0:33:15 - (Reg Prentice): So first, having your email organized by project in a database-driven system like Tonic is a huge step forward, because without having emails organized by project, asking AI to answer a question is going to be very difficult because it's not going to know whether this email relates to a school. Well, I mean, at some point, presumably it will, but at the moment it needs some context around it so we're increasingly making Tonic have more context. 0:33:49 - (Reg Prentice): And I think search like within the context of a project AI can make search more interesting because you don't have to know the exact word. You can ask for a concept and have it bring that back. But that has to be still bound by project, and it still has to respect the security, like the role-based security we were talking about before. Like if someone has filed an email confidential in the project, AI also can't reveal the content of that email. 0:34:23 - (Randy Wilburn): Yeah, so that's a good point. 0:34:25 - (Reg Prentice): In an enterprise sense, it's trickier than in a personal sense. If you just want general information, like say you're interested in Greek mythology, chatGPT is incredible but when the information store is an enterprise information store, there are a lot more overlays in terms of how that information relates to the person asking for it. We're not jumping into the AI bandwagon. We're looking at how we can do more of what we do well, which is to organize information in a way that will benefit AI at some point for the firm. 0:35:04 - (Randy Wilburn): You have to think with AI in mind now as you develop and you start to process things. I'm doing that now and just thinking about what GPT five is going to look like, because that's supposed to be coming down the pike in the next seven to twelve months and as they say, it's going to be a game changer and I can only imagine. 0:35:27 - (Randy Wilburn): I won't call them large language learning models, they're small language learning models, but I'm creating them just on my information. And for me, it's perfect because then I get to process this information in a different way than just the way that you would normally search for data. And so it's kind of amazing, to be honest with you. I'm using Google Notebook which is like my sandbox for playing with language learning models to see what they're like. But I'm in the process of creating one for my website where I have five years' worth of podcast data worth of transcripts that I want to make searchable so people can find out any information that they want. And eventually, at some point in time, I want to be able to do it for the Zweig Letter podcast. So stay tuned, listeners. We will have an ability for you to search because, I mean, I have some amazing interviews from this podcast dating back to 2016 when Mark Zweig and I first started sitting down and recording it. So all kinds of information is valuable to people and the end user. And so I think AI is going to make a difference for us all. I'm stating the obvious, but it's just going to be interesting to see how it plays out in different pockets and so in the pocket that you're in with TonicDM and how that impacts the day-to-day workflow of a design professional, I think it's going to be interesting to see how that plays out. 0:36:58 - (Reg Prentice): Yeah, I would say I'm fairly opinionated on that. I wrote a blog post for our TomicDM.com. I'd say it's emerging, but we have some content we're putting there. So for practitioners who are thinking about how AI is going to affect them as designers. I always fall on
Made with FlippingBook Annual report