Become a Creator today!Start creating today - Share your story with the world!
Start for free
00:00:00
00:00:01
Fast AI and fragile enterprises: Phil Fersht with Gary Hoberman
198 Plays5 days ago
In the latest episode of From the Horse’s Mouth, Phil Fersht sits down with Gary Hoberman, Founder and CEO of Unqork, to break down what’s really happening below the AI hype, and why the race to scale AI might just be deepening enterprises’ technical debt.
Enterprises are discovering that speed alone isn’t the advantage they thought it was. Security, reliability, repeatability, and explainability are becoming the real bottlenecks for AI adoption. It’s exposing a divide between what can be built quickly, and what can be trusted at scale.
What you’ll hear in 30 minutes
- Why AI is accelerating code creation, but enhancing complexity
- The hidden risks of low-code, AI-assisted development, and code replication
- Why security, reliability, repeatability, and explainability matter more than speed
- Why control, not velocity, will define the next phase of enterprise transformation
Guest Snapshot
Gary Hoberman is the Founder and CEO of Unqork, a no-code enterprise application platform used across key industries like financial services, insurance, healthcare, and the public sector. He is a former CIO and technology leader at firms like Citi and MetLife, and has spent more than three decades building and scaling mission-critical systems in highly regulated environments.
Explore More
Gary Hoberman on LinkedIn: https://www.linkedin.com/in/gary-hoberman/
Phil Fersht on LinkedIn: https://www.linkedin.com/in/pfersht/
More from HFS: https://www.hfsresearch.com
More from From the Horse's Mouth: https://horsesmouthpodcast.com/
Recommended
Transcript
Introduction to Gary Hoberman and Uncork
00:00:00
Speaker
Hey, welcome to the latest edition of From the Horse's Mouth podcast. I'm your host, Phil First. And joining me today is a very exciting character in the field of low-code, no-code technology, Gary Hoberman, who is the CEO and founder of Uncork. Welcome, Gary.
00:00:19
Speaker
Oh, thank you for having me here, Phil. This is an honor to be on with you. So not everybody who is joining today is going to be that familiar with Uncock. A lot of people will know of you, but maybe not know exactly where you're differentiating in the marketplace. So why don't we start off it with what you guys are all about, how you founded the firm and what it is you're trying to do in this market.
Gary's Career Journey and the Birth of Uncork
00:00:41
Speaker
yeah I mean, I'm going to, if you could do the wavy lines, felt like we'll pretend we're going way back. So I don't know where you were in 1994, but I was a kid out of college, Stern School of Business in New York and joined a big company called Bankers Trust. And it was, you know, still not around anymore. Right. It's been amazing myself. And i was an engineer.
00:01:03
Speaker
Since fifth grade, I communicated with machines, starting in punch cards and Timex Sinclairs and Vic-20s. And I love this idea that you could communicate and it listens and communicate well, communicate poorly, you know, similar to people. When I started in Bankers Trust, they assigned me a job. They said, go build a money transfer system. It's, you know, and Sybase.
00:01:26
Speaker
the days of ODBC drivers, for those who remember, I did it very quickly and I suddenly realized that was kind of boring and if I do it well, they're gonna ask me to do another one, like life is, of course. And so instead of building the one screen, I built a code generator. This is a code generator which you just populated a Sybase table with field names and ah table names and it built the tables, the fields, the start for start procedures, the visual screens and C++. plus plus Within six months, 700 apps were built.
00:01:55
Speaker
And when you say built in those days, it meant burned on CDs and distributed to business users in the company. I accomplished the climbing Mount Everest. oh This was like, to me, it was like the coolest thing ever. I was nominated as one of the youngest assistant treasurers in the company's time. And I suddenly realized my calling and my calling was not to build that code generator.
00:02:15
Speaker
The calling was, i got a call back from a business user saying, we accidentally sent out two $14 million dollars wires and we don't know why duplicate. Could you go find out what happened in your system?
00:02:28
Speaker
And there was and a join incorrectly joined, the null condition. And, you know, what I realized was there was a bug in my code. It wasn't perfect. And that's okay. I'm going to fix it. I know how to fix it quickly.
