Your AI Project Doesn't Need More Agents

0:08Good morning, good evening, wherever you are, and welcome back to the Enterprise AI Show. This is your host, Aaron, and I'm still down at the beach at my remote massive studio trying to get some things done around the house and also just trying to get some beach time in before the end of summer here in the United States. A quick note on our editing process. We're still working behind the scenes on a bunch of things and we're automating some things and improving some things. And hopefully this should sound exactly like you've been used to. If anything is off, please let me know. Now, moving on to today's show, Brandon has an interview for everyone.

0:44Brandon speaks with Rich Zayad. The main topic is AI transformation projects and why your AI project might not need more agents. And all of that is coming up right after this break. Today's show is sponsored by Nasuni. There's a growing gap in AI right now between what's possible in theory and what successfully works at scale inside an enterprise. The difference comes down to unstructured file data. Many AI initiatives struggle because the file data they depend on is scattered, unstructured, and disconnected from where and how work actually happens. Nasuni changes that. It brings your unstructured file data into a single, secure foundation so AI, both generative and agentic, can access it with the context, governance, and performance it needs in production.

1:32Bring AI to where your unstructured data lives. See what it takes to activate your data for AI and request a demo at nasuni.com slash AI. And we're back. Joining me today is Rich Ziadeh, co-founder and CEO of Abord. Welcome to the show, Rich. Hello. Thanks for having me. All right, listen, you are a professional podcaster and you have a CEO of a board, but you have what I think is a fun backstory and I wanna hear it. So I'm just gonna set you up with, I know you started as a lawyer and then somehow throughout your life, you're now the CEO of a board, which I'm just gonna call digital agency, digital transformation, give us a little bit there.

2:12But I don't think that's the natural path for a legal career. So either what happened or how did you get here? Tell us the story. Yeah, it wasn't my dream to be a lawyer. My dad told me I have a trustworthy face. So I went to law school, essentially. Went to law school, finished law school, took the bar exam, got a job at a commercial litigation firm. And they put me in a library, like writing memos and briefs and stuff like that. And it wasn't the worst job in the world. I wasn't talking to a lot of people. I like to interact with people. And it's 95. I'm out of law school. It's 95. The internet landed like a spaceship.

2:54You go back through the internet timeline. And I was like, I was a nerd. I wrote Microsoft Word macros for fun. Okay, good. And I said, let me go see. I can come back to the law anytime. And at that time, people were making up titles and hiring random people because the internet was crazy. Because no one knew anything? There was no one had any experience in the internet. So you were as good as anyone? I was good as anyone. And I was just technical enough. And I had a good product sense. I cared about good product. And I just tried it and it never came back. That's essentially what happened. Yeah.

3:31So that's the hard left pretty much first year out of law school. That's cool. That's good. You're one of many lawyers I know who left being a lawyer. That's what I would say. Like half of the people I know with a legal degree, they're like, no, I didn't do the law thing. So I don't know, I guess good for you. Yeah, thank you. Yes, it's been fun. I get to do what I love and be in this world. All right, then let's flash forward because I know you did a bunch of stuff, but I think what's real interesting about kind of this moment of time is you guys originally started PostLite with your, I guess you mentioned your co-founder, Paul Ford.

4:02I think a lot of people know Paul, like F-Train, writes a lot, Bloomberg. You can look him up. He's got a lot of interesting things he's done, but it's almost like I'm gonna call it pre-check GPT. You guys had PostLite going, right? It's like way back. And as I understood it, or sort of as I understand PostLite, really you're out there like building apps for clients, doing digital transformation. This is all before AI. So what led you guys to start that company? And where I'm going here is what, why did you sell PostLite? And then why did you start a Borg? I think that's an interesting thing to understand.

4:33Yeah, PostLite was a digital transformation firm. There was always, as we grew and had a phenomenal reputation in New York City, a lot of great clients, just a great roster of clients, great team, great people. And there was always interest. And it had grown to a size for Paul and I. We're builders. We like working with small teams. PostLite had gotten to 130 people, and it was getting big. And the interest was strong. And the winds were starting to change, too, at that point. This is 2021-ish, coming out of the pandemic. And so the right opportunity came along and we sold it. And yeah, that was the why.

