SG EP 36 ENHANCED
===
Richard Ellis: [00:00:00] In 20 25, 90 2% of companies plan to increase their AI investment, but only 1% describe their deployment as mature Bain's. Latest research confirms the gap, while 58% of executives expect AI to reshape their business model within three years. Fewer than 15% have actually redesigned their workflows, the result automation that accelerates confusion.
Richard Ellis: Meanwhile, OpenAI and MIT estimate that generative AI could add up to $4.4 trillion in annual global GDP, but that only materializes when humans and systems work in concert, not competition. AI may amplify our work, but only clarity makes it effective. Welcome to some goodness where we engage seasoned business leaders and experts to share practical guidance to help today's go-to-market leaders, execute, lead, and win in a fast [00:01:00] changing world.
Richard Ellis: My guest today is Ben SCOs, director of Data Science at Cat Labs, where his work spans numerous areas including leading research and development on applications of AI technology. Ben combines a deep technical background with a master's in the history and philosophy of science from University College London, giving him a distinctive perspective on the logic and limit of algorithmic thinking. Well, Ben, welcome.
Ben Scoones: Oh yeah. Thank you. Looking forward to chatting.
Richard Ellis: So, uh, let's, let's just dive right in to AI readiness. One of the things that, uh, the stats are showing is that 92% of companies are investing in ai, but only a fraction have integrated meaningfully. Right. And so, um, mm-hmm.
Richard Ellis: One of the things that you and I had talked about is just this idea of really understanding your own workflow as the true step for readiness. So I'd love to start [00:02:00] there.
Ben Scoones: Yeah, absolutely. Um, so I think understanding your own workflow really, it, that should be a basic requirement, I think for, for probably doing anything within a business.
Ben Scoones: But I think it's particularly important when you're looking to, um, integrate AI technology. Um, because if you think your workflow really is a, a series of steps for how you, how you do some job or how you do some task within your company and. Even though AI is a really powerful, um, and impressive technology, it's can't do everything by itself, right?
Ben Scoones: It needs some help. It needs some direction for you to be able to do that. Um, kind of just like a person would really, if you think about it, you know, AI in some situations is maybe as just as good as a person. Think about those two interchangeably. Um, if you can't tell a person. What to do and expect them to be successful.
Ben Scoones: Um, if you can't define that person's role well enough for them or the task that's required of [00:03:00] them, why would you expect them to do a good job with that? Um, and I think similarly, if you can't do that and you're just expecting AI to be able to do all the work for you. Uh, then you should expect the same results.
Ben Scoones: You're not gonna see success there.
Richard Ellis: No, no, that's a, that's a good way to put it. In terms of just kind of thinking what level of, uh, documentation or structure or, you know, um, process has been lined out for where you're going to apply AI or like you said, really needs to be, you know, processed first before you overlay any kind of tool or technology.
Ben Scoones: Yeah, I think so. Um, and I think, actually that gives me another thought. I think if you are trying to, um, let's say you're, you're trying to make a, a particular process, programmatic. You're trying to make it repeatable in some way. That discovery process really should be, that's something that someone's trying to figure out.
Ben Scoones: They're trying to discern what those steps are, what are the things that are repeatable here? What are the kind of individual [00:04:00] tasks or series of steps that, that I need to do to make sure that this can be done? Again, how do I define that process? And only once you've done that is then where you should be looking at other technology and thinking, okay, I can introduce AI in this step, or I can introduce AI in this whole series of steps.
Ben Scoones: Um, but that's, that's kind of the, just like you said, that you should be beginning with that, and then you should be thinking about, okay, how do I integrate technology into this to make this more efficient?
Richard Ellis: One of the things that, uh, is interesting about ai, and I, I'd appreciate your perspective here, is there's something about it that causes people to, to not take that first step of just kind of thinking, well, let's just throw AI at it.
Richard Ellis: Right? And you even kind of see this from the top down in organizations where there's kind of a mandate, we need to be AI first, right? But we need to start using AI in everything. And they just kind of. Shout it out to the people to go figure it out. Uh, why, why, why do you think that we're taking that approach?
Richard Ellis: Just because it's [00:05:00] AI versus, you know, any other tech, we would say, Hey, let's define the process and the steps and, you know, all of that good stuff.
Ben Scoones: Yeah, I think that's probably, probably a couple of answers to that. One would be. I think people are still figuring out exactly the limits of, of what it can do.
