Live from Los Angeles! As Mike records this podcast episode from Anaheim, CA prepping for the California Association of Realtors annual REimagine event, and Paul Roetzer is coming off the heels of a whirlwind trip to Munich for an event he keynoted, they sit down for a quick 56 minutes to regroup on the latest in artificial intelligence, business, and marketing.
Listen or watch below—and see below for show notes and the transcript.
This episode is brought to you by AiAdvertising. Start winning with AiAdvertising’s innovative approach to maximizing budget and performance. Use AI to optimize campaigns by gaining deep customer insights, drawing out motivations and behaviors, enabling intelligent targeting and ensuring messages hit the mark. Stop wasting time, money, and resources. Let AiAdvertising lead while you take the credit! Visit their website to learn more!
00:03:49 — Major service firms pivoting to AI
00:13:21 — AI and its impact on our work
00:26:07 — Where AI is going based on real-world experience
00:38:34 — AI leaders meet behind closed-doors in Congress
00:43:32 — Dreamforce AI updates
00:46:22 — Google Gemini nears release
00:48:27 — Deutsche Bank and generative AI
00:51:08 — Newsom’s executive order on AI
00:52:39 — Stable Audio launches
Major service firms are pivoting to AI
Two big AI announcements from top consulting and advisory firms this week give us a sneak peek into how major services firms are pivoting to AI. The first announcement comes from EY, which announced the launch of EY.ai, a platform “to help clients boost AI adoption,” according to VentureBeat. “The company said it has invested $1.4 billion as the foundation for the platform, including embedding AI into proprietary EY technologies like EY Fabric — used by 60,000 EY clients and more than 1.5 million unique client users, as well funding a series of cloud and automation technology acquisitions. The announcement also included the fact that following an initial pilot with 4,200 EY technology-focused team members, EY will be releasing a secure, large language model called EY.ai EYQ.”
In another announcement, major AI company Anthropic announced a partnership with another heavyweight consulting firm, BCG. The partnership is designed to bring Anthropic’s Claude model to more enterprises. Anthropic says, “BCG customers around the world will get direct access to our AI assistant to power their strategic AI offerings and deploy safer, more reliable AI solutions.” They continue, “Through this collaboration, BCG will advise their customers on strategic applications of AI and help them deploy Anthropic models including Claude 2 to deliver business results.”
How AI is actually impacting how we work
Is AI really that big a deal for the future of work? A new Harvard Business School paper from 9 authors, including leading AI expert Ethan Mollick, answers with a resounding YES. The paper is titled, Navigating the Jagged Technological Frontier: Field Experimental Evidence of the Effects of AI on Knowledge Worker Productivity and Quality.
The paper details a multi-month experiment working with Boston Consulting Group (BCG) to gauge how AI is transforming knowledge work. Mollick says, “There is a ton of important and useful nuance in the paper but let me tell you the headline first: for 18 different tasks selected to be realistic samples of the kinds of work done at an elite consulting company, consultants using ChatGPT-4 outperformed those who did not, by a lot. On every dimension. Every way we measured performance.” The tasks included creative tasks, analytical tasks, writing and marketing tasks, and persuasive tasks that a consultant might be asked to do for a client.
Specifically, Mollick and his team of researchers found that “consultants using AI finished 12.2% more tasks on average, completed tasks 25.1% more quickly, and produced 40% higher quality results than those without.” But the study also showed that when people used AI for tasks it wasn’t actually good at, they were more likely to make mistakes and put too much trust in AI.
The key, says Mollick, is being good at judging when AI is good or bad at a task. Mollick says that some consultants in the study weren’t good at doing this, and suffered in their performance as a result. But others navigated this dynamic well, he says, “acting as what we call “Cyborgs” or “Centaurs,” moving back and forth between AI and human work in ways that combined the strengths of both. I think this is the way work is heading, very quickly.”
Where AI is going next based on real-world experience
Paul just published a LinkedIn post that contains observations on where we’re headed with AI—and it’s blown up thanks to its unique breakdown of the current AI landscape. In the post, Paul says, “I’ve done more than 60 AI presentations this year; had dozens of conversations with enterprise, venture capital, and educational leaders; and fielded hundreds of questions about AI from audiences. Based on those experiences, here are some observations about where we are, and where we’re going.” Listen, then check out Paul’s post linked below.
It’s another exciting week in the world of AI. Listen, subscribe, and we’d love your review!
Disclaimer: This transcription was written by AI, thanks to Descript, and has not been edited for content.
[00:00:00] Paul Roetzer: no one seems to really have good answers for what's going to happen to search. what are we going to need from agencies? But what we know, what we've heard from a lot of these companies is they need help and they need guidance. And there's just very few agencies and consultants capable of providing that general.
[00:00:17] Paul Roetzer: AI knowledge and strategy. So I think that we just need more people that are AI literate and capable of providing this kind of guidance.
[00:00:27] Paul Roetzer: Welcome to the Marketing AI Show, the podcast that helps your business grow smarter by making artificial intelligence approachable and actionable. You'll hear from top authors, entrepreneurs, researchers, and executives as they share case studies, strategies, and technologies that have the power to transform your business and your career.
[00:00:47] Paul Roetzer: My name is Paul Roetzer. I'm the founder of Marketing AI Institute, and I'm your host.
[00:00:57] Paul Roetzer: Welcome to episode 64 of the Marketing AI Show. I am your host Paul Roetzer along with my co host Mike Kaput. I am fresh off of a keynote Friday, last Friday in Germany. It is Monday, September 18th. I'm still trying to get my bearings straight here. So I did a crazy jaunt to Germany. Igot there Thursday at noon local time, did a talk.
[00:01:21] Paul Roetzer: 9 a. m. local time, and I was back home in Cleveland at midnight Friday, so I was traveling for 28 hours in Munich for 26. And here I am on Monday trying to figure out what is going on in the world. You, Mike, are not in Cleveland, if I recall correctly, right? Where are you?
[00:01:38] Mike Kaput: Right, I'm in, just outside Los Angeles in Anaheim for, tomorrow I'm speaking at the California Association of Realtors, annual conference.
[00:01:48] Mike Kaput: So they're doing a lot of content around AI, so props to them. Really excited about it.
[00:01:54] Paul Roetzer: Nice. And I actually have two talks this week in Cleveland. I'm really excited. I don't have to go anywhere, but we have the cozy, innovation conference Tuesday nights. If anybody's in Cleveland wants to grab tickets for that, I'll be there.
[00:02:07] Paul Roetzer: I'm actually going to talk at Baldwin Wallace college on Wednesday. So lots of talks and you you've got another one after this, right? Aren't you somewhere else this week? Going
[00:02:17] Mike Kaput: straight to Dallas from here for later this week, the NIO summit, which is nonprofits and nonprofit serving agencies. So super excited about that one too.