5:12And then a board is interesting in that PostLite had an arm called PostLite Labs where we would experiment and scratch the itch and tried weird things. And a board was incubating inside of PostLite in the lab. And it got serious enough that we committed a team to it and carved it out and said, this is interesting. Let's keep going. AI is not in the picture yet. And that team was in flight and was employed by Abort Inc. It was a real thing. Again, before the sale, it's going. And then when we sold the business, we, yeah, we did need a breather, to be clear. The process was intense. PostLite was intense.

5:52But we felt a commitment to that team that was still in a holding pattern while we were selling PostLite and came back to them. And frankly, it was still didn't have a North Star exactly just yet. We knew what we wanted to do. But and then AI landed another spaceship, right? And we're like, OK, we're in we know what to do here. It was crazy. But so that's kind of the meandering narrative here of how we got there. We felt committed to the team. We also like to work. That wasn't my retirement exit. It's not a good scene if you're asking me to sip out of a pineapple in the Caribbean. It's not. Not ready for the beach, huh?

6:34Not ready for the beach. I don't like the beach either. But beside the point, I wasn't ready to just not touch tech. Let me put it that way. Paul felt the same way. So then now take me through this. So I was reading through some of the stuff you guys said. So it sounds like you're working on a board and then you say something like AI lands. And then I think it took six months to figure it out. So what I'm really interested in, like what people did looking back on it. So what did you actually do during those six months when you're like, okay, you get your chat GPT subscription and what happened?

7:07It's worth just taking a step back and acknowledging that we like to go into big, messy organizations. That's kind of our DNA. Like in New York city, there's lots of big companies here. Many sectors of business intersect here. And we always looked at AI through the lens of how do I introduce something this disruptive into large organizations? And from very early on, we were extremely suspicious of just typing stuff in a prompt and then fixing a company and how it works. Because companies are, by their very nature, complex, nudgy entities that are dominated by edge cases and weird process and politics and humans, right?

8:00And so a prompt, the idea that you're going to unleash these robots into an org and everything's going to kind of work itself out, we never bought into it. How we translated into what our mission was going to be took time, frankly, because we didn't know. Everything was so insane. Everybody's making weird images and spinning up what looked like really polished apps that didn't really work. Vibe coding left and right. And we were like, whoa, this isn't, I'm not going to parachute this into an org. So we hung back a lot as we were trying to figure out how to insinuate this stuff into organizations.

8:39And so what did you do when you kind of like, how did you figure out, like, how am I going to take this technology to an organization? Because that seems like one of the key things. I think we're still figuring it out. So maybe no one maybe has the answer, but it's, you know, we have these great tools. You've done a lot of digital transformation on PostLite. But then you're like, and you're probably getting calls from everyone, from executives, maybe foreign clients, like, hey, I got to have some AI, the whole AI washing kind of thing. Even wrote a fun post to something like, I think it's something like, tell your CEO to stop vibe coding, right?

9:10The CEO is walking in and we've all had this experience. The CEO is like, you're not going to believe what I created over the weekend. And you're just like, that's, I don't know if that's going to help us, sir, you know, ma 'am. And so, so I guess that's, that's the part I'm like really interested in. And it's like, I think you would probably have a mind of, we got this thing. I mean, all these people calling us and they're probably just saying, I want it. They're just like, I need it. I want to put some AI. What's that phone call? Like, what are you telling potential clients? Yeah, no, it's everyone's lost their minds, Brandon.

9:40I mean, there's just no other way to put it. And I think in an enterprise setting, it's a really tricky place because if you really look at how AI has blossomed inside of companies, it started at the desk. everybody's doing stuff at their desk you know like check it out look what i just did and so you've got engineers coming up with like polished designs you've got designers writing software you've got different people it's sort of the wild west right but organizations and companies ultimately they are going to they're going to audit all of that to see if there's business value bubbling up, right?

10:22And so when we talk to executives, and that's often our entry point, it's directors, VPs, execs, and whatnot, they're coming to us and saying, hey, I have problems I want to solve. They're not saying, I just heard about Opus 4.6, and I'm pumped. No one's saying that when they call us. What they're saying is, I have been in pain in this particular way for eight years. I've been told in the past that it's going to cost me a couple million bucks. And now I'm hearing about all this revolutionary stuff. Do you think you can help? We've had executives come to us and tell us, we've been banging away for 10 months on AI, and we can't get it over the line.

