Marketing AI Institute | Blog

[The Marketing AI Show Episode 82]: Are Paid AI Tools Worth It?, Big Google Bard Updates, and Big Tech’s AI-Driven Quarterly Earnings

Written by Claire Prudhomme | Feb 6, 2024 1:15:00 PM

In Episode 82 of The Marketing AI Show, discover the newest advancements in AI, where certain updates may feel like they are directly from a Sci-Fi movie. Hosts Paul Roetzer and Mike Kaput explore the rationale behind paid AI tools, the newest updates from Google Bard, and Big Tech's AI-driven quarterly earnings calls.

Listen or watch below—and see below for show notes and the transcript.

This episode is brought to you by our sponsors:

Today’s episode is brought to you by Marketing AI Institute’s AI for Writers Summit presented by Jasper, happening virtually on Wednesday, March 6 from 12pm - 4pm Eastern Time.

Following the tremendous success of the inaugural AI for Writers Summit in March 2023, which drew in 4,000 writers, editors, and content marketers, we are excited to present the second edition of the event, featuring expanded topics and even more valuable insights.

During this year’s Summit, you’ll:

  • Discover the current state of AI writing technologies.
  • Uncover how generative AI can make writers and content teams more efficient and creative.
  • Learn about dozens of AI writing use cases and tools.
  • Consider emerging career paths that blend human + machine capabilities.
  • Explore the potential negative effects of AI on writers.
  • Plan for how you and your company will evolve in 2024 and beyond.  

The best part? Thanks to Jasper, there are free ticket options available!

To register, go to AIwritersummit.com

This episode is also brought to you by our brand new Piloting AI 2024 course.

Piloting AI 2024 is a collection of 18 on-demand courses designed as a step-by-step learning path for beginners at all levels, from interns to CMOs.

More than 800 learners registered for the inaugural 2023 edition. The fully updated series, including a new Generative AI 101 course, is available on pre-sale now, and the full course launches this week on Thursday, Feb. 8.

Go to PilotingAI.com to learn more.

Listen Now

Watch the Video

Timestamps

00:05:18 — Are paid AI tools worth the $30/user/month cost?

00:16:33 — Google Bard just got some more big updates

00:27:47 — Apple, Alphabet, Microsoft and Meta’s AI-Driven Earnings Calls

00:42:25 — Apple Vision Pro launches

00:49:25 — Who Owns Your Voice?

00:53:51 — The hottest new job in corporate America? Chief AI Officer.

00:57:11 — The First Human Received a Brain Implant from Elon Musk’s Neuralink

01:02:20 — Microsoft Copilot for Sales and Copilot for Service are now generally available

01:05:56 — Amazon announces Rufus, a new Gen AI-powered shopping experience

Summary

Are Microsoft 365 Copilot, Google Duet AI, ChatGPT Team / Enterprise, etc. worth the $30/user/month cost?

Businesses may hesitate to incorporate tools like Microsoft 365 Copilot, Google Duet AI, ChatGPT Team / Enterprise, etc. but Roetzer claims that they are well worth the cost, if integrated properly.

Roetzer, in a recent LinkedIn post, highlighted the efficiency potential of these tools:

An employee with a $75,000 annual salary costs about $36 per hour, based on a 40-hour work week. Introducing a GPT-4 level tool, assuming a conservative 10% efficiency boost, enables the same work to be done in 156 hours instead of 173, saving 17 hours or $612 monthly. Thus, a $30/month investment in such a tool is cost-effective, effectively yielding up to 190 hours of productivity within the same work week.

Roetzer asserts that the short answer to the question is that, effectively integrating these tools will “be the greatest value in business software history.”

But AI-enhanced efficiency requires companies to do more than just sign up for these services. Embracing these tools demands a strategic approach:

  • Creating Generative AI usage policies
  • Launching comprehensive AI education programs with proper onboarding of the technology
  • Appointing someone to oversee AI operations

Bard’s latest updates: Access Gemini Pro globally and generate images

Google Bard recently received major updates. The advanced Gemini Pro model was integrated into Bard for English users in December and has now expanded to support 40 languages, making it accessible in over 230 countries.

The second major development is the introduction of free image generation capabilities in Bard across most countries, powered by Google's Imagen 2 model, which debuted in December 2023.

Imagen 2, a cutting-edge AI model, is capable of producing detailed and lifelike images. It prioritizes safety by incorporating features to reduce the likelihood of generating harmful content and undergoes extensive testing to align with ethical AI practices.

To enhance privacy and safety, Imagen 2 excludes images of specific individuals and implements a unique digital watermark, SynthID, developed by Google DeepMind. This feature helps identify AI-generated images in Google Search or Chrome.

The model strives for a deeper understanding of the connection between text and imagery, enhancing its ability to grasp context and, in turn, elevating the quality of its output.

Additionally, this technology is integrated into Google's Search Generative Experience and Google Slides for users with a Google Duet AI subscription.

Apple, Alphabet, Microsoft and Meta’s AI-Driven Earnings Calls

Apple, Google’s parent company Alphabet, Microsoft, and Meta all had quarterly earnings calls in the past week—and all three had significant AI components.

First, Apple CEO Tim Cook confirmed that generative AI features are coming to Apple software “later this year.” He refused to reveal more details at the moment, but that likely signals that Apple is making a big move into AI.

Second, Alphabet/Google talked a big game about AI agents. CEO Sundar Pichai said that Google Assistant would “act more like an agent over time,” or an autonomous AI assistant that can do things for you.

Microsoft also showed solid earnings, with revenue from the quarter growing moderately faster year over year. 6 percentage points of growth in its Azure cloud business also came from demand for “AI services.” CEO Satya Nadella also gave a shout-out to Microsoft’s “small language models” that it’s rolling out, which are less costly AI models for customers.

Last but not least, on Meta’s earnings call, Mark Zuckerberg talked about the company’s vision to build full general intelligence.

Specifically, he said: “Previously, I thought that because many of the tools were social, commerce, or maybe media-oriented that it might be possible to deliver these products by solving only a subset of AI’s challenges. But now it’s clear that we’re going to need our models to be able to reason, plan, code, remember, and many other cognitive abilities in order to provide the best versions of the services that we envision.”

Links Referenced in the Show


Read the Transcription

Disclaimer: This transcription was written by AI, thanks to Descript, and has not been edited for content.

[00:00:00] Paul Roetzer: You cannot bet against Google. They have more data than anyone probably, and they have more AI history and capabilities than most other companies, if not all of them

[00:00:13] 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.

[00:00:34] Paul Roetzer: My name is Paul Roetzer. I'm the founder of Marketing AI Institute, and I'm your host.

[00:00:43] Paul Roetzer: Welcome to episode 82 of the the AI Show. I'm your host, Paul Roetzer, along with my co host, Mike Kaput. We are coming to you. It is Monday, February 8th at about 10 a. m. In case anything crazy happens today that we don't get to in this episode. So [00:01:00] you are potentially listening to this coming out on Tuesday, February 6th.

[00:01:04] Paul Roetzer: So, we are back. I am, I am home. Mike is home. We were both coming to you from Cleveland this week, which is great. I was out in Arizona last week, which was really cool for, a bunch of time with attorneys actually, which was a fascinating thing. I'm not going to get into it today, but. of conversations around copyright and IP and implications to, organizations and different opportunities and challenges that enterprises are going to face from a legal perspective with generative AI.

[00:01:34] Paul Roetzer: So a topic for a future episode, Mike, but it was, it was fascinating to. You know, hear their perspective and the things they're thinking about. So, yeah, I, my, the flight back was just like swirling with, with things I wanted to go research and topics to about. So, okay. So today's episode is brought to us by the Marketing AI Institute's AI for Writers Summit.

[00:01:59] Paul Roetzer: You've us [00:02:00] the is coming up in. one month from today. It's March 6th, 12 to 4 p. m. Eastern time. an on demand version, but that's going to be the live. a So, you can join us from anywhere in the world. last year we had over 4, 000 writers, editors, and content marketers join us for the inaugural event.

[00:02:21] Paul Roetzer: this year I haven't looked at registration lately, but I know it was up over probably 1, 500 or more already. So, it's a four hour event. We've got The sessions are going to be opening talk for me on the state of AI and writing. Mike is going to do generative AI writing tools and platforms you should We're going to have a conversation with, Meghan Keaney Anderson from Jasper, the presenting sponsor of the event. We're to have an interview I'm going to do with an IP attorney to be on the legal side, understand the legal landscape specifically related to generative AI, all the things you need to know and understand about copyright and trademark and, maybe getting some patent stuff.