[00:02:27] Paul Roetzer: We should probably share our speaking schedule more. Like we don't talk about it much, but maybe in the newsletter, because I know like a lot of people doing a lot this fall and it might be a chance if people want to like come and say hi, or maybe we'll organize some meetups or something as we're bouncing around, we should probably do that.
[00:02:42] Paul Roetzer: Like just organize some meetups. So yeah, stay tuned. Maybe we'll do that. I'm sure you're listening, Cathy McPhillips on our team, remind us to get some meetups going. All right. So this episode is brought to us by AiAdvertising. Start winning with AiAdvertising's innovative approach to maximizing budget and performance, use AI to optimize campaigns by gaining deep customer insights, drawing out motivations and behaviors.
[00:03:11] Paul Roetzer: Enabling intelligent targeting and ensuring messages hit the mark. Stop wasting time, money, and resources. Let AiAdvertising lead while you take the credit. Visit aiadvertising.com/aipod to learn more. So thanks to AiAdvertising for supporting the episode. All right, Mike, it's an interesting week, like not a ton of like major earth shattering news and AI, but yet some really, really fascinating topics that I think have broader implications to what's going to be coming this fall.
[00:03:47] Paul Roetzer: So let's get started with the main topics. Awesome. First
[00:03:51] Mike Kaput: up, we have two big AI announcements from some top consulting and advisory firms this week. And these give us a sneak peek into how many major service firms are pivoting to AI. So the first announcement comes from EY, which announced the launch of EY.
[00:04:10] Mike Kaput: AI, which is a platform they're building, quote, to help clients boost AI adoption, according to VentureBeats. According to VentureBeat, they say the company said it has invested 1. 4 billion as the foundation for this AI platform, including embedding AI into proprietary EY technologies. They have one called EY Fabric, which is used by 60, 000 of their clients.
[00:04:37] Mike Kaput: And they're also funding a series of cloud and automation technology acquisitions. And the announcement from EY also included the fact that Following an initial pilot with 4, 200 EY team members, EY is also going to be releasing a secure large language model called EY. AI. EYQ.
[00:05:00] Paul Roetzer: That's not easy to say. I
[00:05:02] Mike Kaput: just realized that saying it out loud.
[00:05:05] Mike Kaput: So, in another announcement, major AI company Anthropic announced a partnership with another big consulting firm, BCG. And this partnership is actually designed to bring Anthropic's Claude model to more enterprises. According to Anthropic, BCG customers around the world will get direct access to our AI assistant to power Their strategic AI offerings and deploy, deploy safer, more reliable AI solutions.
[00:05:33] Mike Kaput: Now, Anthropic also says that through the collaboration, BCG will advise their customers on strategic applications of AI, and then help them deploy Anthropic models like Claude to, to actually deliver business results. So we're seeing a couple of really high profile. AI vendor plus consulting firm partnerships here.
[00:05:54] Mike Kaput: First off, is it safe to say that we should expect kind of every major consulting firm to build AI into its client facing
[00:06:01] Paul Roetzer: business?
[00:06:10] Paul Roetzer: Because we had Accenture in June announced a 3 billion investment into their data and AI practice. By the way, Paul Daugherty, the leader of AI practice at Accenture has an amazing book called Human and Machine. So one of our favorites, it's probably like four or five years old now, but. Still a great book.
[00:06:30] Paul Roetzer: So if you're looking for a book recommendation, human and machine is an awesome one. So that was Accenture in June. We had McKinsey announce a partnership with Cohere in July. I know there's been multiple partnerships announced related to OpenAI. So it's just like. Yes, like this is kind of where we're going.
[00:06:47] Paul Roetzer: It's the future of these consulting firms. It's the future of service firms. And I think what we're seeing is every enterprise, the large businesses in particular, they're trying to solve for generative AI, specifically large language models, and they need help. So this becomes a supply and demand thing.
[00:07:04] Paul Roetzer: It's like, who do you turn to, to help you figure this out? So the strategist, the consultants become the people. And then if you flip it, the large language model companies like Anthropic and Cohere and OpenAI, and even like Google and others, they're competing now for the enterprise customers. And the best path to do that is through the trusted advisors, consultants who already have relationships.
[00:07:27] Paul Roetzer: With the brands. So if you're cohere or anthropic and you want to get into these large enterprises, you go through a McKinsey or an Accenture, because they're the trusted advisors to these enterprises, and if you're the recommended large language model for them, then they can wrap services around it.
[00:07:44] Paul Roetzer: And I think we saw, I don't know, probably like five, 10 episodes ago, we talked about McKinsey's large language model offering. And somebody had shared that online. I mean, they're charging like five to 10 million to like build these custom integrated large language models. So this is a massive opportunity for these service firms.
[00:08:00] Paul Roetzer: And then we're seeing not only at the big service firm level, but even down to the smaller marketing agency level or just any service provider or strategic advisor, they're trying to figure out what is the future of their firm. You know, as, as services are evolving, as the needs are evolving, what can they be offering?
[00:08:18] Paul Roetzer: And so large language model strategy and implementation certainly seems to be a good bet for the near future. So, And I just think like the demand is going to keep growing that it's at the higher, the big enterprise now, but it's going to move down market real fast to the point where in the next year, you're going to have small mid sized businesses trying to figure out what are they going to do with language models.
[00:08:39] Paul Roetzer: But these big enterprises have a lot of challenges to implement these things. So again, throw a bit cohere. If you're trying to get a language model into a big enterprise, there's privacy concerns, there's security, there's the fact that this stuff doesn't work like normal software. So it's somewhat.
[00:08:53] Paul Roetzer: Unreliable. We'll talk a little bit about that in one of the next topics. There's the resistance by workforces who don't want to use this technology. Maybe you're afraid of it. There's the rate of innovation. Like you can do build this around a GPT-4. And what happens if GPT-5 comes out or what happens if Google's Gemini comes out and it's better than Chad So there's this uncertainty around the market.
[00:09:18] Paul Roetzer: And then there's the, do we go with a closed model like an anthropic or a coherent, or do we go with a open source model like a llama two for meta? So there's just, this is what these big consulting advisory firms exist to do is to solve these complex business challenges and help brands navigate through what is going to be a very disruptive technology.
[00:09:40] Paul Roetzer: That's going to get introduced into their companies. Yeah, so that
[00:09:45] Mike Kaput: kind of front runs a little bit some later topics we'll talk about but it sounds like what you're saying is that there's not just this whole AI revolution when it comes to, oh my gosh, this technology has broken out, and is really exploding everywhere and it's going to change everything there's such change.
[00:10:02] Mike Kaput: Management involved too that apparently these enterprises are going to be turning to people like EY or BCG to help with. Is that right?