Ben Scoones: Hmm. Obviously that's changing a lot because the technology is developing very rapidly. But I kind of appreciate that idea of, well, let's see if it can do this. And if it does, then that's fantastic. If it doesn't, okay, we can scale back and and limit our expectations a little bit more. So in that sense, if you're being just kind of, uh, experimental with it, then I think that's good.
Ben Scoones: Um, I think that kind of makes sense. But yeah, sometimes I think maybe the, the flip side of that is you can also just be very lazy and think, well, I've seen this technology do something that looks kind of like what I wanna do. Um, or you can even imagine that because it can do [00:06:00] this one thing, it can do something else, which really is, is not related for whatever reason you think that it is.
Ben Scoones: Or you think those things are kind of in the same bucket and so. You just take that lazy mental step of saying, well, I think AI should be able to do it. Um, and you don't really do your due diligence of figuring out, you know, again, what that process should look like, what success looks like, really figuring out exactly what tasks you're wanting it to do.
Ben Scoones: You're kind of just offloading that hard work onto a system or onto a technology that. Is, is maybe not ready for it in some of those situations.
Richard Ellis: Right. And, and I think part of the complexity is just, uh, how rapidly the technology is advancing, right. Combined with the fact that we don't know what we don't know.
Richard Ellis: And so we're moving beyond just kind of machine learning and, uh, automation to now, you know, predictive and, you know, even decisioning and, and just a creative aspect of AI that. Yeah. It, it is just brand new to us. And so [00:07:00] I think it just kind of fosters a little bit of, you know, kind of ad hoc, throw things at something and see if it sticks right.
Richard Ellis: Which is probably not the most, uh, appropriate, uh, way to formulate an AI strategy.
Ben Scoones: Yep. Yeah, I, I would agree. And I think, um, that's. So that's part, I think part of the challenge with all that businesses face, I guess, especially when you're getting that kind of top down direction from, from the board or from investors or whoever it might be where you're saying we need to be using AI in some capacity.
Ben Scoones: Yeah. You're kind of like someone who's got a hammer and you're running around looking for a nail and there might not be any nails for you to hammer in with that. Right, with that tool. Right.
Richard Ellis: Yeah, that's a good one.
Ben Scoones: And it's, it's the same with ai, right? It's, it's not the best solution for every problem, really.
Ben Scoones: You should be Ty instead with what are the problems that we have or what are the types of things that our business does, and then trying to identify specific situations where you know that this is the right tool or you know that [00:08:00] this might be a tool that gives you some benefit of what you are currently doing.
Ben Scoones: In which case pursue it. Um, and certainly it does open up new opportunities, so I don't wanna make it seem as if you only have to do it within, within the lens of what you're already doing. Um, but a lot of the application is two things that you're already doing. It is a tool that is gonna give you some efficiency or some improvement in quality in some way.
Ben Scoones: Um, but you just have to make sure. That you're applying it to, to the right problem. That it is the right solution for whatever problem you're trying to solve. And I think a lot of people are not doing that. I think they're thinking too much about it as the very, a very general purpose thing, which it is, but just a little too general purpose, uh, and not really doing the work to figure out whether it's gonna be good and worth pursuing, uh, in any particular situation.
Richard Ellis: Right. Well, a lot of the conversations we're having these days with our clients, with the teams is. Help me figure out where to start, right? Uh, where and how do we start using AI effectively? And I think one of the things that you [00:09:00] shared, uh, with me offline is, you know, go back to, you know, the basics of thinking about it, uh, as an algorithm, right?
Richard Ellis: And, uh, or a series of algorithms. So maybe we can kind of just do some grounding here on, uh. If you're gonna walk before you run with ai, you know, what are those types of tasks, uh, and kind of thinking algorithmically that you might look for to understand where you might start within your teams, uh, in your company?
Ben Scoones: Ai really, when we're talking about AI in the modern day, obviously that includes things like machine learning, um, but primarily we're talking about LMS and. Really an LLM is an algorithm, right? It's, it's a model, but really a model is, is kind of an algorithm. So an algorithm is just, uh, a series of steps for completing some task.
Ben Scoones: So a very simple everyday example of algorithm is a recipe. So if you have a recipe for baking a cake and to be able to bake a cake, you get these ingredients, you measure them [00:10:00] out, you mix them together, you, you know, bake them for some period of time, and then at the end of it, you should get. You know, you should get your cakes as long as you've done all those things effectively, clearly.