[00:02:58] Paul Roetzer: And then we're going to have a panel [00:03:00] on AI writing in the enterprise, what it looks like to adopt these, tools and platforms within your marketing team and organization, then, we're going to end with a really cool feature. That's going to be AI in action, rapid fire tech demos for writers and creators that Mike and Cathy and I to kind of coordinate.

[00:03:16] Paul Roetzer: and that's going to be probably five to seven different tools showing you a bunch of different use cases. so I'm really excited about. So that is coming up. You can go to AIwriterssummit. com or You can just be at Marketing Institute's site and go under events. You'll find all the information. It is free to register.

[00:03:31] Paul Roetzer: There is a paid option, a private paid if you don't want to share your contact information with the sponsor. And there is an on demand option as well. So that is coming up again March 6th, one month from now. And then also, we talked a little bit last week about our new online education offerings.

[00:03:50] Paul Roetzer: So Piloting AI 2024 goes live on Thursday, February 8th. So Mike and I spent a good chunk of January, about a hundred hours or more. [00:04:00] I may have spent a hundred hours on my own. I don't know about you, Mike, but we did a full update. the Piloting AI 2023 course, the course series. And so it was all redone and recorded in January, 2024.

[00:04:14] Paul Roetzer: So it's content. we added an entirely new generative AI 101 course, which I'm pretty excited to get to people end up being like an hour and 20 minute course. I'm like, I know you haven't seen it yet, but. it was actually kind of fun to build that course. so last year we had more than 800 people go through that series.

[00:04:32] Paul Roetzer: Again, it'll be live 18 on demand courses with quizzes. There's a final exam with professional certification. the series, and and worksheets. Mike previews, what about 65 vendors? I think Mike go through years. So Mike teaches 10 courses in the middle that are all these.

[00:04:53] Paul Roetzer: for Advertising and Analytics and Communications and SEO and just all these different ways. it goes through dozens of use [00:05:00] cases and vendors, that can really help jumpstart your adoption of AI in your organization. So can go to pilotingai.com and learn more about that. Also available on Marketing Institute under the education component.

[00:05:13] Paul Roetzer: So, with that, turn over to Mike and let's get going on our main topics

Are paid AI tools worth the $30/user/month cost?

[00:05:18] Mike Kaput: All right, Paul, so first up, we've got kind of a big question that's on a lot of business leaders' minds. Um.

[00:05:25] Mike Kaput: Are tools like Microsoft CoPilot, Google Duet AI, ChatGPT Team/Enterprise, etc. Are these actually worth the money that we're spending per user, per month on them? Now, Paul, this is a question that you recently posted on LinkedIn as you broke down the benefit being provided by these paid AI tools and all these different subscriptions that companies are racking up.

[00:05:51] Mike Kaput: So you wrote, Consider this, an employee making 75, 000 a year costs approximately 36 an hour to [00:06:00] employ, assuming 2, 080 hours per week or about 173 hours per month based on a 40 hour work week. Now let's assume a very conservative 10 percent increase in efficiency for an employee who has access to a GPT 4 level solution and has been properly trained on how to use it.

[00:06:20] Mike Kaput: And you said you actually think that the average is closer to 20 percent or more for most people in the first full year of activation of these tools depending on their profession. So with a 10 percent efficiency gain, a full time employee could produce the same level of output in 156 hours, which saves 17 hours per month.

[00:06:43] Paul Roetzer: You didn't know you were going to do math when you started listening to the podcast this week, huh?

[00:06:46] Mike Kaput: Well, thankfully we're breaking it all out for people, so just follow along line by line. You don't even have to do the calculations. And the direct savings of those 17 hours is [00:07:00] 612. So when you compare that to these tools being about 30 bucks a month per user, that savings looks pretty good. And you kind of sum this up saying short answer is that these AI tools, when properly integrated into companies with effective education and training, will be the greatest value in business software history.

[00:07:26] Mike Kaput: Can you unpack your thinking here a little more Cause I seem, I see a lot that.

[00:07:31] Mike Kaput: Businesses sometimes aren't always convinced of the value of these tools. They get hung up on 20, 30, 40 bucks per user per month, because naturally so, they're budget conscious, but the math that you do here on the value of these tools when properly applied seems like a no brainer to me.

[00:07:49] Paul Roetzer: Yeah, I mean, the point I was kind of trying to make, and this is, it was funny. I think I wrote this like Thursday morning when I, when I was on my flight to Arizona. So I was up at like [00:08:00] 5am

[00:08:00] Paul Roetzer: and so I put a disclaimer on this, like, Hey, it's really early and I didn't sleep much. So if my math is off, like, forgive me, but it's directionally correct.

[00:08:09] Paul Roetzer: And I think it held up even once I got a cup of coffee I mean, the math still kind of stuck correctly. So this, it was on my mind because. when I was recording the piloting AI courses that I, mentioned earlier, there was actually one of the, courses where I get into this concept that, you know, one way to look at this is the productivity you're going to gain.

[00:08:30] Paul Roetzer: And so as we look at the cost of these tools. It doesn't take much to justify it. So, you know, if it's 30 per user, then actually, ironically, I went into Google Duet AI today, which is for, for workspace, which is their version, you know, of co pilot and ChatGPT team. And it is now 36. per user per month.

[00:08:51] Paul Roetzer: They've they've raised the rate. So it works out perfect. If you have a person making 75, 000 a year and you get them Duet AI for 36 a month, [00:09:00] if you save one hour or if you increase their productivity one hour, it's basically paying for itself. So on the surface, It seems like a no brainer. And then when I went to publish the post, I actually was like, Oh, hold on a second.

[00:09:15] Paul Roetzer: People are going to misinterpret this, that I'm saying everyone should go out and buy 150 licenses to Duet or Copilot or whatever. And that is actually not at all what I was saying. So then probably the most important part of the post. And I think the thing that resonated with people, because this, this post did take off.

[00:09:32] Paul Roetzer: Quite a bit. I think it had, it's almost 18, 000 impressions and 270 some engagements or something like that. So what I said at the end was, that you, you should not. These companies want you to buy 150 licenses. If you've got a 50 person marketing team, they want you to buy 50 licenses. If you've got a 10 person team, they want you to buy 10.

[00:09:52] Paul Roetzer: Like they want you to be buying as many licenses as possible and committing to a one year engagement. That is not what you should be doing. So [00:10:00] what I said at the end was do not go buy. licenses for your entire company or team, just because the math works out until you've done three things. So the first was you have to have generative AI policies AI policies that guide usage.

[00:10:14] Paul Roetzer: If you've listened to this show at all, you've heard us say this over and over again. So if I turn this on for our seven employees, And I don't tell them how they're allowed to use it. Like, are you allowed to use it for internal and external purposes? Do you have to disclose if you've used it for internal or external purposes?

[00:10:28] Paul Roetzer: Do you have to tell me that you used AI to write this email back to me? how are you going to protect the data? What are you allowed to put into it? What aren't you allowed to put into it? So there's just these fundamental things you have to answer for generative AI policies, and even if they're not perfect, you need to give your team.

[00:10:45] Paul Roetzer: Guidelines of how to use the technology before you turn this on for them. You also then, the second thing was, have an AI education program that teaches them how it works and provides onboarding. So if you just, like, if Mike and I get this turned [00:11:00] on, we generally kind of know how this stuff works, like we'll find our way.

[00:11:04] Paul Roetzer: But if you turn this on for people in your marketing team, or customer service, or sales, or your C suite, and you've provided no training for what it actually is, or even what generative AI is, like, People have heard that term, but reality is most people who have interacted with this technology have used a free version of ChatGPT a few times.

[00:11:23] Paul Roetzer: Like obviously we live in a bubble where everyone's using this stuff all the time, and if you listen to this show there's a decent chance you probably live in that bubble as well, but the average business professional and leader has no idea. I cannot tell you, like, I, the last like 10 talks I've done, every one of them I have asked, how many of you use ChatGPT?

[00:11:43] Paul Roetzer: Everybody raises their hands. How many of you use the paid version? Almost no one raises their hands. And so my argument is always, if you're not using the paid version, you have no idea what generative AI is capable of. Like if you're not using the more advanced versions of these tools, then you you have a [00:12:00] misperception of where we are with this technology.

[00:12:03] Paul Roetzer: So number one is have policies. Number two is teach your team what it is capable of, show them, give them, Hey, here's the five use cases you should be using. And here's samples of how to do it. Here's sample prompts. Here's template, like really show them so they get the value. And then the third is have someone who owns the rollout training and oversight of the program.