11:09We've been hired by organizations that have told us they pumped out all this stuff and they just could not get it over the line. And I think if you really zoom in on what happened here, what happened was, I think, was this. My favorite analogy is you walk into a doctor's office and you say, doc, I heard about this new heart medicine. I'm running out of breath when I go up the stairs. I want it. that literally the patient is walking in and saying, I want this medicine. I want this particular medicine. And the doc is like, whoa, how about cool? You saw an ad of a family hanging out in the park because dad's feeling better for about this new medicine.

11:56How about we run tests? How about we get a full picture of your health? How about we understand if your body can tolerate this all of that stuff and so the the script is actually flipped and that's we're back to kind of basics here which is when we meet someone we actually say we'll get to ai it's nice to meet you your organization your problems are interesting your org is interesting we'll get to the solution we'll get to the prescription that you need but we need to talk And you need to do most of the talking, actually. We need to actually understand how your organization works, the tools you have in place, where your data is buried.

12:39It's all the boring shit. It's like, okay, you want an injection of miracle medicine because this stuff has been pitched as, we're done, everybody, stop working, eliminate your own job. We'll give you another one. We're going to cut costs. Back up, go to the beach. Yeah, all of this miracle stuff. And Sam Altman, I think a few days ago, he said plainly, organizations are adopting this stuff way slower than we thought. And that is because, and I can go on and on about that, organizations resist change naturally. They defend the status quo naturally. But more importantly, a lot of the stuff that's being pushed in isn't the right solve.

13:29They're not thought out. They're not carefully architected, not just from a technical perspective, but from a solution fit perspective. And so here we are. And this is good for business, Brandon. I'll be honest, because a lot of people are calling us up and saying, we took a swing, man, and it didn't fly. And so you guys are saying you actually do stuff. And it's true. Once you do the diagnosis and you run the tests and you really pinpoint a solution, it is a game changer because what used to cost$10 costs like$2 now. If you're not going to do that upfront work and do that thinking, good luck.

14:09So I kind of think of all of digital transformation kind of consists of three phases. Most everyone will have some type of discovery, plan, and build phase. And so what I feel like from listening to your own podcast, we'll make sure to put a link in that in the show notes, is that you and Paul are really kind of advocating that. I kind of see discovery has really not changed. You're still going in, doing interviews, talking to people, walking around the office, asking them to show you what they do. And this is just maybe an opening question. I don't know if AI has, maybe it summarizes your notes or something, but I don't know if it does anything in discovery that is new.

14:45Is that a fair way to say it? Do you see any part of that front part of the project or engagement actually changing at all? Yeah, it's good for note-taking and summarizing. And they're like, give me your documentation and then they give you piles of PDFs and I don't have to maybe wade through it as much, right? But I still have to consume it. So maybe I can get good summaries. But the idea that you can just record a meeting and then skip the hard thinking and get solutions out of the other side of that is a fantasy. We go. Our default posture is if you're willing to host us, we want to come to you.

15:27We want to go see how you work. We want to go see how paper and how information travels in your organization. that's a very human process and i think not acknowledging that humans and habits and muscle memory is a huge part of the status quo is a fatal mistake because a lot of this isn't about a bad component that needs to be rewritten a lot of it is about process and people and patterns and it's the basis of digital transformation right that's been probably since the dawn of computing. That has been like the core foundation of it all, right? A hundred percent. And everyone thought they could route around it and everybody's learning a lesson now.

16:11We're feeling vindy. We've been saying this for two years and we sound like grumpy old men, but it's real, it's real. And I've heard it from executives who have said, I wasted a lot of money, man. They just went to town on these tokens and I'm not seeing an ROI. I'm not seeing it. and I think the hard thing, we thought we could skip the thinking. The thing I say on our podcast a lot is AI punishes laziness, punish you. If you think you can skip it, if you think you can just pump in a bunch of docs and say, give me a plan, right? You are in for a lesson, especially I'm speaking particularly in the enterprise context, just to be very, very clear.

