's it's what 100 pages. It's not even a difficult read. Yeah.

No, there's a lot of great info in there.

Jay Combs: I mean, you can have all the technology in the world, brightest minds, but unless you can kind of sometimes put ego aside and convince people and advocate for these ideas in a way that puts ego aside and makes everybody kind of come along for the ride like partner with them as you would say. That's how you get to a good decision. That's how you actually get good decisions.

Skip McCormick: I've even Einstein would talk about Fermi thinks this rather than saying I'm Einstein. Shut up and do what I say. I bet he even said, "Yeah, I have my opinions, but it's based upon Oppenheimer said this or Fermi said that or whatever." That's, you know, that's that's kind of I think that's a good leadership technique is Yeah.

Jay Combs: Like if the book doesn't exist, go ahead and write it. Yeah. The psychology of good decisions. That'd be that'd be a good

Skip McCormick: funny story. When we set up to do AntiPatterns, we knew we had to read every book on the topic. There was a list of maybe 40 books that were relevant to what we were daring to write. And so the four of us set up a Thursday evening beers and darts at a local bar near MITRE where we would meet and every week

Jay Combs: where was it? Was this up in Boston like at Hanscom Air Force Base or was this

Skip McCormick: It was in the DC area. MITRE's Washington. DC. You're down to DC.

Okay. And so we set up this thing and we just it we sort of like a book club, right? We said, "We have to read these books, and there's more books to read than all of us can read together. So, let's read them to each other."

So, so we would say, "You read this book. You read you read the first few chapters of this book." And we just sort of test them out. And then we would show up, order drinks, and play darts while each of us told each other what they had read.

And so, over the course of about a year, we went through all these books and discussed them with a there was I want to say six or eight of us. Well, maybe more. The four authors and six or eight others that all became pretty good friends. Uh, fast forward a few years later, AntiPatterns was out.

It was successful. And we did two more AntiPatterns books. And almost everybody else who was part of that book club is also a published author now. And it was like, wow.

It was like, uh, if you want to write books, you got to read them. And then you got to believe that you could write something that people would want to read. And, and that was, I mean, I think they watched us. They go, "These guys are stupid and they got a book.

We could, you know, Skip can do it. Oh my god." Yeah. So that's amazing.

So I believe I believe in collaboration. In fact, we also met Saturday mornings to uh every Saturday morning to work on whatever chapters we're working on and hand them off. And I remember a lot of times working all night Friday night because I had something that was due and I didn't want to disappoint my co-authors and be the laggard. If I had tried to write that book by myself, it would never have gotten written if even if it could have been written because I would have I would have given myself permission to sleep instead of working all night.

So collaboration.

Jay Combs: Yeah, I think that's one of kind of the core challenges now in especially in the tech world of like collaboration, bouncing those ideas in a positive way with each other. And when you're over Zoom or

Skip McCormick: your reputation on a on a promised result to somebody else whose opinion of view you care about that's important. Yeah. I think that's that comes from military leadership. You know, there was there's a lot of that even as a plebe at the Naval Academy that you could never bilge your classmate.

You know, if you knew the answer and your classmate got it wrong and they asked you the answer, um you weren't supposed to give the right answer. You would ask for permission not to bilge your classmate because because you don't want to make your peers um look bad if you can avoid it. You know, those are important leadership and team building skills. You know the academy experience is about learning how to be a team in a difficult environment and they make a decision.

Yeah.

Jay Combs: No, they do. And I think that like but being able to operate under those you know high pressure ambiguous situations like once you can do that and then when you come to you know less stressful everything just it slows down it's easier. Um, but what do you hope people say in 10 years about the decisions now that you're making with uh Angularis and what you're doing in this kind of transformational AI period? What do you hope people are going to say about what you're doing right now 10 years from now?

Skip McCormick: That it was helpful, you know, useful. That's kind of the goal. It's like, did it matter?

Jay Combs: Well, I got one final question for you. But when you're making a good decision or celebrating a good decision or processing a difficult one, is there a uh go-to drink or recipe or food that you go to celebrate with others?

Skip McCormick: That's a dangerous question for me because I've struggled with weight my whole life. I enjoy eating. My favorite uh go-to food is fish and chips and my wife doesn't eat seafood, so we only ever get it when we go out. But, uh, to go with it, I like I like a Guinness.

And after my cancer, I can't drink alcohol anymore. So, I've switched to the no alcohol Guinness. And, and I vouch for it. It tastes like Guinness, and it's just as satisfying.

Goes great with fish and chips, especially with mushy peas. So, I recommend it. My mother's English, so I grew up uh eating fish and chips and with lots of vinegar. So, that's that's my go-to meal, but there's there's not many foods I would turn my nose up at.

It's a challenge. Um, if you like food as much as I do, it's a challenge not to overindulge. Yeah.

Jay Combs: Well, I love fish and chips and I know in Annapolis, I was actually recently there. My sister lives in Virginia. Went to Annapolis. There's some good fish and chips places down there.

So, appreciate the uh the food guidance. Um, that's it. Thank you so much for your time. Uh, we covered a lot of ground and you've clearly made a lot of good decisions in your life and career.

Uh, so thank you for sharing it and joining us today.

Skip McCormick: I enjoyed it. It was good.