[00:12:24] Paul Roetzer: So if you turn on co pilot and go get your 10 licenses, or Duet AI, or ChatGP Team, or Jasper, or Writer, like whatever the platform is you get and use,

[00:12:35] Paul Roetzer: If no one is making sure that it's being utilized, then there's no value in it. And I think, let me see if I can pull it up real quick, oh yeah, Dan Slagan, so Dan, the CMO at tomorrow.

[00:12:46] Paul Roetzer: io commented on the LinkedIn post and he said, for the right people, 100 percent worth it. I'm a fan of having a usage minimum with clear value that people need to adhere to in order to qualify for the company to pay for it. I thought that's [00:13:00] a brilliant idea. A usage minimum is a great way to put it. So we're going to give you this tool.

[00:13:05] Paul Roetzer: It's going to cost us 30 per month for you to have it. You have to use it. X number of times, or you have to like show that you have increased productivity by X amount, like have a metric tied to the rollout of this. And if you can't do that, if you don't know how you're going to measure utilization or success or value, then don't buy the license for that person.

[00:13:27] Paul Roetzer: So. That again, it just kind of goes back to this idea of what we suggest. And we talk about over and over and over again, the whole idea of the piloting AI series, basically start with a pilot program that has a small number of users who are trained to do it, right. They are given guidance. There's oversight.

[00:13:43] Paul Roetzer: There is measurements in place of how they're going to be assessed on it. And then look at that and prove the value before you go buy it for everybody. So you have 50 people in your company. Put three people in the program, put five people, put a person from sales, a person for customer support, a person from marketing, like [00:14:00] do it cross discipline if you want to, but do this in a controlled way.

[00:14:04] Paul Roetzer: You're going to get pressure from the technology companies to buy everyone the licenses for 12 months. It is rarely going to be the right decision.

[00:14:13] Mike Kaput: So as part of this, you also mentioned it's really important for someone to own the rollout of AI technology in the organization. like what should that role look like? Who should be in it?

[00:14:27] Paul Roetzer: I think for bigger companies, it's going to be like an AI ops type of role. We've seen that emerging as, you know, someone who understands the internal workings of the organization, maybe they're project management, or I don't know, maybe it's HR, I don't know where they're going to come from, but someone who actually learns the product themselves, like deeply understands the product and the capabilities and can do the internal rollout and training, you know, work with the vendors, things like that.

[00:14:52] Paul Roetzer: so in, the near term, it might just be the people who are going to be using the tool. You might just have someone on your team. You're not going to give them a new title necessarily [00:15:00] to do this. But I think as organizations start to really scale this out and it starts moving beyond one or two pilot projects and you're managing a number of different vendors and onboarding processes, that might be where you start to see more and more AI, AI ops titles or something along those lines coming through.

[00:15:16] Paul Roetzer: Or, you know, it might be. generative AI manager. I don't know. Like, I'm not, again, this is, this is to me like the exciting thing. Like if you're kind of on the frontier here and you're like taking our piloting AI classes and you're the only one in your organization, like is really figuring out where to go with this.

[00:15:32] Paul Roetzer: I think you're going to see a lot of people kind of like. develop their own title and career path saying, Hey, I think our organization needs someone that does these 10 things. And that is not a role in our company yet today. And I'm seeing it being called an AI ops manager or director, whatever that is, and go get yourself like a new career path, potentially, you know, being the one that's kind of championing this within the organization.

[00:15:55] Paul Roetzer: And I think the key is if you can tie that to [00:16:00] value creation, like, I think this role can increase our productivity 15 percent in 2024. You're going to get a meeting in the C suite with that kind of claim. So if you can build a business case to show the value and like, hey, we've been testing it. Here's what we've seen over the first three months.

[00:16:16] Paul Roetzer: we saw an increase in efficiency of 20 percent on the writing team. We saw an increase in productivity on the creative team of 30%. The developers are, you know, if you can go with numbers, the C suite's going to listen and they're going to see value in what you're doing as a professional.

Google Bard just got some more big updates

[00:16:33] Mike Kaput: So in our next big topic today, Google BARD just got some significant updates. So back in December of 2023, BARD was, got an upgrade in the form of Gemini Pro, one of Google's most advanced new models. it was made available in BARD in English. back at the end of 2023. Well now, BARD powered by Gemini Pro is available in 40 [00:17:00] different languages and is now available in more than 230 countries.

[00:17:04] Mike Kaput: And as a reminder, we covered recently how a popular AI leaderboard actually ranked barred with Gemini Pro as the second most capable AI model out there right now. Behind only GPT-4 Turbo. It actually beat out some versions of GPT-4 and a bunch of other popular models. Now another really important piece of this is that Google also announced that you can generate images right within BARD in most countries around the world at no cost now using its new ImageN2 image generation model.

[00:17:43] Mike Kaput: Now, this model, Google says, has been trained on higher quality image description pairings. and generates more detailed images that are better aligned with the semantics of people's language prompts. It is more accurate than our previous system [00:18:00] at processing details, and it's more capable at capturing nuance.

[00:18:04] Mike Kaput: which means it can deliver photorealistic images across a range of different styles and use cases. What's also interesting about this is that as Imogen 2 is rolling out across Google's suite of products, they have significantly prioritized the model's safety. They've added quite a few guardrails, technical guardrails, to prevent you from generating, offensive content.

[00:18:30] Mike Kaput: they also conducted extensive red teaming to make sure. They mitigate potential harmful and problematic content. And also, all images generated with ImageInto in consumer products are going to be marked by SynthID, which is something we covered on a previous podcast episode. It is a tool developed by Google DeepMind that adds a digital watermark directly into the pixels.

[00:18:56] Mike Kaput: of the image.

[00:18:57] Mike Kaput: So you can't see this with the human eye, but [00:19:00] basically you can quickly identify if an image has been generated using AI technology. So, Paul, as you're reviewing these updates, how big a deal are these advancements and these new features in BART?

[00:19:17] Paul Roetzer: Yeah, Imogen2 is interesting because, I don't know if people remember back, but, you know, when MidJourney and DALI came out in, what was that, like February March 2022. So before ChatGPT came out, we actually started seeing image generation tools. We weren't really hearing Generative AI is like the category describing all these capabilities until ChatGPT is like really when that that phrase started becoming very popular.

[00:19:44] Paul Roetzer: But the reality is that in spring 2022, these image generation tools started emerging. Google did not publicly release Imogen.

[00:19:52] Paul Roetzer: So they had a tool like MidJourney, like Dolly, but they did not release it, maybe because the product wasn't ready. They they [00:20:00] explained it more as like a safety thing that they didn't put it out.And then we saw the, these capabilities start becoming available through,

[00:20:08] Paul Roetzer: is it Vertex AI? I think is kind of like the thing you can build on with Google. That the average business person wouldn't be using Vertex, but a developer could, could build things on top of that. And so now we're seeing it directly built right into BARD.

[00:20:26] Paul Roetzer: So I think that what we're starting to see is the more comprehensive approach that Google is taking to compete head to head. with Microsoft and OpenAI and others here. And so the ability to have one BART is free. It now has this very powerful large language model, as you mentioned, Gemini Pro, which is currently the second most powerful built right in.

[00:20:51] Paul Roetzer: And now you have a powerful image generation tool. We know they're going to mess with video. They own YouTube, like video is [00:21:00] coming to BARD at some point. It may not be in the form of Gemini Pro. The thing I start to really wonder here is what, where's the monetization strategy? Like, I know I can go pay 36 a month for duet AI, but they're not replacing search and ads with 36 a month for corporate users.

[00:21:18] Paul Roetzer: So like the thing I start to wonder is what is like. What is Gemini Advanced/Ultra? Like, is that a paid version of BARD that I might actually want to use instead of ChatGPT or Perplexity? Because I think what's going to happen is in the coming months, we're going to start to really see the full picture of what Google's vision is for Bard. there's some increasing murmurs that Gemini Ultra may be around the corner, that, that it, we could be seeing the release of the more advanced version of this model, at some time, and then I assume that would probably be a paid version of Bard is what I'm guessing their strategy is. But I went back [00:22:00] to, there was an article in June in Wired Magazine with Demis Hassabis, who People remember he is the head of, Google DeepMind.

[00:22:10] Paul Roetzer: and he was talking about, Gemini. It was like the first time that they really, acknowledged that Gemini was something they were working on. And so I went back to that because I keep trying to figure out, like, what is the grand plan here? Like, where does Google go with this technology? Because their, their search business is being threatened.

[00:22:29] Paul Roetzer: Like, they know that their ad model. probably is being threatened in some capacity, as you and I have talked about. Perplexity is just a better user experience. Like if I'm looking for information, I don't like 10 blue links anymore. Like I don't want to click through stuff. Google knows that, like, but they can't just replace their ad business overnight.