16:53You can make some cool, cute little apps for yourself. It's really good at that. You want your own task manager that has some fun features. You can do that. But this stuff is another animal and you have to go in. Now, I think one of the reasons people are thinking that way, and I want you to kind of, I think you'll have some real hot takes here around is that the word agent, right? I think this word agent has just, I don't think anyone can define it, right? I feel like everyone that gives a keynote, defines the agent in a way that says their technology runs this thing that they call an agent. And not only do you need one, you need millions of them, right?

17:30They're largely going to, if it's not said overtly, it's said subtly, you'll probably be able to reduce your headcount a lot because you'll have agents, right? And then, and listen, when I say this, I always want to my disclaimer, I love Claude. I use Claude every day, huge fan, right? But then when I watch, I think his name is Boris, he gives a lot of talks on YouTube about here's how we use Claude at Anthropic. And it's like, he'll be something like, I've got 15 planning agents going and they're running 30 implementation agents. And then they're looping and I let them run for, I don't know, four days straight and then out pops like magic.

18:06And so I feel like in defense of all of us and the executives that maybe you call, it's like, you hear that and you're like, Anthropic's doing great. And you're just kind of feel like I may be missing out. I don't know, my company, whether it's like I'm making tractors or I'm doing banking or what all the advertising, all the things are happening in New York city. It's like, you just kind of feel like I, we're not doing that. Like, you know, in your organization, you're not doing that. You got everyone on a cloud subscription and they're just kind of using it. So like, I don't know, like, how do you reconcile all of that?

18:34Do you feel, cause I know your stance on a board is very much like AI as a tool, not the boss, right? Like you are utilizing AI, but there are at least like that one, and you can mention Sian Altman is like, there are some of these things, like they get your attention, right? I mean, Claude and Arthropic has my attention. I try to see what they're doing just to know. How do you reconcile all of that? Yeah, look, I don't think these pitches that they put forward really land contrary to our thinking. To me, when Boris, I've seen him talk. I actually saw him at a conference in Aspen. This is the guy who made Claude Code, right?

19:10I think him and his team. And he talks about agents overseeing middle management agents that oversee hundreds of other agents. And all I think to myself is, cool, that's execution. You have a particular mental model, sorry, a particular approach to executing on what you want to make real, right? And what I'm not hearing there is is we have a process where we think through what the Claude Code user community has really bubbled up as the key needs and requirements of how this product should evolve. And then we think about the sort of best ways to introduce new features so that people actually use them and not just cake them on.

20:01And then once we do all that, and that's validated, then when we actually have to build the software, I have my army of agents. Here's what I think has happened. This is an engineering narrative. The engineers won for now. This invention is an engineering invention at its core, right? And it started at the engineer's desk. And the conversation is around how engineers are hypercharged and the whole of humanity must get in line now, right? And it turns out engineers, for the most part, lack a lot of the skills to empathize and understand what the world needs. Not all. There are unicorn engineers who think very deeply about product.

20:48And secondly, there is a sort of defense of how power has coalesced around engineering execution. They're going to defend that territory. And the truth is, when Boris talks, I don't hear him talking about the kind of thinking that goes into evolving their software. My observation, there are so many slash commands now in Cloud Code. I feel like he doesn't even know what it's inventing at this point. I feel like its quality of experience has either plateaued or gone down a bit. Mind you, the models keep getting improved because that's its own. to write, right? But as a product and as a tool, to this day, there is no collaboration in Cloud Code, like in a very fundamental way where people know, and other than sharing your project, like deep collaboration for teams.

21:43How does a designer chime in on a Cloud Code project for their wing of, so they can use their expertise and their craft? That kind of thinking is not interesting to I just exponentially scaled up armies of agents and blah, blah, blah. I have my own view on agents and the whole agent narrative. I would like to ask you, Brandon, I'm noticing that the agent type and the agent narrative has calmed down. And at one point, it was like, they are everywhere and they will be everywhere. And this is about agents because the metaphor was so easy to grasp, the idea of this thing that simulates human behavior.