[00:10:10] Paul Roetzer: Yeah, for sure. I mean, there's just there's so many unknowns ahead. And right now, a lot of the enterprises that we talked to that we hear from, they're just aware they need to start solving this and they don't really know where it's going to come from.
[00:10:23] Paul Roetzer: And I think that there is this disconnect where. The more technical side, the CIO, they're working on larger scale solutions for the organization, but like, we'll talk with the marketing department and they have no insight into what's going on with the CIO's office or what's going on at the broader implications across the organization.
[00:10:42] Paul Roetzer: And they're just trying to say, like, we just want to start writing posts more efficiently and writing articles and improving what we're creating and doing transcription and summarization and like all of these use cases that live. Within marketing and sales and service, these dozens or hundreds of obvious use cases, and they don't want to wait around for a year to figure out what to do.
[00:11:02] Paul Roetzer: And that was, that's what then leads like, well, let's just go get a third party software product, like a writer or Jasper or something like that, where we can just turn this on two weeks from now and start. Piloting it in the organization. So that's I think what we're seeing, but it's definitely going to keep evolving to where a lot of these professional service firms are going to be building right now there it's AI practices, but the reality is it's going to be their whole business in the not too distant future.
[00:11:28] Paul Roetzer: Everything they're going to be consulting out is going to be related to AI strategy and implementation. Are there any
[00:11:35] Mike Kaput: lessons here for services firms broadly? I mean, ones that might not be the size of EY or
[00:11:43] Paul Roetzer: BCG. The same, the same needs are going to exist in every company. So we have our AI for agency summit coming up in November, and this is one of the key things we're going to be talking about is just where is the demand going?
[00:11:56] Paul Roetzer: So when I was building PR 2020 through the years and you and I were there together, it was all about like for us, it was all about inbound marketing, content, build your blog, drive traffic to the site through that stuff. Build a social media presence. Like that was the game plan for.
[00:12:13] Paul Roetzer: 10 years, 10 plus years. And now the question becomes like, does that game plan even work moving forward? Like what is organic search going to look like? And so I think a lot of agencies have really built over the last 10 to 20 years around digital marketing and driving traffic to websites and driving leads and converting leads.
[00:12:33] Paul Roetzer: And, and, and the question just becomes like, what does that look like in the future? And so every service from every agency needs to be thinking about that. And then. The client side, you need to be thinking about what do you actually need from your agencies moving forward? And everyone's trying to figure this out at the same time.
[00:12:50] Paul Roetzer: Like no one seems to really have good answers for what's going to happen to search. And what are we going to need from agencies? But what we know, what we've heard from a lot of these companies is they need help and they need guidance. And there's just very few agencies and consultants capable of providing that general.
[00:13:09] Paul Roetzer: AI knowledge and strategy. So I think that we just need more people that are AI literate and capable of providing this kind of guidance.
[00:13:21] Mike Kaput: So along those lines, actually, our second topic is about answering a really big question on everyone's mind. And their question is broadly, is AI really that big a deal for the future of our work?
[00:13:35] Mike Kaput: According to a new Harvard Business School paper from nine authors, including leading AI expert Ethan Mollick, The answer is a resounding yes. This paper is titled, Navigating the Jagged Technological Frontier, Field Experimental Evidence of the Effects of AI on Knowledge Worker Productivity and Quality.
[00:13:55] Mike Kaput: And this paper details a multi month experiment where these researchers worked with BCG to gauge how AI is transforming. Knowledge work. Now I'm going to quote Malik on some of the results here. And Ethan, as a reminder, was a speaker at our marketing AI conference this past year. He says, quote, there is a ton of important and useful nuance in the paper, but let me tell you the headline first for 18 different tasks selected to be realistic examples of the kinds of work done at an elite consulting company.
[00:14:30] Mike Kaput: Consultants using ChatGPT for outperformed those who did not. By a lot on every dimension, every way we measured performance. Now this included tasks like creative tasks, analytical tasks, writing and marketing tasks, and tasks related to persuasion, which a consultant might be asked to do for a client.
[00:14:53] Mike Kaput: So specifically, Malik and the team of researchers found that consultants using AI finished 12. 2 percent more tasks on average, completed Tasks 25. 1 percent quicker and produced 40 percent higher quality results than those who didn't. Now what's interesting here is this study also showed that when people used AI for tasks that it wasn't actually good at, they were way more likely to make mistakes and put way too much trust in AI.
[00:15:25] Mike Kaput: So the key, says Malik, is being good at judging when AI is good or bad at a task. Malik actually says that some of the consultants in the study just weren't good at doing this and they suffered in their performance as a result. Others, however, navigated that dynamic really well. And he says they're quote acting as what we call cyborgs or centaurs moving back and forth between AI and human work in ways that combine the strengths of both.
[00:15:54] Mike Kaput: I think this is the way work is heading very quickly. Now, Paul, there's a lot to unpack there, but what do you see as these findings meaning for knowledge
[00:16:04] Paul Roetzer: workers? I think it's the start of what we're going to see is a lot of more formal research into the impact. So we've theorized about the impact on knowledge work a lot on this show.
[00:16:17] Paul Roetzer: I get asked all the time about the impact on knowledge work, how quickly, how big, and what would, what we've lacked is like the empirical data behind it. So I think it's important because it starts to show what we have. In many cases made assumptions about we're when we look at it when I project out that I think it's going to affect millions of jobs in the next 18 months.
[00:16:43] Paul Roetzer: I try and do it at a micro scale, like I'll sit there and run thought experiments of like, okay, I have a team of 50 writers. I use these tools. It takes the average writer four hours to create an article with these tools. It takes two hours so you run through all these different use cases and then you stack those use cases across an entire department or team.
[00:17:03] Paul Roetzer: And you come away with, well, we don't need as many humans, not that like writing jobs are gone or SEO jobs are gone or email jobs are gone, but like, we don't need as many people to do the same level of output. So the options are, you're going to increase what you do. You're going to just make more of the thing, or you're just going to use less people to create the same level of output.
[00:17:24] Paul Roetzer: Like those are pretty much the only two options. And so I, the way I've looked at all of this is that. Well, each company is going to have a choice. They're either going to produce more of the thing or start a new thing to redistribute people to, or they're going to choose to take the cost savings and just have less people doing the work.
[00:17:42] Paul Roetzer: So I think that studies like this become very important to try and start making. Much more tangible conversations around what is the real impact going to be and to, to expand it. So if you haven't read, I mean, Ethan's post is a great summary. We'll put Ethan's post in, the show notes. The actual study itself is like 58 pages long.
[00:18:05] Paul Roetzer: Great use case to test the anthropic CLAWD2. Upload the PDF to CLAWD2 and have it analyze it for you. But what they did is they took. 758 BCG consultants, so 7 percent of their consulting force, and they were allowed to use GPT 4 with no special fine tuning or prompting and some very basic training, in essence.