[00:22:46] Paul Roetzer: It is, it is where their money comes from. It's their profit engine. So. They have to know that. And so what is the play here? And if you go back to that article in Wired Magazine, Demis talks about like the [00:23:00] infusion of their, their DeepMind capabilities with the large language model, because one of the things Google has is DeepMind that no one else has.

[00:23:10] Paul Roetzer: And at one point he said, there was a quote he said, where like he, Demis Asavas attributed like 80 to 90%. percent Of the advancements that have been made in AI. Oh, here it is. If you look at where we are in AI would argue that 80 or 90 percent of the innovations come from one of their Google teams internally, either the Google brain team, which invented the transformer, which led to ChatGPT or the Google DeepMind team.

[00:23:37] Paul Roetzer: If you want to see what I'm talking about. I've mentioned this probably a bunch of times on the show, but go watch AlphaGo. So back in 2016, I think is when the AlphaGo movie came out. It's when their AlphaGo machine, defeated world champion at the game of Go, Lee Sedol. And it's incredible. Like, the strategic ability, the reasoning ability.

[00:23:59] Paul Roetzer: And they're [00:24:00] envisioning building that technology into this model. So they are very much moving in the direction of AI agents, the ability to reason tasks, the ability to have chain of thought, the ability to like, solve problems, mathematics, coding, like, And so I think it's really important to, again, to surface this as a main topic, because you can't look at BARD today and be like, eh, it's just not as good.

[00:24:27] Paul Roetzer: You can't look at Google and think, yeah, perplexity is just better. They're going to win. I've said it over and over and over again. You cannot bet against Google. They have more data than anyone probably, and they have more AI history and capabilities than most other companies, if not all of them,

[00:24:49] Paul Roetzer: they weren't ready for generative AI. Like they, it was the innovators dilemma. They had the technology, but to do it would have had to disrupt their own business model, which is very profitable. So they were [00:25:00] stuck and now they're trying to figure out their way around the innovators dilemma, I guess here, where the people have now come for their core business.

[00:25:08] Paul Roetzer: But. I think we just have to keep our eye on this and for me it's all about what is Gemini Ultra or Advanced and how does it fit into this picture. And I do think that in the coming months we're going to have much more clarity around their, their bigger vision for what happens next.

[00:25:25] Mike Kaput: So it sounds like business leaders definitely need to be paying attention to the moves Google is making, especially as your, As a marketer, if any of your initiatives or work touches any type of Google property, right? Like whether you're using their tools in your own work or you are, have an active presence on YouTube, et cetera, Or in search, obviously. What do these updates mean for the average business leader trying to figure this stuff out?

[00:25:56] Paul Roetzer: I don't, I don't know that they mean it too much yet. What I'm, I'll say what [00:26:00] I'm doing, and I don't know if you're doing something similar, Mike, but I, you know, so we have ChatGPT/plus, like I've, you know, paid version personally and then for the, for the company, it is still my, my, my main tool. So the two tools I use.

[00:26:13] Paul Roetzer: In lieu of kind of what I might've previously done only in Google, I use ChatGPT plus a lot. and I also use Perplexity increasingly for, for search. So those are like two daily, daily workflow tools for me they've become. What I think I'm going to likely now do with BARD is I will, my, my normal use cases for ChatGPT.

[00:26:34] Paul Roetzer: I will also probably run them in BARD side by side and compare the outputs because BARD's free, I can do that. Like, it's, you know, I don't have to spend another 20 bucks a month to get, you know, Gemini Pro in BARD. So, that's kind of what I always suggest to people is Anthropic Claude is a really good tool, you know, Bard appears to be a player.

[00:26:58] Paul Roetzer: You have ChatGPT, maybe you [00:27:00] have, you know, Jasper, a writer or something like that. It's going to be a constant game of leapfrog. And I think when you have your core uses that are part of your daily workflow, or even like interesting ones you want to experiment with, run them, run the same prompts. Through each of the tools and, you know, maybe it's every, couple months or when a new update comes out in BARD, go back and say, okay, summarization,

[00:27:24] Paul Roetzer: ideation, like whatever your normal uses are for these tools, go in and run them And maybe like create it and say, here's my three or. four. five prompts. And every time there's a major update to a system, I'm going to go run these same three or five prompts and see the improvement for myself. I think that kind of consistent use of the same prompts that enables you to go and see the improvement on your own is really helpful way to think about it.

Apple, Alphabet, Microsoft and Meta’s AI-Driven Earnings Calls

[00:27:47] Mike Kaput: So in our third big topic this week, we So, earnings calls for quarterly earnings come from Apple, Alphabet, Google's parent company, Microsoft, and Meta, and [00:28:00] they all had a significant AI component. So, first up, Apple CEO Tim Cook. Confirmed that generative AI features are coming to Apple software quote later this year while he refused to Reveal more details at the moment Many people believe this is a likely signal that Apple is making a very big move into AI Second, Alphabet/Google talked a ton about AI agents.

[00:28:29] Mike Kaput: CEO Sundar Pichai said that Google Assistant would, quote, act more like an agent over time. And by agent, we mean an autonomous AI assistant that can do things for you. And the company actually also refers to Duet AI,

[00:28:45] Mike Kaput: which works in Docs, Sheets, et cetera, kind of as an AI assistant. They already call that an agent in some of their public facing comments.

[00:28:53] Mike Kaput: Not to mention the information previously reported that Google is working on an agent called Pixie for its [00:29:00] Pixel phones. Microsoft also showed very solid earnings with revenue from the quarter growing moderately faster year over year. And interestingly six percentage points of growth in its Azure Cloud business also came from demand for AI services.

[00:29:17] Mike Kaput: CEO Satya Nadella gave a shout out to Microsoft's quote small language models, which it's rolling out, and these are less costly AI models for their customers. And users of the company's GitHub co pilot AI assistant for coding are also up to 1. 3 million users from 1 million in October of 2023. Last but certainly not least, on Meta's earning call, Mark Zuckerberg talked about the company's vision to build full general intelligence.

[00:29:50] Mike Kaput: Specifically, he said, quote, Previously, I thought that because many of the tools were social, commerce, or maybe media oriented, that it might be possible to [00:30:00] deliver these products by solving only a subset of AI's challenges. But now it's clear that we're going to need our models to be able to reason, plan, code, remember, and many other cognitive capabilities in order to provide the best versions of the services.

[00:30:16] Mike Kaput: that we envision. So Paul, I want to start with Apple. Can you unpack the significance of Tim Cook's comments during Apple's earnings call? While he was short on details, the company does not talk about AI explicitly too, too often in these types of settings.

[00:30:36] Paul Roetzer: I think we maybe mentioned this on the podcast last we get they, for whatever reason, just avoid the term altogether, like even in their own messaging, they'll talk about it, like machine learning in some ways within like their chips and stuff like that. But they just don't.

[00:30:52] Paul Roetzer: but we don't talk about the technology as AI, even though the iPhone has been infused with it for like 10 years.

[00:30:57] Paul Roetzer: . Hmm. just a decision they make. So I [00:31:00] always wondered like when the pressure was going to hit them to where they have to acknowledge for Wall Street purposes that yes, we actually are building AI and we are a leader in it. And this was the earnings call where it finally came to a head. So yeah, him mentioning AI and specifically generative AI think is a very big deal for where Apple goes next.

[00:31:18] Paul Roetzer: He did, I don't have the exact quote here, but he did say something in the call. Like we like to build things, then talk about them because he was getting pressure, like what exactly are you going to do with generative AI? I think there is the obvious thing of Surrey has to actually become intelligent.

[00:31:33] Paul Roetzer: Like be functional. I think I've said before, like it has to do something more than weather and reminders. Like. 95 percent of how I use Siri is to set reminders on my phone, like, set a reminder for tonight at, you know, 6pm to do this thing. I would love to have it actually function for something way more than that.

[00:31:49] Paul Roetzer: And I do think we are going to get that. the date to watch here, I don't think they've announced the actual day yet. But in June, they have their Worldwide Developers Conference. And he alluded [00:32:00] to later this year, they would be, you know, announcing things with generative AI. My expectation is June is the date to watch.

[00:32:06] Paul Roetzer: There will be probably a lot of announcements. And then I would assume what they would do is probably make those announcements available to developers to build on top of, and then I would expect maybe in their, like September is usually when they. Do updated iPhones and things like that. So I would think that maybe by this summer we'll get a better vision of what they're going to do with ai.