22:25But I'm curious, I'm seeing less of it from what I can gather. But yeah, curious to hear your view on how agents and how they're perceived has evolved. I think it's always been like this artificial construct that people really want to relate to like reducing headcount. And then you have, I don't want to blame. I don't want to say blame is the wrong word, but you have all the hyperscalers. What do they do? Just like everyone does what they want to do. It's like, oh, new technology. You know what I need to do? I need to help you manage this technology. And you need to have a user access and rights and you need to have security and you need to have lots of these agents.

23:00And we have this agent developer console. And if you watch the Google keynote a while back, they created this very complicated agent-based system to make like this, if you will, a marathon running thing. And This is not like, like if you read, she said, Brandon, make a marathon round. I would probably just sit down with cloud code by myself and we would just like hash it out. I wouldn't create any like from happy things. So where, but I think this is where, and I, but I think this even goes back to the cloud code for everyone listening. I know many of these listeners have done the cloud code training, which I've done right now.

23:32It's like, I want to know how to do this. I want to be in it. Okay. And if you go through like one of the main exercises is learn the API. Okay. So I want to learn the API. But really what it is, it's like, here is how you build essentially a chatbot, right? That's really what you learn. It says like, okay, you open up the API and then you send in the system prompt. So you're doing the cloud code and then you get it back, you parse it out and then you go. And so what I thought was very interesting about that, it was like, after doing it all, it was very instructive. And I was like, I don't really know if people really need this in like their apps.

24:08or this is where I, and this is kind of a question back to you, is like exactly what, when you're implementing it with any of your clients, pick anyone that you think is good. It's like, I understand making a call out and you'll always hear something like customer service, right? I get that one, right? Get that use case. That's like the number one. And they're just like, and they go back and maybe they'll do something to reset the password or else, but pick any application, CRM, Workday. Like, I don't know anything, accounting. And it's like, what, I don't know. Like, what am I asking the age?

24:39What am I asking Claude in this case or any of them to do? And this is where I've always wondered, like, and then you can get to these more complicated, like, what is it? Lane loop or lane. I can't remember. I forgot the name of it. I think I'm saying that that lets you do more of this. And I guess this is where I'm coming in. Are people, and this is my fundamental question is like, are people really doing this in these applications? Like I see chatbots showing up on the side of the apps a lot. But then I'm like, what are people really invoking in an API inside an application and then doing and having it come back and calling that an agent?

25:10Because that's what I'm always like, what's really happening there? So I'm curious. I don't know if you've had clients that are actually doing some projects like that, but like what's going on is really what I want to know. I will tell you, the clients guide us in terms of what we're going to build. We don't tell them we're going to build. You look like a dozen agents will sort you out. Like we don't say that, right? Instead, we listen, right? And we need to understand what they need, right? And we've built very few agents. Based on the client roster and the needs that have come across the line, we've not built many agents.

25:45And that's because when we go into these worlds, the worlds, the problems we're solving are just, they're complex from a process perspective, and they're just very hairy. And so how do we use AI? I think this is worth talking about. I can't speak to how AI is researching medicine or how, like there's so many, that's why this stuff is transformative. So let's acknowledge all that. We are software builders, right? And for us, we see this stuff falling into really two categories. One is the spreadsheet, I call it a spreadsheet problem. We'll have people come in and say, I need AI and agents and chat boxes so I can solve my business.

26:30And then we will respond with, how about we get you off of running your business on a spreadsheet? Like boring stuff. How about we have tools that actually read these invoices, which are PDFs coming across the line so that humans aren't making mistakes. those kinds of problems are still everywhere big companies are using old tools old processes and 11 of their salesforce instances and they don't understand the rest of the functionality that is boring stuff and the truth is when we go into these orgs it's like let's go back to the patient the patient's like i want this shot because it's going to give me a strong hard upper body and we're like, you have congenital heart failure.

27:19Why would you, do you want muscles? How about we fix the core first? That's often AI is great to do that faster and cheaper. That is real. Once we know that you need bypass surgery, the cost of doing that is gone down. So let's do that. Now, once we, so that's category one, I call the spreadsheet problem, which is kind of old school basic stuff. Category two is, okay, we got you to a good place or you're already in a pretty good place, but your dashboards suck and you want to be able to ask questions so that your world of data is more accessible and AI is a great way to do that. And a narrower either vector or rag-based solution that understands your data.