00:02:39
Speaker
it was impossible to patch 700 apps without each time deploying a patch and breaking the other app.
Challenges and Evolution of No-Code at Uncork
00:02:47
Speaker
And you suddenly realize back then, over 30 years ago, realized that this concept of code generation, and it doesn't work.
00:02:55
Speaker
And reality is you replicate code, you replicate it. And replication is the thing that's holding back businesses. So I vowed 30 years ago to say, what can we do differently in software? And I went, you know, climbed up the ladder at Citigroup, managing director, when I always built another version of this using XML, still not perfect. went to MetLife as global CIO, built another version of this using, at that time it was JSON and it was pretty cool, data structures, and then left the corporate world in 2017 to create Uncork.
00:03:27
Speaker
And as you said, this category didn't exist when we created the company, low code, no code. When we named it no code, every research group said that's a marketing gimmick. Do you remember those days back in those days when the category did not exist? It was just called software.
00:03:42
Speaker
I remember the RPA thing came out and we were we were kind of the voice behind a lot of that. And that was in 2012 when we started to use the term low code a little bit, but then I didn't hear about no code until I heard about you guys.
00:03:56
Speaker
Yeah, we're still we're still the only ones, pure, pure no-code, let's say. And let me just explain the difference to your question. So specifically, if you think about code generation and writing code, you know, average Java developer does what? Eighth story sprints, eight you know points per sprint or every two weeks. And that's kind of the average for Java. You know, code itself, I like to think of like you're creating a house. You're standing up to two by fours.
00:04:22
Speaker
Before you do that, you'll agree with the business on blueprints. It's gonna take three years to build this house. It's gonna cost you $3 million. dollars It's gonna be great. It's gonna be everything you want at the end. Here's the kitchen, here's the furnishings, here's the... And you know what happens. You get into the house and nothing is what you wanted.
00:04:36
Speaker
And you're like, wait, wait, the light's not shining here. Can we move the window? The answer is no. Too hard, too expensive, too difficult. Can we change the finishing on the cabinets? Absolutely not. That will cost you $50,000. And that's the world of software.
00:04:49
Speaker
It doesn't matter Agile, Waterfall, that's just software. And so when you think about low code, low code was this concept of like a panelized home. Like we'll pre-build 70% of it in the factory and we'll ship you these walls and you put up the walls, but then when we get it all done, we'll still come in and we'll do the finishing touches.
00:05:06
Speaker
The problem with that home is the second it's built, you still have all the repairs. it's gonna be bastardized to your taste liking and it will never work. And so the concept of no code was very different. It was, what if we actually, by the way, of course there's code. First off the word no code was probably the wrong words we use because of course there's code. The code is behind the scenes. That code is locked in about 56 components And those components like a text box or a number or a panel, these components themselves contain code that our business and technology users never could see. They don't get access to.
00:05:46
Speaker
And the reason why that is, is we were able to make this honeycomb structure of components push the code down so no one sees it but our team. And what it basically means is instead of 700 apps generated that each one has their own unique problems and support and maintenance teams, basically becomes a single code base you maintain across every client, every user, which means we could provide the security, the liability, all the illities, everything you talk to your clients about as far as what's important to enterprise software.
00:06:18
Speaker
But we do it in one code base we centrally maintain which means all customers are scaling and growing together, which is very different. The second you let a user write code, it's creating a branch. It's creating a unique snowflake. When you want to upgrade the base, that house, it will break.
00:06:39
Speaker
And that's, that is the major difference between on cork vision and everyone else. it basically means every single customer using on cork is making every other customer more secure, more reliable, more efficient, you know, more compliant without ever thinking about that.
AI's Role in No-Code and Software Development
00:06:55
Speaker
And that was the vision I had that was back to 30 years ago. It was the major reason why we came down cork. How are you finding today's market based on everything that's happened in the last three or four years, the whole hype curve of AI, and now there's all the noise around Fyde coding, and you see films like Replit, Cursor around the place. Has that helped you guys, or has that created more confusion in the market for you?