[00:18:28] Paul Roetzer: And then they did a bunch of pre testing and surveying to establish baselines. And then they asked the consultants to do this wide variety of work for a fictional shoe company. that BCG kind of devised to represent an actual BCG client. So this is kind of how the process works so you can get a sense of kind of what went into this study.
[00:18:52] Mike Kaput: So if I'm a knowledge worker listening to this or reading the paper It occurs to me, like, how do I actually start putting this into practice in my own career, like, it sounds like this ability to move back and forth between AI and human work in ways that kind of combines the strengths of both is really, really important for careers moving forward.
[00:19:14] Mike Kaput: So, like, what skills should I be thinking about to get good
[00:19:18] Paul Roetzer: at that? Yeah, maybe, I mean, let me go through a couple of the other things that jumped out to me. Maybe we'll back into this answer because it is, I mean, it's the thing we're all trying to figure out is like what matters moving forward. So a couple of things he said that stood out to me is AI is weird, which is something Ethan says all the time.
[00:19:37] Paul Roetzer: That no one actually understands the full range of the capabilities of these large language models because there is no instruction manual. So the testing and experimentation with the tools. So in terms of a skill needed right now, you need the skill of being able to run lots of experiments and being analytical and being strategic and being curious.
[00:19:57] Paul Roetzer: Like, so the reason Ethan can. Talk authoritatively about this is he has spent the last nine months testing these things and building use cases and testing prompts and experimentation. And like, that's it. Like, that's what made him an expert in this thing. And so I think for moving forward, that is one of the most important things is the curiosity and the willingness to experiment and then share with your team what you're finding.
[00:20:21] Paul Roetzer: The other thing he talked about is, again, this. Deep realization that these things are imperfect by nature. So he gave the example of writing a sonnet. It is amazing at writing a sonnet, way better than I would be. But if you asked it to do it in 50 words, it can't because it doesn't actually know word counts.
[00:20:41] Paul Roetzer: It thinks in tokens, like fragments of words. So it's really hard for it to do that. It's also hard for it to do math. But it's really good at ideation. So there's, like, it's the discovery of what are they actually good at. And you can't figure that out without experimenting yourself. The other thing I thought was interesting is they talked about that the consultants Who weren't as skilled, kind of leveled up with this technology more.
[00:21:10] Paul Roetzer: So the, like the most skilled people saw a boost, but it was the people that were kind of probably more like mid level in their capabilities, they saw the greatest increase. So it works as this weird, like skill leveler. Where it really raised, they said the consultants who scored the worst when we assess them at the start.
[00:21:29] Paul Roetzer: So in the benchmarking phase, had the biggest jump 43 percent when they got to use the AI. And so that's, he talks about like, not, not, not. And he says, I do not think enough people are considering what it means when a technology raises all workers to the top tiers of performance. The other thing they talked about is this whole idea of like not knowing what it's capable of and not knowing where its flaws are is the people who came to rely on it too much, what they called kind of falling asleep at the wheel.
[00:21:59] Paul Roetzer: They made more mistakes than normal because they assumed it was good. It's like, I think of this like the self driving car. I know my Tesla makes all kinds of mistakes. So you keep both hands on the wheel and you're paying attention at all times just because it's driving itself doesn't mean you're checking email.
[00:22:15] Paul Roetzer: But if you don't know that you're getting a Tesla the first time and nobody tells you that you're like, Oh, full self driving amazing. You click it, you go and you start checking your phone and whatever. And the thing crashes. You had no idea that it actually wasn't perfect. Like you assumed it's called full self driving.
[00:22:30] Paul Roetzer: It drove you. That's not what it does. So this understanding of the technology becomes really critical. So I just think that where we go in the future and what matters right now, what we know is being able to work with these tools, figure out what they're capable of and then maximize the value you can create.
[00:22:50] Paul Roetzer: With the use cases, it's really good at, there are going to be things it's not good at, and that's the whole jagged frontier is like, we don't actually know what's within its capabilities and what's outside of its capabilities. And so for the foreseeable future. Figuring those out and then applying them in your company saying, okay, it's really good at summarization.
[00:23:11] Paul Roetzer: What are all the places we do summarization? It's really good at transcription. It's really good at writing transcripts like video transcripts, like, or scripts. Like, what are the things it is really good at? And let's double down on those. And I think that's what really the frontier looks like right now and the people are going to get the most value out of these tools.
[00:23:32] Paul Roetzer: Yeah, and
[00:23:32] Mike Kaput: based on my own observation and experimentation people throw around this word intrapreneur to talk about employees that have that kind of business owner mindset and are treating their role as if it is a business with the bottom line and a goal and a mission to meet. And I think that mentality just lends itself naturally to this type of experimentation where if you are viewing things with an intrapreneur mindset, even if you're not an owner of a company.
[00:23:59] Mike Kaput: You're really going to back into this much more easily, I think, where you're going to say, oh, okay, of course we have to run experiments. Of course we have to double down on what's actually driving performance, et cetera. So that's really, really helpful in my experience.
[00:24:11] Paul Roetzer: Yeah. And we do get asked all the time, like, what are the careers of the future?
[00:24:13] Paul Roetzer: And it's not a cop out for me to say, I don't know, go figure it out. Like you got to find your own career path. Like it takes being a domain expert in your field and understanding what these things are capable of. And once you connect those dots, like, now you can figure out what is the career path. Again, is it an AI ops role to help other people figure this out?
[00:24:33] Paul Roetzer: Is it an AI educator where you're teaching people about these things and running workshops? Like, there's all kinds of career paths to be defined, but I haven't really seen a great paper yet that says what those are going to be.
[00:24:46] Mike Kaput: So to kind of wrap this up, I am curious, do you have any thoughts on the flip side of this equation?
[00:24:51] Mike Kaput: What should companies and leaders be looking to hire for as it relates to kind of skills that Malik is outlining here?
[00:24:59] Paul Roetzer: Yeah, I think it does go back to just. You know, seeing what people are capable of and what their comfort level is with these technologies. And I go back to when we were hiring at the agency, all those years, we were looking for good writers and so we would give them writing tests.
[00:25:14] Paul Roetzer: Well, writing tests are useless now. And so now when we hire at the Institute, We still do a writing test, but we also have them do prompting within GPT 4, whatever tech we're guiding them on, and we want to see how the prompts, how they develop their prompts, how they improved on it, how they kind of communicated with the chatbot, and then Like have them compare an actual writing sample versus the machine writing sample.
[00:25:39] Paul Roetzer: What did it do bad? Like you're looking for people who can infuse AI into their work and be comfortable with it and want to seek more knowledge about how to continually improve there. What you don't want to do is hire someone who's like, I don't want to use those tools. Like, okay, that is, that's just not going to fly.
