# The AI hype cycle and what's really changing - **Episode:** [157](https://agents.truthliesandwork.com/episodes/157.md) (2024-12-26) - **Type:** other - **Tags:** ai, hype-cycle, technology - **Time:** 20:37–26:08 (start 1237s) - **Listen:** [https://truthliesandwork.com/157?t=1237](https://truthliesandwork.com/157?t=1237) --- [20:37] Leanne Elliott: In terms of AI, I've spoken to a couple of tech, tech leaders recently and they seem to have different opinions on AI. I speak to some people who are like, it's going to completely revolutionize the way we work. It's going to make a lot of, of certain types of jobs redundant in terms of they're not needed, not necessarily actual layoffs, Um, and then I've heard other people say it, it, it's a hype, it's another hype cycle. It's, it's gonna normalize out. Where do you sit on the argument? [21:03] Al Elliott: I think that there is AI. Okay. So I think that like we've been going through a hype cycle on AI, which is like everyone thinks it's gonna change the world instantly and then it doesn't. Um, but that way the hype cycle usually works is it, it gets so much attention and investment that some of those investments end up working and paying off. So I, I, so I think we're gonna see more of those things paying off soon. Um, I think there are some places that it's paying off tremendously today, but the challenge we have is that like, you know, even if we had AGI today, um, like general intelligence that could do anything, and let's say it looked like a Slack bot, like you could Slack with some, something, and instead of it being a human being, it's like this AI and you give them tasks and they do stuff. Culturally, it still takes time to get through change. Um, it still takes a lot of time. And, um, I actually like ChatGPT. I'll give you an example. ChatGPT, incredible at helping you with grammar and rewriting stuff. It's so good at this very specific thing. So why, why is anyone writing anything that doesn't have perfect grammar? It's like they have to take it and they have to copy it. They have think to do that. You have to change a workflow, and that is hard, and that takes time. And I do think everyone will eventually get there, uh, but I just think it is like culturally, it's, it's slow. We have this cultural speed that things can go at, and even if there's brand new technology that can change the world, it's just, it always takes time for people to get comfortable with it. Um, there's also gonna be really, really weird failure cases. So I'll give you an example on the support side. I think support's been something that people point at. Yeah. And say like this is probably the most likely area to be impacted by AI first. Because if you're writing in and you have a support question as a consumer, if you get an answer in real time, you're going to prefer that. If the answer is really good, that's going to be great. If it's any hour of the day, that's going to be great. If it's in any language, that's going to be great. All these things like the knowledge base that has the answers that people have trouble working through, or the APIs if it's a software product, whatever. It seems like a pretty automated thing. Like you should be able to do that at some point. The issue is, let's say you have a support organization that is, um, 50 agents effectively. Uh, what happens if you have a huge problem and instead of it really alerting you, you end up with like going from 50 agents to 1,000 agents and you don't even know? [23:38] Leanne Elliott: Yeah [23:39] Al Elliott: Like you could have an outage and think everything's fine. And it's just because there's like, you're infinitely scaling this thing that used to require humans. And so also humans are pretty good at asking pretty nuanced, complex things. I just think that like we probably will use AI. We are certainly using AI at Wistia to help us on the support side. But to the place where you get where there's like no support people seems pretty unlikely to me. I think, I think jobs will change [24:08] Leanne Elliott: Yeah [24:09] Al Elliott: A lot in a lot of places and probably a lot of unexpected places. But I think they're going to change. And I think that's really what that means. Like another example is if you look at graphic design, we're seeing on Fiverr like that less people who are like average or not great graphic designers are not getting as much work because a lot of one-off stuff that you would've needed to go get somebody to help you make this thing with, now you can do yourself. Uh, but the very best graphic designers are, I would bet, gonna end up getting more work because there's more people who, um, actually are able to get some graphic design done. And often that's what convinces you to do even more of it. So there's like that, there's gonna be a lot of, it's definitely gonna change work and it's gonna change jobs, but I'm not convinced it's gonna be super, super fast. I think it's gonna take some time [25:01] Leanne Elliott: Not really. Now you've got the context. He doesn't really think it's all hype, but Chris's point is that whilst we are excited about AI, any change with humans inevitably is going to take some time [25:13] Al Elliott: It's such an important point about culture taking time to change. I think a lot of businesses learnt that the hard way this year [25:20] Leanne Elliott: Yeah, 2024 has been a real year of transition, and with change, of course, comes opportunity. Our next 2 predictions come from Louise Camejo and Amy McCown. Louise is co-host of the runaway success that is Content is Profit podcast and also has a professional podcast studio. That they rent in Jacksonville, Florida. Amy is a leading consultant specializing in workplace mental health who has worked with huge organizations such as the EU Parliament, Samsung, Burberry, EY, and many other FTSE 100 organizations. We'll hear from Luis first. He thinks 2025 is going to be a really good year for business. Amy is also cautiously optimistic and predicts the businesses that finally kick old-fashioned ways of working to the curb will be reborn and thrive