00:07:24
Speaker
I mean, it's gotta be the most exciting time for a technologist any time in existence. This must have been like when the first mainframe came out. It has to be like at that level and the internet. And just for me, when we think about AI and what it means,
00:07:38
Speaker
We're looking at it very different from everyone else. Now, with on cork before AI, we were actually targeting and scoring 60 story points per developer per sprint, which is six to eight times faster than then cho a Java developer.
00:07:54
Speaker
So 600% better, not 30, 40%. So, you know, when you think about the old aspect, it's kind of crazy because sometimes I feel like a Morpheus. Are you a Matrix fan, Phil and all? Oh yeah. yeah So you could be Neo, I'll be Morpheus. There's just code all around us. We're creating more and more of it.
00:08:12
Speaker
AI-assisted code is just creating more and more code. Vibe coding is generating code. And the second you generate that code, when you want to change it and the AI isn't able to make the change or it doesn't have the tokens to rewrite it the correct way, you're going to have to understand the code.
00:08:30
Speaker
you to have to debug it. going have to actually jump in and on this. The big thing will be in this world of AI generated code, the hackers will look at this world and say, wow, you know what? If I could break into one single open source library that's been used across all these firms. there will be replication across the world in the millions that becomes an annuity for years that they get to do. For us, AI is exciting. we brought on Dave Ferrucci, who to in my mind is the top AI officer in the entire world who built Watson to beat Jeopardy. He joined us as our head of AI and CTO.
00:09:04
Speaker
And the first time he demoed what's currently in Alpha called Uncork AI. I didn't sleep for a week because what I finally saw was there it is a world where AI could not just replace the role of the developer, as we've been doing, we've been focusing on, and that's what Vibe Coding does, is it takes out the developer.
00:09:26
Speaker
But what we just deployed to Alpha actually knocks out the engineer as well, which means guardrails, practice, design, principles, how do you build reusable, all of those things that an engineer needs needs to think about. How do i actually use a service? And so for us, what we've been building and we deployed is this concept that you could actually communicate and just focus on your business requirements, your logic in English, nothing else.
00:09:54
Speaker
And what's different about that and vibe coding and different about that and ai bolted on top of low code is they're still replicating code. And the problem with replication and the survey we did with HFS, which was incredible. the Problem with replication is, you know, 93% of the CIOs we surveyed and CTOs, you know, said tech that is holding back their company. 93%. And like, when you think about those numbers and that stats,
00:10:23
Speaker
All you're doing with Vibe coding and AI assisted code is generating more code. So how do you generate cleaner code, better code? And that's the piece. So it is the most exciting time in our lives for sure.
00:10:36
Speaker
Do you think CIOs are readier than ever to rethink how they build up IT? I mean, i was talking with someone from one of the leading fast food restaurants last week.
00:10:48
Speaker
And he said, we are so beholden to SAP and Microsoft and Ariba, these types of tools. They just feel completely strangled by these applications. They can't get out of it. But then I'm talking with somebody in the retail business, and it's completely different mindset.
00:11:06
Speaker
And I realized in the retail space, if you don't completely pivot how you manage your technology, you're screwed. Whereas there's something like the fast food business, maybe you can afford to go a little slower. And the bottom line is it feels like some companies are moving very, very fast. Others are still very stuck in nine test of the 90s, the ERP world.
00:11:31
Speaker
You know, it's interesting. So my view of AI. So there's a lot of people that think AI is going to knock out 20, 30% of the workforce in each company, automation. And you're reading about that. What I personally see happening is AI is going to enable the best products and services to take more market share.
00:11:50
Speaker
And those products and services which are subpar will go out of business. it's gonna be a complete different cut in my mind of the market. Because, you know, we recently demoed Phil to ah a CEO of a very large financial institution. What we showed her was this ability that in Oncork, we built an onboarding application, institutional onboarding application, where it basically enables you to do the work of opening the account in maybe 30 minutes versus six months.
00:12:19
Speaker
And that's the difference today. And we used Gen AI inside of it to basically help us get this done. Her reaction was she turned pal. And I'm like looking, I'm like, wait, is there something? and she goes, you know, we don't have the best product and service tonight.