Ben Scoones: An an LLM is the same thing. It's just what you put in is a series of words. And then the LLM does a bunch of stuff to manipulate those words and it's predicting the next word or the next sequence of words. And that's a little bit of a simplification of what's happening, but, but you can effectively think of it like that.
Ben Scoones: It's an algorithm for predicting the next word or the next sequence of words, summit. What's incredible about it, I think, really is how well that lends itself to kind of higher order functions. Um, higher order tasks. It's not just predicting words. You can use that to do planning, think strategically for generating new ideas, um, for taking actions with tools.
Ben Scoones: So writing code and then being able to call those tools and get back, you know, truthful information from APIs, all that kind of stuff. So. [00:11:00] I think it's, it's pretty remarkable that just that kind of simple capability leads to those things. Um, but it is important to remember that those higher order capabilities don't come from nothing.
Ben Scoones: You can't just necessarily prompt the LLM and say, I want you to do this, and it will be able to do all those things You have to think about. How you're gonna construct from this ability to predict sequences of words, how you're gonna construct those higher order capabilities. So what are the steps that you, as a person who can speak and understand concepts and can reason that kind of thing.
Ben Scoones: What are the steps that you need to take and not shortcutting that? And just thinking about it very obviously as a person, but really each individual step. If you can spell that out, then you're gonna get better performance from the LLM. And so really you're building an algorithm for doing this task.
Ben Scoones: Which is that series of steps. Again, that's what an algorithm is, and the LLM is either us doing all of that and you're kind of telling it what steps there needs to do, or it's just one of those steps and it's [00:12:00] performing that particular step very well and in an automated way.
Richard Ellis: That's good. I, I like that and, and I, like you kind of pushed into, you know, some of the planning and the more esoteric tasks that we might be involved in almost intuitively without really thinking about it.
Richard Ellis: That was one of my aha moments when using, you know, Gemini or chat, GPT, et cetera, as at first I was kind of thinking about what are the steps I'm doing on my computer that I can, you know. Have AI do more effectively for me or instead of me, or automate, et cetera. But then it kind of dawned on me that there's, there's, there's brainstorming that I do.
Richard Ellis: Off my computer, right on flip charts, on a whiteboard and, and to use AI as a sounding board and a collaboration teammate to test ideas. You know, that's part of my workflow that, uh, that I just didn't really think about. But if you kind of map out your workflow, like you were saying, then that starts to give light to a lot of opportunities [00:13:00] to, to plug AI in or leverage it as a tool effectively.
Ben Scoones: Absolutely. Yeah, absolutely.
Richard Ellis: Well, one of the things that, um, so we've talked about, you know, that mapping out the workflows, mapping out processes, designing processes, and really understanding what that process is, is a great first step and prerequisite. Uh, and of course, you know, Bain has, uh. Published in one of their studies that the companies that are seeing the highest ROI from their AI efforts are those that mapped out their processes before automating.
Richard Ellis: So great first step for everybody to take, but, um, what, what's in turn? It's, in my mind, it's not just processes, but it's also data and having access to data. Uh, where does that data come in? And, and, uh, how important is that in terms of, uh, in terms of an organization's ability to be, you know, effective with, with an AI strategy?
Ben Scoones: Yeah, I think that's a good question. So data, [00:14:00] data comes into it really in, in two, just, uh, two distinct points. So one is obviously when the model is being trained, um, and for the vast majority of people, that is not gonna be something that you ever get involved in. Having good quality training data, understanding what the model is, learning effectively, learning language from, uh, that's relevant, that can determine, um, its capabilities in different areas.
Ben Scoones: Its familiarity perhaps with certain domain specific terminology. Um, it's, you know, safety. How harmful is that content that you're providing it with? Maybe what, uh, what kind of truth claims are perhaps being made at some of this content. Um, not that it has an understanding of truth, but. That's, you know, whatever relationships it's seeing between particular terms or concepts represented in the language, that's what it's gonna be predicting when it's generating text.
Ben Scoones: So that's one side, again, that's not gonna be relevant for most people. The other side is how does data in some way, feed into [00:15:00] you, uh, making better use of this model? So, and there are, I think, again, you could probably divide that into two or, or maybe more, um, different, different categories. But one way is.
Ben Scoones: I guess we'll talk about retrieval, augmented generation. So that's why you're bringing in knowledge from some external knowledge source and you're giving it to the model because the model needs that knowledge to be able to perform its task. So a chat bot that is retrieving policy information, um, when it's providing customer support.