[00:32:29] Paul Roetzer: And then probably by the fall of this year, we might actually start to see a next gen operating system looks like, how it's going to be infused into your phone, how Sury becomes smarter. And then I wonder at what point we start to see the vision for what it, what is the iPhone evolve into, you know, especially with the Vision Pro coming out.

[00:32:47] Paul Roetzer: We'll talk about that in a couple minutes in the rapid fire, but I really start. to wonder like, what is the iPhone, how are AI agents going to be infused into this? As you alluded to earlier, what does Surrey look [00:33:00] like? And I think we're going to get a lot of answers from Apple this year. And I'm, it's the one company I just keep waiting for.

[00:33:06] Paul Roetzer: I mean, Google, I'm extremely interested in what they do and how they evolve. But I think Apple is often just overlooked in this space because they don't tout. AI all the time. And so they're kind of forgotten about, but they're a major player.

[00:33:23] Mike Kaput: So speaking of Google, they definitely made some noise about AI agents kind of combined with what you were talking about related to BARD and Gemini. How do you kind of see their, rumblings about AI agents kind of fitting into the overall strategy or vision of that business?

[00:33:43] Paul Roetzer: Yeah, so AI agents, if again, if you're new to the show and you haven't really heard us talk about AI agents before, the basic premise here is all of these major AI labs are all trying to build models that can take actions on your behalf. So right now, the language models we have today [00:34:00] that are, you know, that we see and experience through ChatGPT and BARD, and CLAUDE.

[00:34:05] Paul Roetzer: They can't take actions. They can just build lists of things. They can't go on book flights or send your emails for you, per se. They can write your emails. and so they're all moving toward a phase where these things can take actions on our behalf. And I think the significance of a Google and or an Apple is you already trust them with all of your data.

[00:34:25] Paul Roetzer: So we talked, I think I mentioned last week, this multi on is like this AI agent company that I started. tracking last week. And to use it, you have to give it access to your browser. Like you install it as a Chrome extension and then it sees everything you do. And so the question becomes like, well, do I trust this company?

[00:34:46] Paul Roetzer: Like who is it? Like who started this company? Who's funding it? What government is maybe behind it or what sovereign wealth fund is fun? Like, I have no idea. And it might not be any of those things, but I don't know. Google and Apple, if they [00:35:00] come to me and say, Hey, we now have AI agents that can take actions on your behalf.

[00:35:03] Paul Roetzer: And it's Google. It's like, well, I'm, you already have everything you have. I use chroma for everything. but you already know everything I'm doing. So I am far more likely to experiment with a tool or to adopt a tool that's coming from a major technology company that I already trust with my life. Like they have all of it stored.

[00:35:24] Paul Roetzer: For the startup ecosystem, it's not probably great that these companies are moving in that same direction because I think a lot of these companies are just going to get rolled out. Like they're, they're just going to get obsoleted or, you know, maybe they get lucky and get acquired for some talent acquisition, but it's hard to look at things being built in the startup community that Apple and Google and others aren't going to just.

[00:35:48] Paul Roetzer: Come out with themselves and obsolete those things. So, yeah, I don't know. I'm in, I pulled a Reuters just because for context, like one of the ways that we used to look at this stuff. So before everyone was talking about [00:36:00] AI, like before ChatGPT, when we were just. Building the AI Institute and trying to create awareness about this.

[00:36:05] Paul Roetzer: One of the ways we used to gauge the companies that were serious about it was, how many AI machine learning engineers they had on staff. So I would go into sales navigator, LinkedIn sales navigator, and I would look at and keep track of how many AI engineers were working at these different companies.

[00:36:20] Paul Roetzer: And the other one was. Mentions of AI on earnings calls and prior to ChatGPT, it wasn't very common. So according to Reuters, and we'll put the link in the show notes, the term AI or artificial intelligence has been uttered on 38% percent of conference calls held by s p 500 companies in January. So that, that was this year for Q4 earnings.

[00:36:44] Paul Roetzer: Reuters analysts, transcript shows. That's up from 34% percent at the same point during third quarter. But the one I thought was awesome is I looked at, what was it, last year in January, so January of 2023 was 15%. Now, [00:37:00] there's a common mistake people make when calculating percentages. I see this all the time, even in like mainstream media articles, but to go from 15%.

[00:37:09] Paul Roetzer: To 38% is not 23%. So the assumption is third, 38 minus 15, that that's not how it works. The percentage points. So to increase from 15% to 38% is 153% increase. It's remove the percentage signs go from 15 to 38 is 153%. So mentions of AI on earnings calls. In a 12 month period is up 153% percent and the leaders for, the January calls, alphabet and Microsoft at 53 and 52 times each

[00:37:41] Paul Roetzer: So again, like I think, you know, when we're talking about this topic, part of it might be of interest to you as an investor, like where are the stocks to be, thinking about the companies to be tracking. But just to understand. One of the things I always say when I, you know, I do my keynotes is

[00:37:59] Paul Roetzer: [00:38:00] There is a necessity for a next gen professional and a next gen leader because the technology companies that power business have bet everything on this.

[00:38:09] Paul Roetzer: And so when you look at earnings calls and you see the significance of the mentions by their CEOs and CFOs, it is all of their thinking about. And so. What has to happen is this technology that they're building, AI agents and intelligent virtual assistants and all these things, it will get diffused into society and into businesses.

[00:38:27] Paul Roetzer: And so, you, as a business leader or a practitioner, marketer, whatever you do in your career, the opportunity for you is to be the one to figure this out. It's not going anywhere. Like, it's only going to increase the amount of times they're talking about it, and the amount of products they're rolling out, and how quickly those products come.

[00:38:47] Paul Roetzer: Hopefully it's interesting context for people. I personally, I, no investing advice here. I'm not telling people what stocks to buy. I personally have avoided meta stock forever.

[00:38:59] Paul Roetzer: Like, I just was [00:39:00] not interested in Facebook as a company because I knew that. They had one of the leading AI research labs in the world, and they didn't seem to care.

[00:39:09] Paul Roetzer: Like it was just basically being built into their newsfeed and into other products like messenger and wherever else they're using it.

[00:39:16] Paul Roetzer: but it wasn't the focus of Zuckerberg The Metaverse wasn't, I didn't care about the Metaverse. I thought 10 years off if they were it was once they pivoted and and started truly building their messaging around their AI capabilities.

[00:39:31] Paul Roetzer: That was when I actually said, Okay, now this is an interesting company. And so, I have personally started taking a much greater interest in Meta as a stock. And the day after their earnings, it jumped 21%, percent like Friday morning, I was at this talk and I'm looking at this, like, Oh my God, 21%! Now part of it was because they put out their first dividend, I think, ever.

[00:39:53] Paul Roetzer: But, I think in large part, it's the AI thing. And so I'll read one quick thing and then we can get into rapid fire. [00:40:00] on the call, Zuckerberg address the opensource issue. And so we've talked on this podcast many times about closed models, like Google and OpenAI where they're not sharing what they're doing.

[00:40:11] Paul Roetzer: You can't go build on it. Like you could an open source model. And then you have the companies like Mixtral and Meta and others that are creating these models and just. sharing them so people can build whatever they want on them. And so Zuckerberg addressed this in the earnings call. And so I'll just read this because I think, and again, like we won't unpack this now.

[00:40:29] Paul Roetzer: Maybe we'll come back around to this, but I think it's a really important thing for people to hear cause I've wondered this myself sometimes. So he said, I know some people have questions about how we benefit from open sourcing the results of our research and large amounts of compute. So I thought it might be useful to lay out the strategic benefits here.

[00:40:45] Paul Roetzer: The short version is that open sourcing improves our models and because there's still significant work to turn our models into products and because there will be other open source models available anyway, we find there are mostly advantages to being the open source leader [00:41:00] and it doesn't remove differentiation from our products much anyway.

[00:41:04] Paul Roetzer: then he gets into specific, strategic benefits. First open source software is typically safer and more secure. That is. One opinion, like there are people who do not agree with what he's saying here, that is typically safer and more secure as well as more compute efficient to operate due to ongoing feedback, scrutiny, and development, um, big deal, efficiency improvements and lowering compute costs also benefit everyone, including Second, open source software often becomes an industry standard, like a Linux is kind of what they're thinking about here.

[00:41:35] Paul Roetzer: And when companies standardize on building with RStack, then it becomes easier to integrate new innovations into our products. And then third, open source is hugely popular with developers and researchers. We know that people want to work on open systems that will be widely adopted. So it helps them with recruiting, you know, talent and things like that.

[00:41:52] Paul Roetzer: So, you know, I think again, topic for another time, but there's always this debate of, well, isn't open source just going to give everyone, including [00:42:00] bad actors, the power to do all these things? Yes, it will actually. and I tell you like, you know, have her just spending time with a bunch of attorneys for a couple of days.