00:12:35
Speaker
And the only thing we're relying on is a very long switching cost. So that switching cost is what's protecting us staying in business. You just took out the switching cost. You just showed us that someone could move. So I think in the retail market, in fast food, i actually think it's going to be a different cut that we all experience. It's going to be a world of the best products win and those don't make it easier to switch to them.
00:13:01
Speaker
So they could go out of business. It could be a very different shift in the market that we've never seen before. Now with that, you know, Phil, like as a CIO, I went from banking to after about 20 years in banking, I went to be a CIO of an insurance company. And i will never forget, you know, visiting one of the businesses in claims and watching our technology that I now owned went down. 2,400 claims operators in this building basically stopped working for the next 30 minutes.
00:13:33
Speaker
And I asked, I said, how long is that? How many times has this happened? It was happening maybe 20 30 hours a month. And in banking, i could tell you the number of P1 outages I personally had in my software. It was like one because we deployed something on a Thursday. We didn't think we made a mistake. We're stupid. We deployed on Thursday and knocked out Tokyo trading.
00:13:53
Speaker
yeah but but like That was stupid. It didn't happen again. But when you think about it, like, I think that that whole market is about to change and legacy software as we know it,
00:14:04
Speaker
And I am more excited about taking over and refactoring mainframes and Java and.NET and COBOL and so we're doing this across the board and moving that to Uncork and shutting it down.
00:14:15
Speaker
Like it is so needed in this world to move forward. And, you know, it's the thing that's keeping all CIOs up at night. It's like, that's that's probably the most exciting thing.
00:14:25
Speaker
So you're seeing more opportunity where really old stuff and old mainframes and new systems, and then maybe even some of the ERP systems at this moment? but To us, ERP, you know, is like what,
Disruption of Traditional Systems by No-Code Solutions
00:14:41
Speaker
60, 70,000 tables and relational tables. And to me, that just means it should have been a NoSQL structure.
00:14:47
Speaker
It should have been designed in a different way. We recently competed in one of the, you know, in ah HR, you won't think of us as HR. And we've done a few deployments in HR, you know, review systems and, you know, 360 feedback, and we've done compensation, we've done consultant payroll, even in public sector.
00:15:04
Speaker
We've, you could build anything. So we've, we have customers who've used Uncorrect to build their payroll system. And, you know, we recently saw a um an RFP out there for what you would think of as a very large scale HR system where you would see the SAS players, the Workdays and the Oracles, PeopleSofts and all the big And we decided to see how far we'd get. so we actually, like, and it was interesting because we went through the process and I think we were in the third round when they said you need to provide this functionality.
00:15:35
Speaker
And it took us four days, one person to build from scratch everything they needed. And I think the idea that SaaS is not gonna be disrupted by AI is wrong.
00:15:47
Speaker
I actually truly believe we're about to see a world where AI enables you to describe your ah HR process custom to you. And the software is just going to form around you, secure, a compliant, reliable, without ever supporting it again or maintaining it again.
00:16:02
Speaker
And that's the maintenance piece is the huge thing. People, 80% of tech budgets are in maintenance. It's not in the bill that we keep talking about. That's interesting. Yeah, so you mentioned the study we did. One of the things that really hit home with me was the sheer volume of services spending, outstripping software spending.
00:16:24
Speaker
it's got two to seven times in certain instances. Tell me a bit more about how that's going to evolve, because I know you work with a lot of the big SIs and some of the big consulting shops. How do you see the services industry evolving around around sas and software as we look at the next couple of years?
00:16:42
Speaker
Well, you know, I absolutely agree with your concept as, you know, services and software are going to combine into one, which is could take the form of a outcome. You know, we we will create give you an ah HR system that meets every one of your needs or give you an institutional onboarding system. Included in that is the services component because we've gotten so good at doing that and the services might be delivered by Deloitte or UI or...
00:17:06
Speaker
any of the others. like mentioned Or maybe they're doing the opposite and bringing the customer the solution and all of the technology needed is just componentry that they assemble and build it together. But, you know, so when I think about the past, because I've lived through, I've done the largest, you know, Salesforce or ServiceNow deals for Citi and MetLife. I brought them in, actually sitting there with Frank Sloatman or Fred Luddy in those times, or you know we had, and the reality is like these the software is good for when it's designed, they all are.