Ben Scoones: That's an example that that might use rack for augmented generation. Uh, because the model in itself is not gonna know your company's particular policies. When the customer asks a question, if you just ask Chan PT what is, um, I don't know, what is Amazon's return policy? Or something like that, maybe it will have that.
Ben Scoones: 'cause it's probably easy to find. But for a lot of companies, it's not gonna know that you're gonna need to provide it. Um, so that's one way when you're trying to provide that knowledge. Uh, but [00:16:00] another way where data is important is if you are trying to provide, if you're trying to fine tune a model, perhaps.
Ben Scoones: Um, or if the model is having to interact with data in some way, then having examples of how the model should be behaving, what's the input and the output that you're expecting, um, to be able to build up your evaluation set to be able to validate that the model is performing well. Uh, data's important there as well, and I'm sure there's, there's probably more areas that I could talk about, but, uh, I think certainly that first one with the rack.
Ben Scoones: Um, is, is gonna be a very common one for pretty much everyone. Any task that involves using some kind of knowledge that's not fairly general, um, you're gonna want to be including that in a model in some way. 'cause again, just like you said, you're mapping out these processes. What are all the individual things I do.
Ben Scoones: Don't forget to put in everything that you already know. Right. It's necessary to complete that.
Richard Ellis: It may be in your head. It's not just about, yes,
Ben Scoones: exactly. Yep. It's not just about the flow and the individual tasks, it's about the content of that flow, the content of those [00:17:00] tasks, what knowledge is required there, so you have to make sure that's a part of the spot.
Richard Ellis: That's really good. Well, we're just wrapping up a, uh, a sales onboarding project for a client. So we're, we're defining their, uh, sales new hire onboarding process, you know, from a a, a one week intensive to the 30, 60, 90 day structure of learning and, and being onboarding to being on onboard and, um. You know, one of the things that was, uh, you know, immediately apparent is that this particular company was reliant on a lot of quote unquote tribal knowledge.
Richard Ellis: Uh, there was a lot of who we are and what we do and why we exist and the markets we serve. That just wasn't written down in a clear, cohesive way, and so we were having to put some structure and, and do some content creation around that. And that's all for, you know, a, a sales new hire. But the same kind of applies to your AI agents.
Richard Ellis: If you want them to be able to be a smart assistant and agent [00:18:00] and teammate for you to help you with your strategic planning, so you know, and so forth, then you need to feed it that. Type of content so it knows your company, your markets, you serve, your customers, your solutions inside and out. And so I think it can expose, you know, a real opportunity for, for a company to get their arms around, you know, that kind of data and, and content to feed for the rag.
Richard Ellis: Right?
Ben Scoones: Yeah. So I think something that's kind of interesting, just as a side note. You bringing that up about, uh, organizations taking their documentation very seriously. Something I've seen, uh, within the past year or two has been documentation as code. Um, so I dunno if you've ever heard of that, but that is the idea of applying.
Ben Scoones: So when you, when you're maintaining a code base, it's very important that you're managing any changes that get made to it. Making sure that nothing gets introduced that's incorrect or buggy or, you know, it's gonna cause a problem with the, with the software. You can do the same thing with your documentation.
Ben Scoones: Um, you can use those same kind of tools to make sure that [00:19:00] documentation is being maintained very rigidly. And I, I don't know whether this was around before, uh, AI became a big thing, but I think the importance of that now is even greater, um, than, than it was previously because this is something that is being ingested by a program by an LN or you know, by an AI system.
Ben Scoones: And you have to make sure it's correct. You can't have people kind of just changing it willy-nilly, um, because that's gonna affect the way these, uh, these systems function. So that, I think that just kind of speaks to your point of the importance of that.
Richard Ellis: Yeah. That's good. That's good. And I think all of these are just, you know, extra elements of thinking through ai, uh, from a number of different angles and dimensions to, to make sure, you know, you're, you're not, uh.
Richard Ellis: Misapplying it, you're not missing an opportunity to get the max effectiveness out of it, et cetera. Well, let's move into just some practical applications in the go to market engine, which is what we're all about these [00:20:00] days, right? So sales, marketing and customer success. And, um, you know, you've, you've authored an ebook, uh, that highlighted some applications, but, uh, where might be some areas where go-to-market teams can actually see a lot of value initially.
Ben Scoones: Yeah, I think a, a few of these, these are gonna be common to a lot of different use cases, but I think you're seeing some kind of interesting applications in go-to market and, and kind of the last couple. So first one I would say would be just your day-to-day tasks are probably gonna be made much more efficient with the use of ai.