28:06So you can say, how are sales going in Michigan? And it'll actually give you back a plain English response. That's really cool. That's smart. That actually is additive. It's targeted. And that's AI living in your world every day. You're actually hitting LLMs. You're actually using this amazing technology in a truly defined, targeted way. You notice in both of these, which we found is north of 90 % of our delivery today, don't really involve agents. Now, look, are there cases where agents can flag things and automate processes and shoot out emails in the morning? Of course. We're not religious about this.

28:48And I think all of this is colored by what I said earlier, which is what's going on with you? How do you eat every day? You exercise. It's like you. And that's very human, actually, because it requires us to listen to other people and go into their cultures and into their worlds and understand them. I hope I'm not sounding preachy here, but it turns out everybody's kind of coming back this way because you know this probably better than me. The pie chart of software, the majority of it is in these companies. Consumer software, I don't know the number, but it's much less than 50%, 70%, 80%. And so they want that pie, Brandon.

29:33Like everybody's, they're slicking their hair back and they're going out with their shirt button, their silk shirt button. All these AI companies, they're going to need that pie. I think that's the forward deployed phenomenon, right? That's where everyone now is, they call it forward deployed, which sometimes people would call that sales. Other people call that solution architects. Some people call that consultants, but it's just like, it's really just a fancy way of saying what you said at the beginning. It's like, we need to go to your place. We need to sit down with you. You need to tell us what you do.

30:00And if it makes us all sound better that we call this forward deployed. Soviet. The title is somewhat irrelevant. As long as everyone shows up. That's exactly right. And that is the right formula. And that's, I think we're going to see a lot more of that. All right. Now on a more broad digital transformation, I want you to feel free to really disagree with this, but I believe no successful digital transformation project can start unless it has a real burning problem. Like you are in some type of crisis. As someone who's been on, so I like to say on this podcast and other podcasts. I used to work a lot in migrations.

30:34And when the person would call up and be like, we're thinking about migrating, we could save money. We're maybe interested in this new technology. Really nothing would happen. When they would call up and say, the data center, real estate has been sold and it's going to, we need to leave by May 1st. Everyone was extremely motivated. So it's someone that's been through many of this. So I almost feel like if someone asked me, I want you to tell me if this project's going to be successful. That's all I would do. I'd go in there and I'd say, what are the current problems? What are the current incentives?

31:05And are these changing? And if not, I'm probably going to say, it's like, we may do something, but I don't know if it's going to be that effective. What is your take on that? I pretty much agree with you. Listen, there is nothing that aligns humans better than a deadline that no one can control. nothing it is i think that's a perfect way to say it yes yeah and that could be and i've done it and i've done it and i've created them as as a c as the ceo i'll be like we've got an event september 18th where we're going to announce our new thing and there's people coming to the building right if i didn't have that like this externality that we have to commit to it's harder to impose it sometimes it's we're getting acquired and we have to transition our systems by this date or we don't get an earn out.

31:56We don't get this check cut to us. Beautiful motivator. These are things that align people. Otherwise, people revert to their default velocity, right? That's just human nature and how people work. I've been working this way for 11 years. We've gone into government agencies where they're like, let me show you how I work. And they're proud of the craziness. yes they're proud of the nuttiness right and unless you have a sledgehammer of a stakeholder that's going to make change happen it's very hard it's very hard to pull off so i mostly agree with you i will say there is the occasional in like really ambitious stakeholder who's like we have to transform we're not dead today we are not in pain but our competitors have gone to a new place and we're not there.

32:49And if we don't do something, like it's coming for us, right? That kind of urgency is important. The org often kills it. You probably know this as well as me. There's nothing funnier than the innovation group inside of like large organizations, right? It's almost like this playpen where there's little wooden blocks and there's little toys and toy airplanes. And they're like, they're innovating over there. And that poor group, because there's no urgency tied to them, just can't ever do anything, right? Yeah, nothing happens. Nothing happens and it dies. And then your competitor shows up and just eats your lunch.

33:29And then it's like, we better get moving. Now it's time. And then they go and hire McKinsey to come up with a new strategy or whatever the hell. Yeah, urgency. That's how it works. We need it. I always think it's just a great shortcut. I think it's for any executive or anyone. And when I'm involved in projects, I just think this is like, back to your discovery question. That's the first thing we should be talking about. And if we don't have it, like we could acknowledge it, what you said. And then what you said before, it's like, it's no one really likes to do it. It's like almost like being a parent when everyone's having fun where it's like, we don't have any urgency.