00:17:40
Speaker
And then it created the services industry to customize it for your needs. So here's your box, you stay in, you're perfect in that box. It's amazing. When you wanna get out of the box and do more, there's this large scale services industry that was born around those products.
00:17:57
Speaker
And that is something which, you know, when you think about that that concept and that ability, ai is going to disrupt that. That's the piece, which is, you know, if you could actually just speak what you want to maintain it in English, what you want, um the roles of business and technology, you know, kind of blur a little bit and technology becomes architecture and governance controls compliance.
00:18:21
Speaker
And that's actually how it started back in the day. And the business becomes, I'm going to own my outcome. I'm going to own the actual end state. And that's going to be interesting to see. going to be exciting to see.
00:18:34
Speaker
Yeah, I talk about it's like changing the wheels of a car while driving. So there's a lot of multi-billion dollar services shops out there right now. And some of them desperately want to launch their own platform businesses.
00:18:49
Speaker
And HCL have got a $2 billion dollars platform business. there so now card ownd I think here some other big software firms, ah services firms are looking to do it because the valuations are so much higher.
00:19:01
Speaker
But it also tells you there's, I think, a conflict between ultimately companies like Salesforce and ServiceNow trying to get into the services space with their agentic platforms and all this type of stuff. And then service firms who generally think, like, these clients need more hands-on help. They can't just buy software and rely on very expensive consulting all the time. We're starting to see a real change happening with some services firms of all who are generally looking at sort of almost ring fencing emerging platform businesses and trying to brand them, trying to develop themselves.
00:19:39
Speaker
It's fascinating the areas that this market is evolving into and I think if we look at a year, year and a half, we're going to get a whole different set of winners in this space. Yeah, and you know, Phil, we get to see, we have 7,000 production apps across financial services, insurance, public sector, healthcare. care When we look at the apps, we actually get to see AI in use.
00:20:00
Speaker
We get to actually see it used in the applications, besides from what I was describing earlier, where it helps you create and maintain the software in use. It's real. If you think about, if you just minimize everything you read about AI and get rid of the photo generation, video generation, voice generation, all all the the creative aspect, what you get down to is it's a machine that could read a document and give you insights in the document. And just, if you stop there,
00:20:30
Speaker
That's probably 30% of the jobs in the enterprises today is just in that case. If you just look at how do we use this to basically take a claim in understand the claim, get the data from the claim we need to process. We no longer need to send this to an offshore to be indexed.
00:20:48
Speaker
We don't need to actually now store a PDF, have it indexed into a system keyed Like there is so much efficiency there and that we're seeing. If you just look at what it's good at, not trying to do not, you know, I'm not sure if you're seeing this, but the word agentic is overly used across the board. And I think every company has a budget set aside for agentic without knowing what it means.
00:21:11
Speaker
You know, if you just look at deterministic, like I'm going to give you from a former CIO's perspective, 98% of the business could have been automated, no operations with deterministic systems.
00:21:25
Speaker
It wasn't a technology challenge. you know It's normally a business
Component-Based Architecture and Cloud Partnerships
00:21:29
Speaker
problem. it's only ah you know And so you know i think that we're we're overused, but we we built our own agent. We see it working. We know what it could do as well.
00:21:38
Speaker
And it's it's amazing what's out there, actually, what you see. And it's not just... that The one thing i would I would tell your audience is we are personally seeing models that are good within region slash product slash use case.
00:21:52
Speaker
So we have a matrix of like, if you're in Asia or if you're in Europe or if in US or if you're doing claims or if you're doing new business or if you're doing mortgage, like here, are all the different models are getting good at this, but we're also seeing them leaf frog each other.
00:22:07
Speaker
And of course the foundational models are getting stronger. So the one thing you don't want to do is ever bake a model into a system. Yeah, know can take it with you there. Um, um I mean, I see ultimately three technologies impacting. We call that the services and software flag trap, but you've got LLMs coming in as a real tool for content.