Ben Scoones: You are doing so many kind of administrative type tasks while you're working with. Emails or some kind of text-based data or voice data or whatever it might be, day-to-day generating presentations. These are things where AI can just kind of sort into the work that you are already doing and probably just make it easier and quicker for you to do a lot of that work, uh, and maybe provide some quality improvements as well.
Ben Scoones: So [00:21:00] that's, that's one area and that's across the board. So many different jobs, so many different sectors will have things like that. Another is, uh, deeper and more efficient research, I would say. So when, uh, LLMs kind of first became available, they were not very good with current knowledge. They were not very good at doing research.
Ben Scoones: You're really seeing a change in that recently. Um, that's a, a problem that a lot of, uh, model providers have put time and effort into solving. And a number of models that are available now have these research modes that you can go into. And so you ask you a question and you do it in research mode and it will pull up to date information, whether that's academic papers or websites or blog posts or you know, whatever else.
Ben Scoones: Maybe it's got access to some other documentation if you, um, provide it with something I don't know. Um, and you can get a really. Nicely referenced and cited, um, research paper effectively on this, on this topic. Mm-hmm. [00:22:00] Um, so you might not be doing that from an academic perspective and go to market, but you might be doing it, um, for researching a particular company and how they're doing in the marketplace and what their distinctives are and all those kinds of things.
Ben Scoones: And it just makes that job so much easier for you and, and it probably improves the quality of your work because you spend less time doing that. You probably find a lot more sources than you would've found previously. Absolutely. Um, and you can focus your effort on reviewing the content rather than having to find the content.
Ben Scoones: Right. So I'd say that's the second area. The third I think is ideation. So you talked about that a little bit. Mm-hmm. You do brainstorming as part of some of what you do, and you can use, uh, uh, you can use AI to kind of bounce those ideas off of. And that's, I think, a really interesting application. I dunno how many people are exploring those kinds of things, but just asking for, gimme 10 examples of this, or maybe gimme some ways that this is true and.
Ben Scoones: Something I like about that is it doesn't necessarily [00:23:00] matter so long as you've got your wits about you. Whether the ideas it gives you are all correct. It could give you 10 ideas and two or three of them could be, could be wrong,
Ben Scoones: completely wrong. Sure.
Ben Scoones: But it starts you thinking about it. It means that you're not having to go through that cognitive effort of starting from nothing.
Ben Scoones: And instead you are, you've got the, you've got some content to start working on. You're like, oh, okay, well that's kind of an interesting idea. This idea made me think about this other thing. Maybe that's something that I can use it. So it just gets that process started. It gives you something to wipe from or something to improve.
Ben Scoones: Um, and maybe again, gives you some ideas that you wouldn't have thought of if you were just doing it by yourself. Absolutely. So I think that's something that's, that's worth exploring. Um, another would be, uh, co-pilots. Right. So any, any job or any job function where there's a fairly specific flow and where it's already digital.
Ben Scoones: You see this with coding and I know there are a number of, uh, co-pilots in the go to market space. You're doing the same kinds of tasks. [00:24:00] Just being able to ask a co-pilot, can you just do this for me again? It's kind of like that first example. Improved efficiency and day-to-day tasks, but these are much more bespoke to some industry.
Ben Scoones: Okay. Um, and probably have much better features, I would say. So. You can take a large portion of your work or this, again, this particular task and just kind of give it to AI and it's maintained by someone else. Um, the quality hopefully is, uh, is gonna be good in most use cases. Um, so that's another area where you can see efficiency.
Ben Scoones: Uh, and then I think the final one, which. I mean, there's a lot of exciting applications. This one is, is exciting in my mind as well, is, uh, you are democratizing skilled work in some way. Mm. So I think coding is, is a good example of this. I'm not necessarily advocating for vibe coding or something like that, but, uh, having these, let's call it like, uh, analytics assistance, right?
Ben Scoones: Mm-hmm. Um, if you're working in marketing, you probably [00:25:00] have a number of customers in the pipeline. All of that information is gonna be stored in a database, but you wanna be able to retrieve that information. If you don't know how to code and write sql, you can't pull that directly back from the database yourself.
Ben Scoones: So you have a barrier there. You are the skilled party or the knowledgeable party about how you would use that information, but you are not skilled in the way that you need to be to get, to get that information from the way in which it's stored. And so AI can kind of come in there as um, I guess mediator.