[00:42:09] Paul Roetzer: I got to ask this question like a dozen times. about like the dangers of open sourcing. So, I don't know. Fascinating thread to pull on in a future topic, but that is why Meta's doing it. It was the first time I think he's given like a super clear answer of why they see the value.

Apple Vision Pro launches

[00:42:25] Mike Kaput: So diving into some rapid fire topics today. First up, we just saw the, launch of Apple's much hyped Vision Pro VR headset, which started shipping in the U. S. last week,

[00:42:40] Paul Roetzer: and Spatial Computing.We probably shouldn't refer to it as VR. It's their category. They don't want to call it VR, or AR.

[00:42:48] Mike Kaput: Yeah, so this is an important point that we'll get to because it combines a little VR, a little augmented reality, and they're going to wrap this term spatial computing around their [00:43:00] headset and The headset itself starts at 3, 500 and it's actually the first new Apple device to hit the market since the Apple Watch, which is why everyone has been desperate to get their hands on it.

[00:43:15] Mike Kaput: We've already started to see some early reviews. you know, in the show notes we'll link to literally thousands of words of reviews that pick apart kind of every detail of the device so you can Take a look at every individual element of this, but the overall idea here is exactly what Paul mentioned, that Apple wants to usher in this new experience with this device that they're calling Spatial Computing, and this essentially means seamlessly working and playing in both virtual reality and augmented reality, so basically putting apps and media and games in some cases overlaid onto the real world. Now, from some of the reviews that we took a look at, it seems like [00:44:00] the Vision Pro is being praised for its ultra high quality display and its incredible pass through technology. So this is what allows you to kind of see the world around you. along with those VR and AR elements overlaid on the real world.

[00:44:15] Mike Kaput: It's also being lauded for the fact that you don't need a controller to manipulate the environment around you or the apps or the windows that you have open using the headset. You simply use your hands. However, the Verge, which is largely complimentary of the device, notes that, quote, There is so much technology in this thing that feels like magic when it works and frustrates you completely when it doesn't.

[00:44:39] Mike Kaput: So. Basically, where they got was this is truly a stunning consumer facing headset, but we're still in such early days with figuring out, okay, exactly how are we applying this device? How often are we using it? Where are we using it? And what does it mean for kind of our habits when it comes to actual computing and getting things [00:45:00] done online?

[00:45:01] Mike Kaput: So, Paul, what was kind of your initial read on the Vision Pro? Anything excite you about this?

[00:45:07] Paul Roetzer: Yeah, I kind of like, I'll break it into four quick thoughts. One, the technology is remarkable. Like I watched the video of how it's made, which was pretty incredible. Like the whole manufacturing process that goes into this. I know they've been working on it for years. I think there was a. Was it Vanity Fair?

[00:45:22] Paul Roetzer: Maybe Tim Cook did an interview where he, I think he talked about like Steve Jobs was actually involved in some of the early conversations around this thing. So they have had this vision to build this device for a very long time. so I am anxious to try it. I don't know that I'll buy one. a lot, sometimes I'll read reviews and it's like, Oh, I got to try this thing.

[00:45:42] Paul Roetzer: But the thing that comes back to me is like. I feel like I would use it for like the first three days and just be like in awe of it. And then it would probably sit on my desk Cause it doesn't seem clear yet. Like, is this supposed to be a consumer device? Is it, is it mostly going to be for enterprises and business users?

[00:45:59] Paul Roetzer: [00:46:00] and I don't know that. Like we know yet, but what Apple has is a massive developer community, and I could see some killer apps being built where it starts to become more clear. The true value of this thing. The idea of being able to integrate, like, interact with photos and videos and just see them like in, you know, widescreen display and watch shows like That sounds cool to me, but I also find this product quite depressing.

[00:46:25] Paul Roetzer: I don't know, I saw so many things on Twitter of people like out in the real world, just like walking around with these things on, sitting on the subway with them on, you know, sitting at the restaurant with them on, and I just There's something dystopian about that to me.

[00:46:39] Paul Roetzer: Like, I don't know, I think it's different when it's glasses and I feel like you and I are interacting, but if a guy walked into the office, Mike, and you were just walking around with your vision pro on, I would think like that, that's just weird. Like, I don't know if he's like, I'm talking to him, but like, is he looking at five other screens right now while we're talking to each other?

[00:46:58] Paul Roetzer: Like, you know, phones are [00:47:00] always with us and they're distracting. And like, it's kind of rude when we look at our phones when we're with other people, but I know when you look down and look back up. We're we're now together like we're conversing and making eye contact if you're keeping your vision pro on Like what like I heard one that was I thought was super depressing It was like the New York Times lady that covered it and she was talking about how they're going to enable you to like watch movies together So you're sitting on a couch together Both in your vision pro watching the same movie instead of like just being in the room together So, I don't know I find the concept of walking around with these big goggles on to be somewhat depressing I

[00:47:44] Paul Roetzer: think if they were if they are used within like an environment like an office and I'm using it as productivity because it sounds like It's an amazing potential productivity tool.

[00:47:51] Paul Roetzer: That's cool. If we start seeing zombies with like fancy expensive goggles walking around society. I I think [00:48:00] we're going to have taken a wrong turn from a technology perspective. which leads me to my last point, which is this is very much V one. Some of the reviews I've read is like, there's parts of this, which is just the future.

[00:48:13] Paul Roetzer: Like we are living in the future. And then there's glitchy things that don't usually come with Apple products that just don't work. And so I've heard the explanation of some of this looks like they worked on it for 10 years. And some of it looks like they worked on it for 10 days to like get something out the door.

[00:48:29] Paul Roetzer: So the question for me is like, what comes next? Like is this technology just V1 and three to five years from now, they figure out how to integrate most of these capabilities into a pair of glasses, like the Meta Ray Ban glasses. Then I think we are talking about a shift in Everything like if, they can put it into a form factor that doesn't feel dystopian, and I think what they'll do is they're going to look at all the usage, what are the tools, the apps, what are the capabilities people are using and how do we [00:49:00] constantly kind of evolve this? But

[00:49:01] Paul Roetzer: I, don't look forward to like five years from now where, you know, you're on a trip and half the people you're going past are in goggles. I hope that that is not our future and I don't think it will be.

[00:49:13] Mike Kaput: Yeah, Yeah, hopefully we don't go full ready player one over here.

[00:49:18] Paul Roetzer: Yeah, it would be really bad move.

[00:49:19] Paul Roetzer: It's a great example, and I, that book was amazing, but it was depressing as hell. Yeah, Yeah,

Who Owns Your Voice?

[00:49:25] Mike Kaput: yeah. Uh Well, in another interesting Sci Fi becoming science fact, we're considering the following question and trying to kind of solve for it. which is

[00:49:40] Mike Kaput: Who owns your voice and likeness in the age of generative AI? And Paul, this is something you brought up and posted about as we were using Descript, which we use to edit our podcast, to train synthetic versions of our voices to make it easier for Kathy on our team to edit the [00:50:00] podcast and also some of the webinars and other content we create.

[00:50:03] Mike Kaput: This is a feature you can access right in Descript, which is affordable video and audio editing. technology. you asked me, I wonder who owns the voice synthesis and then you wrote on LinkedIn, from a legal perspective, anything that an employee creates is the property of the employer, at least given my understanding of the law.

[00:50:23] Mike Kaput: So do employers therefore have a right in perpetuity to use someone's voice and likeness that they willingly created for the company. You know, we didn't sign any waivers or legal documents related to the AI voice training, which made you think that corporations need to update their employment agreements to address whether or not AI voice and video models are the property of the company or the employee.

[00:50:46] Mike Kaput: So based on the comments on your LinkedIn post, it sounds like there's a lot of unanswered questions around this topic. Like, what are some of the considerations that you've kind of been thinking about since you wrote about this?

[00:50:59] Paul Roetzer: [00:51:00] Yeah, so real quick, like the practical use case here as an example, if people aren't familiar with this concept is, you know, we, we train a voice by reading these like 30 seconds of text and now in Descript is our voice is stored. And so the way you edit in Descript, like if, so this podcast, the transcript and the video will go into Descript and then our team can make edits.

[00:51:20] Paul Roetzer: In Descript, you can edit just by. deleting and adding words. So like you can cut out a part of the video. So let's say I just like, whatever, I had a sneezing fit or a coughing fit right now. They could just go in to that segment where I was coughing and delete it. And it cuts the video. It edits the video for them.