[00:25:55] Paul Roetzer: So right now, I mean, you're really just looking for people who are curious and motivated to solve for the future and help kind of figure it out together, what it looks like.
[00:26:06] Mike Kaput: So speaking of what the future looks like, our third topic today actually concerns you Paul, because you just published a LinkedIn post that contains observations on where we're headed with AI.
[00:26:18] Mike Kaput: And it's really blown up, thanks to kind of its unique breakdown of the current AI landscape and what that means for the future. So in the post you wrote, quote, I've done more than 60 AI presentations this year, had dozens of conversations with enterprise venture capital and educational leaders and fielded hundreds of questions about AI from audiences based on those experiences.
[00:26:43] Mike Kaput: Here are some observations about where we are and where we're going. So my first question is. Where are we and where are we going? Can you break that down for us?
[00:26:51] Paul Roetzer: So, I don't know if people find this interesting or not, but the context around this post, because some people, like, maybe they think I wrote this with like, ChatGPT or something.
[00:26:59] Paul Roetzer: Again, I don't use AI writing tools to write anything for me. So, this is all me. And I had gotten back from Germany at midnight Friday, as I mentioned to start. So, I'm like, jet lagged and we had just got new cats and one of my cats had to go to the vet. And so, I'm Saturday morning at 9 a. m. Sitting, waiting, it ended up being an 11 hour wait, it was kind of a crazy day.
[00:27:20] Paul Roetzer: But I'm sitting in my car with only my phone, and on my computer I have nothing. And I was like, I should, like, I'm thinking about the trip to Germany and what just happened, and like the conversations I was having with some major brand at this event. And, and I was like, what do I, like, have to say about this?
[00:27:34] Paul Roetzer: Like, what do I actually, have I learned? And I started actually, instead of writing about the Germany thing, I started reflecting on the entire year. And all of these conversations I've had with just, Some amazing people, amazing leaders, amazing companies, but educational leaders, venture capital firms, enterprise, like all these different organizations we've talked to, small businesses, large businesses.
[00:27:56] Paul Roetzer: And, and I just started kind of writing, like, what have I, what do I think is happening right now? Like, what is the state of all of this? And so I literally wrote this thing on my phone. And I was like, Iguess I'll just publish this. I don't know. And so I put it on LinkedIn. And that was it. Like it was not some crazy thing.
[00:28:11] Paul Roetzer: I've been working on this forever. It was just sort of spur of the moment. I had nothing to do because I was just sitting with my cat waiting 10 hours to see the veterinarian. So Ikind of summarized them to three points. So the first was The vast majority of organizations are just getting started, even the organizations, and this is critical, even the organizations that have been doing AI for the last 10 to 20 years.
[00:28:35] Paul Roetzer: Think of like a Google. They have been doing AI for 25 years. Literally, they just turned 25. So the companies that have been using machine learning for the last decade or more to do predictions, recommendations, build personalization, even those companies that were on the frontier. We're not prepared for generative AI and the transformative impact that it's having across every industry.
[00:28:58] Paul Roetzer: So that was like the first observation is even these companies that people think are the leaders in this space were caught flat footed when it came to generative AI. And most of them that I talked to are still playing catch up. Like they still haven't solved for exactly what to do about this. So that led.
[00:29:15] Paul Roetzer: To the second thing, most organizations, including some of the largest enterprises in the world, are still thinking about AI as a technology challenge and opportunity. They are not addressing or even considering the significant near term need for comprehensive change management throughout the organization.
[00:29:32] Paul Roetzer: So again, they're coming to us saying, which tool do we get? Which use case do we solve for? And I'm coming back and saying, have you thought about the impact of this on your organization? Like again, you I'll, I'll show that, Microsoft 365 co pilot video when I do talks, like the minute and a half demo of what co pilots going to be and how it's going to be infused into docs and, or Microsoft Word and Excel and Teams and all these areas.
[00:29:57] Paul Roetzer: Like we're just all of a sudden, everybody's got AI, the accountants, the lawyers, the HR professionals. Yeah. And I'm like, are you even thinking about that? And the answer is no, they're not. Like they don't really grasp that it's just like this completely disruptive technology, like general purpose technology across their entire organization.
[00:30:15] Paul Roetzer: It's not just use cases in tech, in marketing and sales and service. So that leads to the third thing, which is there's a lack of urgency to solve for this wider impact. And so I was like, why, like, why aren't people. Thinking about this at a higher level than just tack and use cases. What I have arrived is like the leadership at most of these organizations don't have a baseline understanding of what AI really is.
[00:30:40] Paul Roetzer: And the effects that it's going to have across workforce and operations and tech stacks and culture and products and services and their ability to compete. So it always comes back to the education and training. There's, there's a lack of like baseline understanding. And this is what I see over and over again, because the talk I do most times the keynote is, is some variation of an intro to AI.
[00:31:03] Paul Roetzer: Like, it's like, what is it? What are the, what is language? What is vision? Where are the use cases going to live? What does this mean to your organization, to your industry, to your people? Like, it is not deep in the weeds. Like, it's just high level. And people just, like, they don't even know what to say after the talk.
[00:31:19] Paul Roetzer: Like, they'll come up to me, like, I had no idea. Like, this is... We had been talking about AI. We'd been thinking about it. We were doing some of it. We actually, I've talked with people who have AI centers for excellence, like in their companies, like big companies that weren't thinking about the implications, like in this sense, like this broad business implication.
[00:31:39] Paul Roetzer: And so like that. To me is the biggest thing is like, you go do this, talk to a groom of 500 people and they're just staring at it. Like what in the world? And then they they'll come up to me and say these things. Like, I had no idea. Like that was the craziest thing I've ever seen. And to you and me and other people who live in our bubble.
[00:31:58] Paul Roetzer: It's like, really? Like, is it like, I feel like I should stop doing this talk and I should move on to the next talk. And then I realized like, no, people still need this like future of business intro level talk, because that's where the majority of the market still is.
[00:32:14] Mike Kaput: Now, I know you unpack that quite a bit, but I want to dwell.
[00:32:18] Mike Kaput: for a minute on it still this idea that the vast majority of companies are just getting started with AI because I talked to so many people who still find it totally shocking even when we're talking because their company needs help getting started with AI. So Often, I think anyone in the audience or that has followed our work sometimes thinks their company is very, very far behind and everyone else is ahead.
[00:32:44] Mike Kaput: Can you talk a little bit more about that? Because I think there's a disconnect in people's minds. They say, I hear so much about AI. I see all these AI influencers giving advice. I hear everyone, giving talks about it, publishing articles about it at these companies. Of course, they have to be far ahead, right?