Ben Scoones: Mm-hmm. I dunno exactly the best term for it. But it can help you out with, with doing those things. It means you don't need those skills. You can focus instead on everything that you do with that information, rather than having to worry about how to interact with those. Those systems instead. So I think that's, I think that's a pretty exciting one as well.
Richard Ellis: And I think that's really interesting. And just thinking about it from the standpoint of addressing some of your skill gaps, right? Which normally you do by surrounding yourself with skilled people. Uh, but as we grow and [00:26:00] scale our, our teams, you know, maybe you don't have to hire that researcher, that coder, you know, that analyst, right?
Richard Ellis: You can le lean on. AI and agents to kind of do some of that work and kind of bridge that gap. So just to summarize, I think those are great five, uh, practical ways to apply AI that teams can kind of just write this down as a checklist and, and look for these, uh, opportunities for, for improvement or for value add.
Richard Ellis: So there was efficiency, uh, research. Ideation, which is one of my current favorites, uh, copilots and then democratized, you know, skills or access across, uh, a a lot of data and, and, and doing things that you maybe can't inherently do right now. Really good. Other than just kind of thinking about. You know, looking for those opportunities of, of potential goodness with ai, starting with mapping out workflows and [00:27:00] processes to make sure you are, you know, applying AI where appropriate.
Richard Ellis: Any, any kind of final tips or, or practical advice you would share with our listeners on getting started?
Ben Scoones: I think probably one point that I would add to the importance of a workflow is it gives you a sense of what the output should look like. You, if you understand how to do that job, then you also know what it means to do that job.
Ben Scoones: Well, hopefully. And that's a very important thing. Uh, it's easy to, it's easy with AI technology to kind of prove a concept and say, it seems like it can do this type of thing. Well, it's very different to implement that in a production, uh, or sorry, in a product or in a feature. Hmm. That takes a lot of work.
Ben Scoones: Uh, and you have to be very sure that it's gonna do that job well, that you're giving it to do. So one of the ways to do that is to evaluate it and to be able to evaluate it. You have to understand what that workflow looks like, what [00:28:00] success means. Um, and so I guess that would be some advice. Make sure you're doing that, uh, and make sure you're doing it well.
Ben Scoones: That ML ops piece, machine learning operations is just as important here as it's always been for, for those types of systems.
Richard Ellis: I think that's great, and I think that's, that's where the human expertise element comes into play. So it can't take away all of our jobs if we're needed to, to validate it Right.
Richard Ellis: To uh, to judge it. Yeah. That's to get the best out of it. Yeah. And, uh,
Ben Scoones: yep, that's always true. Someone's gotta decide what good looks like,
Richard Ellis: right?
Ben Scoones: At some point. At some point.
Richard Ellis: So I'd love to do, uh, I'd love to dig into the human aspect of AI usage, uh, on another episode, if you're game.
Ben Scoones: Absolutely. Yeah.
Richard Ellis: All right, well, let's wrap up this one.
Richard Ellis: Uh, as, as you know, the way we like to close out the show is just sharing some general goodness that may or may not have anything related to what we discussed today. So in your life, [00:29:00] uh, these days, anything that's brought you a little extra goodness lately?
Ben Scoones: Well, I will make it related to what we talked about.
Ben Scoones: Okay. Fair enough. Um, so. A book that I have been kind of reading all and off recently, it's called Non-Comp Computable You by Robert Marks. And that book is, I think, has been very, um, informative for me thinking about some of these things, thinking about maybe more the philosophical side of ai. It's a little old now.
Ben Scoones: I think it came out in 2021 or 2022, so it was kind of a little before some of these, uh, kind of modern lms, um, became widely available. So some of it is, is maybe a little bit dated in that regard. But I think philosophically it's very interesting and really it just explores that idea of what do humans do that machines can't do?
Ben Scoones: And. It's not just can't do now, but what will they never be able to do? Alright. Um, and that's, that's what the title was all about, non computable. If something is non-comp computable, a computer, can't do it. And that's, that's basically it. [00:30:00] So if anyone's interested in that, I would recommend that. And certainly been helpful for developing my thinking topic.
Richard Ellis: I like it. I'm gonna have to add that to my reading list, so thank you for that. And also thank you for being with us today. Really appreciate your time.
Ben Scoones: Yeah, thank you for having me. It's been fun.
Announcer: Some goodness is a creation of revenue innovations. Visit [email protected] and subscribe to our newsletter.