[00:51:38] Paul Roetzer: So the use case in the voice synthesis is let's say I'm talking about our AI writers summit, and I say, June 6th instead of March 6th. And I just kind of screw up and lose myself for a minute. The team can go in and just highlight where I said the wrong date and change the date and it'll sound like I said it seamlessly.

[00:51:57] Paul Roetzer: So it's using my voice to [00:52:00] edit. text, which then changes the video and the audio. So that's why you would do it. Like if you're wondering why would we create models of our voice? but yeah, so I, so I did put it on LinkedIn and then ironically, again, I was just with a hundred attorneys, like, and this came up, I was using this as an example.

[00:52:18] Paul Roetzer: And so I think like right now, the general guidance I was hearing from people is it's probably covered like under existing agreements in terms of, you know, the broad blanket of how these things are written. But. A lot of attorneys I've talked to are like, yeah, we probably should revisit employment agreements because this, this is just one example in the generative AI world.

[00:52:38] Paul Roetzer: Like if you do the Haygen thing and you create, like, let's say you're a customer service rep or a sales rep or an executive and you create a deep fake of yourself on Haygen that you use to send personalized messages to new customers or something like that. who owns that? Like, when the executive leaves, like, does the deepfake go with the [00:53:00] executive?

[00:53:00] Paul Roetzer: Like, I don't know. So I, my big takeaway was corporate and employment law has a lot of work to do. It's good news for attorneys. Like, I think there's all kinds of gaps probably in how these things are currently structured. And then I think I mentioned to our team internally, or someone asked about, in a comment, like, what about like speaking engagements, you want to make sure that we're building in, you and I do a lot of public speaking.

[00:53:24] Paul Roetzer: that nothing can be used that we create for, you know, some sort of synthetic purposes. I don't know. So just an example of how wild this stuff is going to get and how, how few answers we have, because honestly, like most people just don't even know to ask the questions. Like we just sort of take what's coming and don't really think about it.

[00:53:42] Paul Roetzer: so yeah, I don't know. It was an interesting one. It was just sort of a random thought we had after the podcast last week.

The hottest new job in corporate America? Chief AI Officer.

[00:53:51] Mike Kaput: according to the New York Times, there is a new hot job in corporate America, and that job is Chief AI Officer. [00:54:00] According to a report recently from the New York Times, many people have long feared that AI would kill jobs, but a boom in the technology has instead spurred law firms, hospitals, insurance companies, government agencies, and universities to create what has become the hottest new role.

[00:54:16] Mike Kaput: in corporate America and beyond. The senior executive in charge of AI. Now the senior AI executive role they mention includes titles like chief AI officer and VP of AI and they are essentially tasked with coordinating AI efforts and rollouts across every function and division. The Times actually reports that companies in very diverse industries including Accenture, Ashley Furniture, and Equifax have hired for these roles in the past year.

[00:54:49] Mike Kaput: So there's still not a ton of them. Not every company has this type of role, but it does seem like they're taking on an increasing popularity, uh, in [00:55:00] the corporate world. So Paul, I'm curious kind of what you think of the value is, if any, is in having a chief AI officer or an executive in charge specifically of AI at a company.

[00:55:14] Paul Roetzer: Yeah, I think it'll probably start becoming more common. It made me remember back to October when the Biden administration issued their, you know, executive guidance on AI and they require this position. So section 10, advancing federal government use of AI requires the creation of a chief AI officer for each agency within the government.

[00:55:34] Paul Roetzer: I think, you know, actually I'm looking at it right now. So chief AI consistent with the guidance described in subsection 10. 1 shall be represented on it. Inter Agency Council, requirements to designate at each agency within 60 days of the issuance of the guidance, the Chief Artificial Intelligence officer who shall hold primary responsibility in their agency, coordination of officials.

[00:55:56] Paul Roetzer: And it gets into like responsibilities. So, I mean, it's going to be a thing in the [00:56:00] government! So I, yeah, I think it's going to happen for sure. I don't know if it's necessary, like the way I've historically looked at AI is, it's going to be the job of everybody. Like your CFO is going to need to basically be an AI officer for finance.

[00:56:15] Paul Roetzer: You're, you know, heads of strategy are going to need to be AI people. Like. But my only concern with this is that we think that by having a chief AI officer, we can centralize all the learning and knowledge about AI into like that person's role and the people that report to that person. When our feeling is no, that we need to, everybody, we have to level all the company up.

[00:56:37] Paul Roetzer: I don't care what their role is. So that's the one thing I wondered down the road is like, Is it an unnecessary thing? But I could see certainly in the near term why it is. Because like, again, I I think of AI as like an underlying operating system to all of business. Like there, there isn't going to be a function of business that isn't affected by AI. I guess it makes sense that you might need a leader [00:57:00] who's figuring out how to, centralize or decentralize this technology and the guidance and the policies and the principles. But I don't know, if the CEO should be the chief eye officer in my opinion.

The First Human Received a Brain Implant from Elon Musk’s Neuralink

[00:57:11] Mike Kaput: So another story this week shows us that the future is coming at us pretty fast because Elon Musk has posted on X Quote, The first human received an implant from Neuralink yesterday and is recovering well. Initial results show promising neuron spike detection. As a reminder, Neuralink is Musk's company that is developing implantable brain computer interfaces.

[00:57:40] Mike Kaput: In another post on the same day, Musk said that the company's first product is called Telepathy, and it quote, enables control of your phone or computer, and through them almost any device, just by thinking. Initial users, he says, are going to be those, they're kind of looking at people who have lost the use of their limbs as like a [00:58:00] therapeutic technology or use case for it.

[00:58:03] Mike Kaput: So Paul, like, let's zoom in and talk about your thoughts on this announcement specifically and then maybe just quickly zoom out like what does this kind of mean for this whole story of AI? Like do we just end up at some point merging with technology like the effect of acceleration as people assume we will?

[00:58:20] Paul Roetzer: Yeah, so first, apologies if you weren't aware that Elon Musk owns a company that's implanting computers into brains. Like, if that scares you a little bit, like, yes, welcome to the future. so I think Neuralink's an interesting company. I've been tracking it since its origins. it is part, like, amazing and inspiring, part terrifying.

[00:58:48] Paul Roetzer: they, so their visions, I think part of the reason like I find this disturbing, is, there's a book called Pentagon's Brain. So if you're a longtime listener to this podcast, you may have heard me mention [00:59:00] Pentagon's Brain before. But, this isn't a new thing, to be trying to find ways to,

[00:59:08] Paul Roetzer: I don’t even know how to say this, create superhumans through machine interfaces with your brain. So if you, if this is an interesting topic to you, go, go read Pentagon's brain. I apologize in advance for what you will know and how you'll look at the world once you're done reading Pentagon's brain, but it's a, it's a factual story about DARPA, the Defense Advanced Research Project Agency and their work to basically emulate the human brain to create super soldiers.

[00:59:37] Paul Roetzer: So. The thing that worries me about this is DARPA does a lot of things. for the U. S. government, for the Department of Defense, that is under the guises of, medical intervention, like to, to help soldiers who have been harmed and things. But that isn't their mission. Their mission is the protection of the US

[00:59:57] Paul Roetzer: citizens, like of the United States of America. [01:00:00] And everything they build is to protect the United States, defenses and, and, and weapons. So any other benefit is, is actually just

[01:00:09] Paul Roetzer: like. PR. So when you read the mission statement of Neuralink, it is create a generalized brain interface to restore autonomy to those with unmet medical needs today. So again, the guise of this is all about helping people, which it will. But the second part of that mission statement is and unlock human potential tomorrow.

[01:00:30] Paul Roetzer: So if you ever listen to Elon Musk talk about humans, he sees deficiencies. in humans, and that our thoughts move faster than our words, when they come out of our mouth or when we type them. And he thinks, and this company thinks, that we're going to end up in a future where the only way to stay competitive with super intelligent AI is to become them.

[01:00:53] Paul Roetzer: Like, and so that, it's weird. And again, like, I always say to people, I always warn people when I go give talks, like, it's going to get weird for a little [01:01:00] bit. Sometimes on this podcast, it's got to get a little weird. And you need to know that in addition to, running X/Twitter and Tesla and SpaceX and the Boring Company and all these other things, Elon Musk has a company, where he believes that part of the way you protect the future of humanity is to merge humans with machines and Neuralink is his path to do that.

[01:01:23] Paul Roetzer: So it may help with some conditions right now, but that isn't why the company is being built. So. I feel like I've apologized like four times already today for like, if this is new to you, I'm sorry, go get like another cup of coffee and just go back to your normal daily life and forget about some of this stuff.

[01:01:42] Paul Roetzer: But I, there's a lot of people I talk to who are fascinated by this sort of stuff. So here you go. You get a little bit of this when you listen to the podcast too.