00:22:31
Speaker
You've got Argentic as a, if it's real Argentic, it's the orchestrator. And then you've got Vibe Coding as the producer, right? Do you see what you're doing As a ladder, just brings it all together for clients. Is that where we're going? Yeah, we are the glue. So it's Uncork AI. We were first naming this all different components. We basically said it's just Uncork AI. So in Uncork today, you could create a visual diagram showing the process flow and the automation. Step one, Phil gets it. Step two, Gary. Step three, it goes to an automated system and an LLM will kick off and actually operate on it to tell you next steps. And it's dynamic.
00:23:05
Speaker
which means you could actually orchestrate a agentic workflow, which is what wouldn would be called within each step. You could then also bring in various models. You could bring correlation models, which are deterministic. You could bring in lm's and And within all of those, you could basically orchestrate this. Now, the thing that no one's talking about is security, reliability, repeatability, explainability.
00:23:29
Speaker
That's what's critical to enterprise. That's the most important thing is, can I show a regulator what happened? And when we built a lot of the reasons why banks brought us in to do onboarding, wealth management or institutional is we could repeat a transaction.
00:23:45
Speaker
So if you generate code, if you allow a user to write code or override your code like low code does, You can't repeat the transaction because that code might have changed or the runtime it runs in might have changed without you knowing. You have no control.
00:24:01
Speaker
With the abstraction layer we did in components and this honeycomb shapes of all different components, we could actually repeat a transaction and store it as a vault. Say, here's what happened. In AI, that means we could actually show you what were the inputs and outputs and explain that back to the user what happened.
00:24:18
Speaker
In that case, you can look at um underwriting in an insurance company. And you could say, that's a process which it really is following a series of a runbook. Data is going to come in at all different forms. The LLM will extract it out. IDP will extract it out easily.
00:24:34
Speaker
goes into a processor. It's looked at and basically assessed. Correlation engine kicks in and scores all the different parameters for you to say what's most important. And maybe a human does do a final check, but you know it's no longer waiting months to get processed.
00:24:50
Speaker
That's real today. That's not futuristic. That's today. And we see that regularly within all our clients. Yeah. So what do you see as the role of the hyperscalers as things evolve a bit more?
00:25:02
Speaker
How do you see that evolving as we look at the whole convergence of services and software? Yeah, to me, first off, like we, day one, we said we're cloud agnostic. So Amazon, Google, Microsoft, we're in all three marketplaces.
00:25:15
Speaker
Amazon is our largest partner and we are partnering them right now on their transformation product, which is pretty cool. So using their transform Kero suite to basically automate on Quark build, interesting which is, that's that's a really good.
00:25:30
Speaker
um Then we look at Google, we're launching our AI on Gemini. We tested all the different models and foundational models. We picked Gemini, we launched it. We're using their security suites. And of course, Microsoft, we're integrated to Teams, we're integrated. If you get married in New York City, that's on Quark and Teams. That's actually doing the the marriage license and issuance. um And so they're all three have their strengths and the weaknesses, I would say.
00:25:54
Speaker
In the future, I think they're all also gonna have to figure out their role within the AI space. And I don't think it's clear yet. In that, you know, we are being asked by ah from our clients to offer Uncork inside their cloud, inside their Amazon, their Microsoft, their Google, instead of us hosting it for them. And we call that Uncork on the Edge. It'll be available this year.
00:26:17
Speaker
It's high in demand. And the reason why we're being asked that is because they're starting to be, especially on the financial service side, a mistrust of SaaS hosting. And, you know, it combined with, hey, I could use AI to build that myself.
00:26:32
Speaker
So the ability to basically package on Quark and deploy it inside someone's firewall, behind their firewall, while we still maintain it for them remotely, that's what we call on Quark on the Edge. And so i actually think that that abstraction over the cloud opens up a lot of questions for the hyperscalers and we're the direction.
00:26:51
Speaker
And we are, you know, we're excited by partnering with them. They are they are our largest partners. tonight And so I think that they have a huge role in the future. It's interesting to see all their strength and weaknesses play out in the AI world that we're entering. Fantastic. Well, Gary, this has been phenomenal hearing just how you're evolving during this market, because it's sort of The market's playing to your strengths more than the other way around, I feel. What's your advice maybe to young developers today?