33:57So just what you said, where you did it with a date and another group, you can come up with it. And sometimes you have to like consciously decide we don't have it. Here's about what we're going to do. And now if someone's like, I'm not sure we want to do it. Then you get to like large bureaucratic process will probably persist for a long time. People come to us because they tried internally very often. Oh, absolutely. They hit the minefield, right? And they're like, all right, I'm done. We're like a startup for hire in a lot of ways. All right, I like it. We're insulated from your politics. All right, listen, we're almost out of time here.

34:28But listen, I want to finish on a fun lightning round. You're in New York City. You have a lot of hot takes. You do a lot of podcasting. So let's just have some fun here. I'm going to fire some questions and whatever you want to say. So question one, are we in an AI bubble? Yes. Okay. Question two. No, I could expand. You can expand. Go. Yes. And yes. And why are we in an AI bubble? We're just aiming for the wrong thing. I think it is transformative. Like every tech, we tend to get ahead of our skis and then second go around, it starts to get good. It's funny. I almost say it this way. It's like, we probably are, but we're probably not in the bubble we think we are.

35:06If everyone's talking about the thing, it's like, that's probably not the problem. Right? So this week is everyone was talking about, actually, it's one of the questions we'll get to it. It's AI watermarking, right? This is the whole thing where Claude's make it easy to figure out if your text, is that a good thing or a bad thing? It's good, but I don't know if you need it because everything is so mid. I don't know. I'm seeing less goofy images of elves riding motorcycles these days. So hopefully it'll correct itself, but I think it's a good thing. All right. This is where, back to the bubble question.

35:36Should GPUs be treated as securitizable assets? This is the whole Jensen-Wang And you can basically, if you will, banks will now loan money against GPUs. And that'll be a way for more companies to have access to GPUs. Is this a good thing or a bad thing? Oh, who cares? Just obfuscation nonsense. Because they're trying to... You're not worried? I feel like this is where people make the mortgage back thing. It's like, we should care because we may all care later on. Because they're like, oh, they're hiding money. We didn't know about that. Yeah, we're reaching layers of abstraction that are getting worrisome, to be perfectly honest.

36:10and everyone's kind of lost their minds and they talk crazy. And this is when things start to get indecipherable for the masses, you start to worry, right? And that's, I feel like that's happening here because the numbers are not real numbers anymore. There's these abstract giant 3D figures at this point. So I'm, I, it's above my pay grade, but I'd rather stay, I, I, this has me drifting into my rants about crypto and how that's obfuscated and complicated and messy so that everybody doesn't know what the hell's going on. But yes, continue. Sorry. No, I like it. All right. And then here, we'll end on just a fun one because you're the quintessential New Yorker, I think.

36:50So what are three things everyone should do in New York City? What would be, if I have one day, what are the three things I should do? In one day. Cats' Deli, obviously. All right. Go see a ball game. There's minor league baseball in Brooklyn too. So no excuses. All right. Or the Mets and Yankees. The third, go to a dive bar. All right. Just some good ones. I don't hear that a lot. I don't hear go to a dive bar. Ask AI for right. It'll give you recommendations. It's good for that. That's fantastic. All right, listen, I really appreciate you coming on the show today. So if someone wants to do discovery with you and your team, how do they get started with Aboard?

37:33Yeah, just reach out. We answer everything. Hello at Aboard.com or just go to Aboard.com. There's a form on there as well. And check out our podcast. We give advice, honestly. A lot of times it doesn't go anywhere. We like talking to people. We like going into people's worlds. That's just, we're anthropologists in a way. All right. I love it. Yeah, definitely check out the podcast. That'll be, it's mostly Rich and Paul. And they sometimes have some guests. So they're fun, fun listen. You'll get more of a talk like this. But with Paul, he's always ready to say something funny. So with that, I want to thank everyone for listening to the show.

38:06And we will talk to you next time. Thanks for listening. Check us out at theenterpriseaishow.com for past shows, newsletters, and all things Enterprise AI.