[01:01:52] Mike Kaput: We'll always mix in a dose of sci fi, hopefully. All right, our next rapid fire topic.

[01:01:59] Paul Roetzer: Back to [01:02:00] marketing, sales, and customer service.

[01:02:00] Mike Kaput: Yeah, back to our regularly scheduled programming that isn't going to help you lose sleep at night,

[01:02:07] Paul Roetzer: hopefully.

[01:02:07] Paul Roetzer: I got asked when I was at the thing in Arizona last week, I must have had five different people ask me the question. How do you sleep at night?

[01:02:13] Paul Roetzer: Like they They were asking questions about this stuff. And so we talk about it. and They're like, Oh my God, like, can we change the subject?

Microsoft Copilot for Sales and Copilot for Service are now generally available

[01:02:20] Mike Kaput: Well, hopefully some sales people are going to sleep a lot easier because. Microsoft just released a product that makes AI for sales and using it to make your job better a lot easier. They just released Copilot for Sales, which is their AI assistant for sales enabled use cases. And this AI assistant works right in Microsoft apps to do things like summarize emails and meetings in Outlook and Teams.

[01:02:49] Mike Kaput: Add leads and update CRM records directly from Outlook. Write emails in Outlook using your data from your CRM. Surface buying intent, budget, authority, need, timing, write in [01:03:00] Outlook. Create pitch decks and create data visualizations. You can also get sales tips, info, and answers to customer questions while you're on a Teams call.

[01:03:11] Mike Kaput: So for 50 per month per user, your team gets powerful sales AI right in the apps that you already use. at the same time, Microsoft has also, in the same announcement, released Copilot for Service, which unlocks similar AI powered use cases. In customer service says Microsoft Copilot for service unlocks an organization's trusted knowledge to accelerate onboarding and case resolution, improve efficiency, and automate tasks for agents in their flow of work without costly development time.

[01:03:48] Mike Kaput: Organizations can simply point to their data and in a few minutes unlock generative AI powered conversations across their knowledge bases. So, Paul, this really aligns [01:04:00] pretty well, I think, with the first main topic we're talking about, because it sounds like any one of these capabilities in sales or service could be Potentially save companies thousands of dollars per year.

[01:04:11] Mike Kaput: I know I'm less familiar with service, but I know in sales alone with some of the companies we talk to and work with on AI use cases, some of these are consuming dozens, if not hundreds of hours of employee time each month. What did you think of this?

[01:04:27] Paul Roetzer: Yeah, I agree. I think you just go, go re listen to the first segment. You know, look at your sales team, your service team, find the repetitive data driven tasks that they're doing that these tools now enable. Go get a pilot group of people to test it with three to five specific use cases. Pick a few of the things that it's able to do.

[01:04:45] Paul Roetzer: run it over 30, 60, 90 days, benchmark performance before and after make your business case. Like don't go buy for everybody in sales and not teach them. Follow the process we talked about, the three steps we talked about. the other thing that comes to mind is like, [01:05:00] if, if your organization has Salesforce or maybe HubSpot sales or like whatever your CRM is.

[01:05:06] Paul Roetzer: You're looking at this saying, Oh, Ooh, that would be really nice. Like, do we, does HubSpot going to build those capabilities too? Do they have those capabilities? So I think it, you know, when advancements like this happen, it cause you to step back and take a look at your technology stack and say. You know, is, should we be making a change or is the company we're with going to be building these same things?

[01:05:25] Paul Roetzer: We know Salesforce is building them. I know HubSpot's making efforts to build things. so yeah, I don't know. I mean, again, if, if you're in sales customer service, you have to look at this

[01:05:36] Paul Roetzer: stuff. Like the savings could be massive. The productivity gains could be enormous. So you owe it to your company to Do some homework, but do it in a strategic way.

[01:05:44] Paul Roetzer: Like we talked about, educate people, get a pilot test going, you know, get someone to oversee that program, make sure adoption is being done the right way, and then make a business case for whether you're going to you know, go all in and scale it up.

Amazon announces Rufus, a new Gen AI-powered shopping experience

[01:05:56] Mike Kaput: All right. And our last topic today, we [01:06:00] have a new AI tool that could change the way we all shop, at

[01:06:03] Mike Kaput: least if we use Amazon because Amazon just released an AI shopping assistant called Rufus Rufuss is trained on Amazon's product catalog and customer reviews and you can ask it anything as you shop on Amazon's site or in the app. So you can ask Rufuss to do things like conduct product research, like what should I consider when buying headphones, compare products and categories.

[01:06:29] Mike Kaput: You can ask questions about products, discover products based on context, so like you could ask, what do I need for cold weather golf? You can get recommendations. such as, what are the best dinosaur toys for a five year old? So essentially you're using conversational AI chat, much like you would with ChatGPT, layered over Amazon's entire product catalog and customer reviews.

[01:06:52] Mike Kaput: So this is currently in beta, but Amazon says it's rolling out to more US customers in the coming weeks. So Paul, what did you make [01:07:00] of Rufus when you were reading this announcement?

[01:07:02] Paul Roetzer: I think it's interesting to see if people use it. You know, it's like, okay, is this instead of Google search? Is it instead of perplexity? Like again, this complication of all of these tools doing roughly the same things. But in this case, it's trained on a specific dataset, do, do Amazon customers want it?

[01:07:19] Paul Roetzer: I don't know. Like it'll be really interesting to see how it plays out. I think it's a logical

[01:07:25] Paul Roetzer: And I think looking at your proprietary datasets and saying, how can we infuse generative AI into this to improve the customer experience? This is a smart play. Now I just wait and see what the adoption is.

[01:07:35] Paul Roetzer: Total side note. I went to see if I could test it as we're sitting here, like if it was live yet. And so I searched Amazon Rufus and apparently I'm guessing this is where the name came from. This is just fun trivia. there's a webpage on Amazon called Our Friend Rufus. And I was like, Oh, they, this is cute.

[01:07:51] Paul Roetzer: They made a back story of. their AI and it's like, no, it was actually a dog. So it says for, for years, Rufus was a fixture at amazon. com, dating back to the early days in the [01:08:00] company's history. He belonged to Amazon's former editor in chief and principle engineer, and he accompanied them to the office every day.

[01:08:06] Paul Roetzer: He's been affectionately called Amazon's shortest volunteer worker. And then it has a little bio on Rufus, including his vital statistics that he was born in 1994. Perfect pause, kindergarten graduate. So,

[01:08:21] Paul Roetzer: I, think the name Rufus may have actually come from a dog that used to walk the halls of Amazon in its early days.

[01:08:26] Paul Roetzer: And Rufus passed away in 2009. Just, rest in peace, Rufus. Oh, Rufus may be immortalized.he has been immortalized, yes. He fetches things, like, it makes sense. It's good branding if that's where it came from.

[01:08:39] Mike Kaput: You know, honestly, as much as I love, science fiction and science fact, I am kind of glad to get an AI tool. that's, like not. Blatantly just named after not Jarvis.

[01:08:49] Mike Kaput: Yeah. Yeah. Like, cause that's where all the names so far. where you're like, Oh, okay. I get it.

[01:08:55] Paul Roetzer: I don't know if this is where it came from, but we're going to like, we're going to end this podcast today, believing that [01:09:00] Rufus is named after a wonderful dog that used to roam the hallways of Amazon. And we're going to end on a positive note. after our dystopian conversations from the tape. I love it.

[01:09:08] Mike Kaput: Well, Paul, thanks as always for breaking down what is going on in the world of AI this week. I would also encourage people to check out further our newsletter, go to marketingaiinstitute. marketingaiinstitute.com/newsletter, because that is where we cover even more stories each and every week, including both what we've covered on this episode and all the stories that didn't make it due to time constraints.

[01:09:33] Mike Kaput: There's like a dozen this week. dozen

[01:09:34] Mike Kaput: a dozen of them easily.

[01:09:36] Mike Kaput: And that's increasingly the case. So you can get in one easy to read email digest every single week, a complete picture of what's going on in AI. So Paul,

[01:09:47] Paul Roetzer: thanks again.

[01:09:48] Paul Roetzer: Thank you, Mike. Talk with you next week. Thanks for listening to the Marketing AI Show. If you like what you heard, you can subscribe on your favorite podcast app, [01:10:00] and if you're ready to continue your learning, head over to www.marketingaiinstitute.com.

[01:10:04] Paul Roetzer: Be sure to subscribe to our weekly newsletter, check out our free monthly webinars, and explore dozens of online courses and professional certifications.

[01:10:12] Paul Roetzer: Until next time, stay curious and explore AI.