00:27:21
Speaker
Maybe in a way we i had a great part with Gen Z guy of the other day who was just giving me a feel for what it's like being gen z Gen Z in New York City at the moment. But um tell me a bit, what's your advice today to young developers, young folks who want to evolve their careers in our space?
00:27:42
Speaker
yeah To me, it's logical thinking, design, architecture principles, understanding the big picture. Technical debt has two different definitions. One is the software which actually is behind patches, the software that's out of date, software that you need to patch it or get upgraded and it's waiting there to do. The other definition is it was designed poorly.
00:28:04
Speaker
The business needs to change that one module each year for compliance reasons to comply with the government regulations. And it wasn't designed in a way to make it easy to change. So it's impossible. I think understanding that mentality is not even a computer science principle.
00:28:18
Speaker
It's a logic principle. It's ah it's a reuse principle. is i ah I saw an article come out recently, which, you know, and this would be, Phil, if you ever want to do a really good debate, I'll get you the name who wrote this, but it was basically the concept that apps should be, um there should never be open source or libraries, there should be no reuse of an application in the AI world.
00:28:40
Speaker
It should just be written, each time you need it, it should be written. And the reality is whoever wrote that has never, ever gone through an ethical hack, never actually understands technical debt and upgrades. And the reality is if you have 7,000 apps like I owned as a CIO, like each day there are 600 vulnerabilities that feed every week created in the world.
00:29:03
Speaker
How are you going to maintain that across each replication, every instance? It's impossible. It would be good debate is a componentized modular architecture on quark versus a, you know, app by design, app principle.
00:29:16
Speaker
I truly recommend, though, I would say my son my son is also a university sophomore and i tell him all the time, computers, data science, understanding data principles, understanding transformations of that data to make sense and insights, it'll play out in finance. It'll play out in computer science. It'll play out in AI.
00:29:35
Speaker
It's a common skill. And that's where I would say focus. That's absolutely where it is. When you think about it, I'm a little younger than you, but i'm not a lot. I remember 94, I was just finishing my last year at university. The access to information kids have today that we didn't have back then, it's incredible.
00:29:52
Speaker
I mean, they probably know more than we do about most things. They just don't have the experience or the emotional maturity to maybe get on in the space. So it's sort of like if you can get that balance between bringing that young talent and mixing it with experienced people who are really embracing what's going on right now, I think you can have a tremendous time. Is it so is it experience or is it work ethic? And I'm curious what you think because, you know, as a entrepreneur, like I managed 10,000 people and then I created a company with one employee and then got to you know, so the reality is you get to see the same employment concerns, the same employment profiles
00:30:30
Speaker
We are remote first because our talent is dispersed and we search for the best talent wherever they are. um But we also expect just as long as you're getting your job done, doesn't matter where you're located as well as there is another whole principle though, which is return to work five days, being in the office, you know, it's interesting to see the dynamic of the two and where that the way that plays out.
00:30:54
Speaker
But yeah, it's cool. I think work ethic is really important. Passion. Anyone who says they don't have time, they're really saying don't have the passion for something because people have always make time.
00:31:07
Speaker
Hey, people will always find that time together to see a Taylor Swift concert or whatever they like to do with their spare time. It's about passion. And whether you work at home or you go to an office, it's about you have to now manage your own careers and these emerging.
Future of AI and Business Dynamics
00:31:23
Speaker
really advanced apps that are, you know, my I'm an analyst. My job has never been so much fun as it is right now. I've got tools that freaking know me. They literally know me. They're practically talking to me right now. I get my morning digest of, Phil, this is what's going on in services and software today. And I'm like, wow, this thing has educated me.
00:31:43
Speaker
And it's enabled me to be way more dynamic at the job that I do. And I believe that enthusiasm I have is paying off to my clients because they're looking at me thinking, this guy knows what he's talking about.