[00:33:03] Paul Roetzer: Yeah, no, I think it's critical because if. Like I'll often ask upfront who's tried ChatGPT. It seems like such an obvious thing. We ask it on our intro to AI for marketers free class that I do every few weeks on zoom. And we've asked it in a survey and it's like 95 percent of people have used ChatGPT.
[00:33:20] Paul Roetzer: You ask that same question in a room of CEOs or HR professionals or accounting professionals, you get like five people raising their hand. So yes, there is a. We have, what, 55, 000 subscribers at the institute right now. So you broadly ask all of those people, it's going to be 90 percent plus, and they all think that they're behind everybody else, but you ask non, like.
[00:33:44] Paul Roetzer: Yeah, at the frontier level marketers about this stuff, like business executives, C suite people, maybe they've tried it, like, but they very rarely do you find the people who are experimenting with it or infusing into their workflow daily. And so I do think that. There is within the world that we live in every day, this assumption you're just behind everybody else, that your company's behind because you don't have an AI council or you don't have generative AI policies, or you haven't formally figured out pilot use cases, or you haven't found a new AI tech and you think you're just like way behind.
[00:34:19] Paul Roetzer: That is not the case. And it is like, I've talked with a lot of software companies and you would think SAS companies would be the first in line that would be figuring this out. They're not like, there are a lot of SAS companies that are behind. There are a lot of venture funded companies that are behind.
[00:34:36] Paul Roetzer: So when you get into like healthcare and financial services. Manufacturing, everybody is at the starting point. And I just, I think it's so important because the opportunity is so significant if you figure this out in your industry. And I think the other thing we always talk about is I don't care if you're the intern or a marketing manager or an HR manager or a CMO, like it doesn't matter what your role is.
[00:35:02] Paul Roetzer: Every book, every organization needs people who go and solve for this and then bring that knowledge back to everybody else. And so I think from a career perspective, it's such an exciting time when you realize how important this topic is to the future of business. And that there are very few people capable of connecting the dots in organizations and that you can be one of those people.
[00:35:25] Paul Roetzer: That's a really exciting thing to, to know. And so that, I think, was like I kind of tried to end that post on a very positive note, which is... Like people who are curious and seek this broad knowledge and understanding will be able to connect the dots in their companies and their industries to build smarter companies.
[00:35:43] Paul Roetzer: And then in the process, like take advantage of these kind of unparalleled opportunities we have. And, and like, I mean, there was some negative comments on there. It's like, fine. I like, cool. Like I get that some people don't agree with that premise, but I believe quite deeply in it. Every, every indicator we look at appears to agree with this is massively disruptive.
[00:36:04] Paul Roetzer: People who compare it to the metaverse or blockchain, that's just a joke. Like they're, they're not the same thing. So I think we're entering a phase where it's really hard to dispute that AI is going to transform every aspect of business and every knowledge workers profession. And, and the sooner we kind of accept that and figure out what to do about it, the better off you are all going to be individually.
[00:36:26] Paul Roetzer: And we are going to be collectively. Like as a society. So yeah, I mean, I'm glad the post resonated with people and got people talking, whether they agree or disagree, like good, like let's have conversations, let's figure it out. But that goes back to the importance of that Ethan Mollick talk topic. We started with is we need more data.
[00:36:43] Paul Roetzer: Like we need more proof of this working. Can you more people sharing what they're doing and what they're seeing? That's awesome. So
[00:36:51] Mike Kaput: before we dive into rapid fire, can you just really quickly talk to us about the change management component of what you posted? You said that most leaders aren't thinking about it.
[00:37:00] Mike Kaput: They're just thinking about the technology and what tool to use, like what should they be thinking about when it comes to change
[00:37:05] Paul Roetzer: management? Yeah. I mean, like simply, let's say you get an AI writing tool. So great. You go, you start there. Well, everybody in your organization writes. So you have marketing, sales service, HR, finance, legal ops, everybody, everybody writes emails, proposals.
[00:37:19] Paul Roetzer: articles, whatever it is, job offers, job descriptions. So if you, if you go get a writing tool, awesome, you have the tool. What are they supposed to do with it? Like, how does it change their workflow? How does it change their job description? How does it change how they, they train and educate the next generation of talent?
[00:37:39] Paul Roetzer: Does it affect your overall tech stack? Do you need five of the other tools that you had that you're paying 10, 000 a month for if this tool replaces those? It's just, it's so transformational across every aspect. That you can't just get a tool and start a pilot project and then say, okay, good, we, we did AI.
[00:37:55] Paul Roetzer: No, you didn't. Like you just threw a grenade into it. Like now nobody knows what their job is and they're afraid this thing is going to do their job for them and they have no training of how to use it properly. And so that like change management with any technology is critical, but with a general purpose technology that can be used by anyone in the organization and affects what they do and how they do it.
[00:38:17] Paul Roetzer: There's a lot of other work that comes with. Doing that the right way. So that's the change management is like realizing this changes everything. And you have, you have to plan for that and you have to help your team along in that process.
[00:38:33] Mike Kaput: All right, let's dive into our rapid fire topics this week. So first up a big meeting on Capitol Hill.
[00:38:40] Mike Kaput: So tech executives, including Sam Altman, Elon Musk, Google Sundar Pichai, Mark Zuckerberg. Bill Gates and NVIDIA's CEO Jensen Huang just met with U. S. lawmakers in a closed door forum organized by Senator Chuck Schumer to discuss AI regulation. This was the first in a series of meetings aimed at educating lawmakers about AI amidst the rapid developments happening in the field.
[00:39:08] Mike Kaput: The executives reportedly pushed differing agendas in the room. There was disagreement over issues like open source AI development, and this was all according to the New York Times who talked to sources that claim to be in the meeting. Musk, in particular, warned of existential risks from AI. He said, quote, if someone takes us out as a civilization, all bets are off.
[00:39:31] Mike Kaput: Pichai and Zuckerberg highlighted AI's potential benefits as well as the need for Transparency. And overall, some lawmakers criticize the open the closed door nature of this forthough Schumer said that it encouraged open debate. So Paul, on one hand, I can see some people seeing this as maybe a photo op, a PR move, but on the other, the sheer firepower of the feet, the people in this room is just incredible.
[00:39:57] Mike Kaput: Like, how important is this meeting? Do you have any speculation on maybe what else might have been discussed or what could come out of it?
[00:40:05] Paul Roetzer: There was very little released about it. I was actually kind of surprised how little is known about what was actually said. I haven't heard any like murmurs about how different people interact.
[00:40:17] Paul Roetzer: There's a lot of big personalities in that room. Some of them don't get along very well. So just like nothing, like there really wasn't much to go on. So, I mean, I think it's... It's important that conversations are happening. I did see just in the last like 24 hours some additional, things starting to bubble up about the executive orders.
