Episode transcript The transcript is generated automatically by Podscribe, Sonix, Otter and other electronic transcription services.
Hi everyone, here is Ricardo Vargas, and this is the 5 Minutes Podcast. Recently, one of the things that has stuck in my mind is when meta, the fundamental research area of meta, released the self-thought evaluator. And this was part of several research that they were doing. But this one raised my attention because it talks about something that I think is absolute. It's a disruption inside a disruption, which is the concept of using synthetic data to train large language models. So instead of using, I would say, real images, images produced by humans, uh, or text produced by humans, it will generate new text based on text, for example, created by AI. And why this is so relevant and disruptive for all of us. Because most of us said that one of the biggest assets in the Project Management field is our ability to understand previous experiences, estimate future projects, estimate future risks, and estimate and generate lessons learned. So, when you start creating lessons learned that are based on a project that did not exist, this is, of course, scary. It looks like magic, but it's much closer than what we think. My guess? Of course, I'm not an ultra expert in this type of development, but I would guess that probably one year, one year and a half from now, we will start seeing, for example, um, insights on lessons, insights on risks that are not based on human experience because, for example, if you ask me, Ricardo, let's do a nominal group technique, or a Delphi to create, for example, a list of risks in a part of our project.
One of the key things that is, I don't know, my competitive advantage or any project manager's competitive advantage is your previous experience, right? You managed a project five years ago, and then you learned this, and you managed another one three years ago, and you learned that. But imagine now you have a project generating information on top of something that never, ever existed. It's just a combination of algorithms and neural networks that generate. And the most important, these insights and these results are not generic, you know, and not generic, because most of the time when we use, I would say this, traditional models, we receive an okay, a decent list of risks, for example, if it's a risk or lessons learned. But they are more generic, as I said many times, like pasteurized. But when you use this kind of model, you can have insights on the top of something absolutely new. And I need to be honest, as many times when I talk about artificial intelligence, I have mixed feelings. At the same time, I think, wow, this is fantastic because this may improve a lot. But on top of that, this is absolutely disruptive because this will reduce and may reduce the competitive advantage of many organizations because what is your competitive advantage is probably the knowledge you have, the lessons learned; you have what you learn.
But when this becomes synthetic, what will it be? For example, let me talk about myself. What will be my competitive advantage? If a computer can have insights based on experience that did not exist but that is very relevant. So this is something it looks like we are talking about, Matrix or something sci-fi, but it's not. It's much closer than we think. It will mean that when we use this, we will have endless possibilities. So instead of you creating lessons learned on top of your 100 projects, a thousand projects, and 500 projects, you have unlimited projects that did not even exist, but they are just emulations on top of emulations to generate lessons learned. And then not only lessons learned, but this is really something that will be extremely disruptive. And this is absolutely close to what we are talking about now. If you want and have more interest, you can check the original paper at Cornell University talking about how they built this competitive model. And it's really, really something that is in the cutting-edge frontier of how we manage and think about that data. Think about that, and I hope you enjoy this podcast. See you next week with another 5 Minutes Podcast.