00:31:55
Speaker
And I think this is the same for everybody. If you can generally embrace what's happening around you and get really good at what you do and leverage these tools in a way that's really smart, I think we're going into a really exciting future. And all this crap about the AI bubble about to burst and everything like that,
00:32:15
Speaker
If you think today, if OpenAI decided we're going to cancel everyone's subscription for a month, and they went to, hey, Gary, what are you spending, 20 bucks a month? It's now 60. You're going to spend 60 bucks, right?
00:32:30
Speaker
This stuff is now completely impermeated in our lives, in the way we do business, the way we do everything. And if we think that this is one big bubble that's just going to collapse around us, maybe there's an economic.
00:32:43
Speaker
impact at some point. I'm sure they could be like cyclical, but I don't believe this AIS bubble. I think this is generally a huge load of excitement. It's permeating every industry, banking, insurance, where you go into some of these industries now, it is so embedded.
00:32:59
Speaker
I just think we're going into a completely new world. I would just feel like I said, if you look at ches just what it can do today, reliably, consistent, and just IGP extraction of data from documents. So we're still in a world where documents exist.
00:33:14
Speaker
We're still in a world where paper and documents are being used. Even if you move the whole mainframe to Uncork tomorrow, those documents are still going to come in and you got to get them into the systems. You can't imagine how much of the workforce population is doing just reading something and translating it.
00:33:29
Speaker
That's it. yeah If you just think about that that, that's trillions and trillions of dollars of value right there. Thinking about those functions differently is what intelligence... And Phil, if you're an underwriter and I ask you, hey, Phil, like a week ago, you underwrote this marine transit company here. You know, why did you approve them?
00:33:48
Speaker
You know, what were you thinking? And you couldn't tell me. But if you could open a log and see the report and see what the AI did and the steps it took and what it looked at in the analysis, that's actually very powerful. That's actually now enhanced better than a human could do.
00:34:02
Speaker
And that's where we have to focus is not just trying to replace what we're doing, but how do we make it better? and And be excited by that. I've always thought that if I could eliminate my own job as a CIO or CTO or whatever position I had, I always thought there's a bigger and better job I'm going to do that's more meaningful.
00:34:22
Speaker
And I think if everyone thought that AI would be everywhere. Like that's, that's I think they call it the termite principle is what I heard. But yeah. Yeah. But you and I are lucky, Gary. It's exciting. Because don't have, you andnna the more we can eliminate our jobs, the more fun we're going to have.
00:34:39
Speaker
Because we're answerable to our clients. We're not answerable our clients. ah You know, it's funny. So I wake up every morning excited for the day ahead. And every time I pitch on Cork to a CEO a CXO, COO, and I still do all those pitches. It's exhilarating because I'm watching almost like Morpheus show Neo, here's what the world is around you. And that distraught world with the nuclear fires. and And that's your technical debt, that's your legacy, and it's holding you back. And it's the first time, Phil, the CEOs are saying, i actually know what tech debt means.
00:35:13
Speaker
Because I used to do this pitch and have to stop and define tech debt for them. Now I don't. Now it's like, I get it. Actually, this makes sense. They all know it's holding them back. And that survey we did shows that as well. So it's a good place to be. in You need that passion. You need the grit.
00:35:28
Speaker
I always said New York City is the place of passion and grit. You could do anything there. And you need the excitement as well when you wake up in the morning for doing something new, which I know you have. And it's very clear in your with your results, of course, and what you've been achieving is amazing at HFS. So it's awesome.
00:35:43
Speaker
Yeah. Thanks very much. Good. Well, thank you, Gary. I look forward airing this to our audience. And thank you for just shedding the whole new perspective of what no care actually means to the industry.
00:35:58
Speaker
Thanks for tuning in to From the Horse's Mouth, intrepid conversations with Phil First. Remember to follow Phil on LinkedIn and subscribe and like on YouTube, Apple Podcasts, Spotify, or your favorite platform for no-nonsense takes on the intricate dance between technology, business, and ideological systems.
00:36:18
Speaker
Got something to add to the discussion? Let's have it. Drop us a line at fromthehorsesmouth at hfsresearch.com or connect with Phil on LinkedIn.