[00:40:40] Paul Roetzer: We were sort of promised earlier in the summer. So if you remember President Biden's office said that there would be some executive orders related to AI that would be coming this summer. If I'm not mistaken, summer ends in like four days, I think fall's official. So I don't know if it's actually coming in the next few days, but I have heard, some stuff starting just through like Twitter threads and stuff that, the, the executive orders might actually be coming and we have no idea what those are.
[00:41:08] Paul Roetzer: So just something to keep an eye on. I do think that we will see some movement this fall. Again, I don't expect any laws or regulations or anything would. Like crazy to happen, but executive orders are a possibility, but we have no guidance of what those might be. The second thing is I watched this amazing presentation from Bill Gurley.
[00:41:26] Paul Roetzer: So he spoke at the all in event last week. I think it was last week, right? The all in podcast, but there's some of it on, and he had this incredible, talk called 2, 851 miles. And I would recommend watching it. It's about 35 minutes long. We'll put the link in the show notes. But they published the full video online, and it's all about regulatory capture.
[00:41:50] Paul Roetzer: And, and this is the thing that really worries me about the direction this is going. Regulatory capture is in essence, the tech companies themselves influencing what the regulations are. And what tends to happen, which Bill highlights in great detail and very humorously, is that when regulation comes, It is the big incumbents who win almost every time.
[00:42:15] Paul Roetzer: So the fact that the people in the room are Google and Meta and OpenAI and like, Microsoft and all the big players. It, they're trying to influence regulation. They're, they're claiming they really want regulation. And as Bill highlights, when big companies ask for regulation, they're just accepting it's going to come and they want to control what it is.
[00:42:40] Paul Roetzer: And so you can certainly look at all of this and say, this isn't going to work out. Well, this is not going to benefit society in the end. It's going to benefit the big players. They're just going to make more money as a result of whatever. The regulations are, and so this is, this is a non political thing for me, like I'm, I'm not saying this from any side of the aisle, I'm just saying like overall, I'm just really skeptical that whatever the government does is going to end up making a major impact in a positive way, I think it's just a really messy thing for the government to have to solve right now.
[00:43:18] Paul Roetzer: But I guess we need as many conversations as possible. I just hope that it isn't a situation where the big companies end up dramatically influencing what the regulations are. And it most likely I assume would head in that direction.
[00:43:33] Mike Kaput: So another news dreamforce Salesforce is mega popular annual conference wrapped up last week.
[00:43:39] Mike Kaput: And AI played a big role in the product updates that were announced. The star of the show was Salesforce's new Einstein Copilot, which is bringing generative AI to every part of the CRM. Salesforce also rolled out Einstein Copilot Studio, which allows companies to customize their Copilot instance and build AI apps on top of it with no.
[00:44:04] Mike Kaput: Code the studio. Interestingly also includes a prompt builder, a skills builder that gives you control over where gen AI is used in your organization and a way to build models. Salesforce also talked up its quote Einstein trust layer, which is a foundational layer of security and privacy features across all of its AI products.
[00:44:30] Mike Kaput: And these features basically make sure that the models don't train on customer data, and there's also features to put guardrails and controls around the instances of co pilot and Salesforce's AI being used across companies. So Paul, this comes on the heels of HubSpot announcing its AI roadmap inbound a couple weeks ago.
[00:44:50] Mike Kaput: What jumped out to you as important for business and marketing leaders
[00:44:54] Paul Roetzer: to understand here? Yeah, I mean, I think most of what they've been announcing, we've been hearing about all year. I mean, there was announcements going back to like February and March of both HubSpot and Salesforce that they were going to be building this stuff.
[00:45:06] Paul Roetzer: And I think the big question on everyone's mind right now is like, okay, when is this stuff going to be available? Like everybody just keeps talking about these things and debuting that they're going to be building these things. And Ididn't see timelines of when all these features are rolling out within Salesforce, but I expect that that is the thing.
[00:45:21] Paul Roetzer: A lot of people are most interested in that and the pricing of these things. I do, I'm intrigued to look into the ability to build apps, the no code stuff. Like, I think that's going to be really important moving forward is that the non developers like you and I can actually go in and build stuff. And it does seem like AI is going to enable that.
[00:45:40] Paul Roetzer: I know Rept we've talked a lot about more of a startup, well, they're not really a startup, but , more of a VC funded organization that's enabling these kinds of things. They want to enable a billion developers kind of thing. Hmm. So, yeah, I think to me, those kinds of tools are gonna be critical.
[00:45:58] Paul Roetzer: And then just seeing this stuff work, like Ithink. Now with Microsoft and Copilot and, Google Workspace and HubSpot and AI. And it's like, we all just want to see if this stuff is going to do what it's promised to do. Because right now we're just theorizing about how disruptive this is going to be.
[00:46:17] Paul Roetzer: And it takes getting this in people's hands for us to realize, is it actually going to be that way now?
[00:46:24] Mike Kaput: So speaking of getting things into users hands, Google is actually nearing the release of its conversational AI system, Gemini, which is designed to compete with ChatGPT. So this comes from reporting from Reuters and the information.
[00:46:40] Mike Kaput: Gemini is a collection of models that will power search chatbots and content generation across text images, code, and other applications. So Google actually plans to make Gemini available to companies through its Vertex AI service, which we've talked about previously on this podcast. Apparently a small group of companies reportedly already has access.
[00:47:04] Mike Kaput: To Gemini, so no exact date on when this is rolling out, but it sounds like it is meant to be Google's GPT 4 slash ChatGPT. How big a deal is Gemini in your mind, Paul, like both for the overall AI landscape, but also for Google's competitive position in AI.
[00:47:22] Paul Roetzer: So little is known about it. I mean, everything that we're hearing here, we've seen different things mentioned by Google over the last few months, like Demis Hassabis has done interviews where I think they've mentioned Gemini.
[00:47:32] Paul Roetzer: And so there's some indications about what it is and how big it might be. I, it seems like it's generally accepted that it will be bigger and more powerful than GPT-4. I think that's the bet that Google has to make is they have to come to mark with something significant. As we talked about last week's episode, they obviously have the data to train something bigger.
[00:47:53] Paul Roetzer: They have the computing power to train something bigger. They have chips to train something bigger. So I think that, it's critical to Google that this is. Way better than what we've seen from Bard so far and I would imagine that they're putting all of their resources behind making it such so I see this as a very, very important and it sounds like.
[00:48:18] Paul Roetzer: There's a chance we might, we might see the first glimpses of it this fall. But again, these are just reports from the information Reuters, nothing confirmed from Google yet.
[00:48:29] Mike Kaput: So investment bank Deutsche Bank is aggressively experimenting with AI to transform its business. They're apparently running 25 AI pilots that are going to be launched in 2024 across HR.
[00:48:43] Mike Kaput: Tech and investment banking. So this is one big example of this AI powered digital transformation that we're seeing across industries for the bank. Some of the use cases include automated client briefings and a chat bot to handle corporate banking inquiries. Now, in one use case. The bank actually demoed an AI system that could use internal data to perform the role of a junior banker, and they used dummy data from a fictitious company to test this out.
[00:49:13] Mike Kaput: Now, according to Business Insider quote, the tool presented a client briefing, a report to prep investment bankers before generative AI, what would usually take a team of junior bankers a day or two to put together was produced in seconds. The bank has said it aims to double or triple its AI related workforce.
[00:49:35] Mike Kaput: Business Insider says about three to 400 people currently work on AI related projects. And they also indicated they would retrain existing employees who are affected by this technology. So Paul, we've talked at length about consulting and advisory services. This is AI?
[00:49:58] Paul Roetzer: Yeah, I mean, it's going to have a massive impact on financial services, but I just think this is so representative of what we talked about earlier, like, this is what we're saying, like, this is change management. This isn't like a tack in a pilot project. This is like, oh, wow, this changes what these bankers do.
[00:50:12] Paul Roetzer: It changes their future of knowledge work in our business. And the fact that they took a very specific thing, ran tests, like analyzed the impact it was going to have, this is what companies need to be doing. Like I'm not a big supporter of Deutsche Bank, like I, I, nothing about what, what else they're doing, but like, this is the kind of model you should be following within your organization.
[00:50:31] Paul Roetzer: Pick a role, pick a job, pick a team and run the experiments yourself. Don't wait for industry reports to come out or us to say. It's 20 to 30 percent and run with that in your deck for your CEO. Go run experiments yourself and figure it out. And you may see that 20 to 30 percent is conservative. It may actually be 70 percent or 90%.
[00:50:52] Paul Roetzer: You have to solve this stuff. And so this is again, what I think every organization should be doing. Don't believe the report. Don't do it yourself and figure it out and then start doing that across your entire organization. So yeah, I mean, I just think it's cool that we're seeing this kind of data share.
[00:51:10] Paul Roetzer: So California
[00:51:11] Mike Kaput: Governor Gavin Newsom just signed an executive order to get state agencies ready for AI. And that order includes directives to work on things like study the risk AI poses to California's energy infrastructure, create state guidelines for procuring generative AI tools, and provide training, AI training for state government workers.
[00:51:32] Mike Kaput: Now, Newsom's deputy comms director told Politico that the executive order is the state's first step toward understanding how to govern AI. Now, Paul, this seems like a significant step forward, but it's definitely one of the first we've heard among state governments that seems to be this extensive. How fast do states need to be moving on this issue?
[00:51:53] Paul Roetzer: Yeah, we, we've talked about this before when we were doing, I think it was like around the 40s or 50s episodes, we were talking a lot about like laws and regulations, and the assumption we were making, and that seems to be holding true, is like this, the states are going to move faster than the federal is going to move, and so while it's interesting to watch what's happening on Capitol Hill and what executive orders may come, I do think that watching the state level is going to be a better indicator in the near future.
[00:52:19] Paul Roetzer: And certainly doesn't, it makes sense that California is kind of one of the first movers in this space. So yeah, I mean, I think it's really important to keep an eye on this and you're going to have, especially for the bigger enterprises that, do business in California. Like you need to have people in your organization are paying attention to this stuff and thinking about what, how it's going to impact how you use AI moving forward.
[00:52:40] Mike Kaput: So last but not least today, Stability AI, the company behind the popular Stable Diffusion Image Generation tool, has just released Stable Audio, which is their first AI product that generates music and sound. So now anyone can use this to generate short audio clips using just text prompts. And interestingly, users are not going to be able to ask the model to generate music in the style of a real popular artist like...
[00:53:08] Mike Kaput: You can't say, give me music. That sounds like the Beatles. Now, stable audio has a free version and a 12 per month pro version in the free version, you get 20 audio generations per month of tracks, and the tracks can be up to 20 seconds long. The pro version gives you up to 500 generations and up to 90 seconds in a single track.
[00:53:30] Mike Kaput: So Paul, I want to ask, it seems like. AI generated audio is a big area to watch moving forward. Like how big a
[00:53:36] Paul Roetzer: deal is a release like this? There's gonna be a bunch of these offerings for sure. We've talked about generative AI when we're thinking about it. There's lots of categories, but we think in terms of.
[00:53:48] Paul Roetzer: Text or language, images, video, audio, and code. And so what we've said is like last year was breakthroughs in image generation was the big thing, and certainly the language side, specifically with text tools, and then this year with runway, we started seeing major breakthroughs in the video side of things.
[00:54:08] Paul Roetzer: And audio also seems to be having it's moment this year. There's been breakthroughs announced from meta Google, stability. So a lot of the AI research labs are working on the generation of audio. So I, again, I don't hear this and think, Oh, I stable audio is the only option. But it is representative of the breakthroughs that are happening.
[00:54:28] Paul Roetzer: And I think audio is going to be something whether you're a marketer or you're, some aspect of an organization or you're a producer, a creator, like whatever you're doing, just assume that AI is now capable of helping you create background music and other audio files for your work.
[00:54:46] Mike Kaput: Yeah. And like we mentioned in respect to video generation tools, if you go and try one of these tools and you get a few seconds of audio and you're like, this sucks.
[00:54:54] Mike Kaput: Like That what's all the big deal or the hype about, I mean, yeah, wait, wait a few weeks, wait a month because these things move so fast. It's going to be very, very rapid innovation. I would say moving forward.
[00:55:06] Paul Roetzer: Yeah. They seem to have cracked the code on how to do this. It's really just now a question of like the applications and the companies you use, but yeah, go try a few of them.
[00:55:14] Paul Roetzer: It's a, if this is something you need in your daily workflows, there are tools now to help you do this.
[00:55:21] Mike Kaput: Awesome. Paul, appreciate you breaking down this week in AI for us. I know we really appreciate the insight and the time you take kind of understanding all these topics.
[00:55:31] Paul Roetzer: Yeah. Safe travels. Good luck with the talks.
[00:55:33] Paul Roetzer: We will, reconvene in Cleveland someday, although I think I'm on next week. So I don't know when we'll see each other again, but we'll do it, virtually next Monday. All right, everyone. Have a great week. Thanks Mike, as always for putting this all together, we will talk with you soon
[00:55:47] Paul Roetzer: thanks for listening to the Marketing AI Show. If you like what you heard, you can subscribe on your favorite podcast app, and if you're ready to continue your learning, head over to www.marketingaiinstitute.com. Be sure to subscribe to our weekly newsletter, check out our free monthly webinars, and explore dozens of online courses and professional certifications.
[00:56:09] Paul Roetzer: Until next time, stay curious and explore AI.