This week, you might notice something different about the podcast—we’ve changed the name to The Artificial Intelligence Show and dropped the word ‘Marketing’ from the title.
Since our first episode, our audience has expanded beyond the marketing industry. In part of our evolution and our movement toward AI Literacy for All, we transitioned to our new name to simplify and align with our current audience. Nothing else has changed beyond our name; we will still follow the same format, release schedule, and dedication to exploring the constant changes in the world of AI.
Welcome to Episode 83 of the Artificial Intelligence Show! Today, we delve into Google Bard's rebranding to Gemini, explore the inevitability of AI Agents, and discuss Sam Altman's trillion-dollar request to reshape AI.
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.
The course provides an overview of the basics of AI and how it works; deep dives into practical frameworks to get started; a brand new generative AI 101 course, and dozens of sample use cases and AI tools that will accelerate your adoption and success.
The series includes approximately 9 hours of content, complete with on-demand short courses, quizzes, downloadable resources, a final exam, and a professional certificate upon completion
The courses were recorded in January 2024, so you’ll get all the latest information.
Go to PilotingAI.com to learn more, and register today!
Listen Now
Watch the Video
Timestamps
00:07:03 — Meet Google Gemini
00:21:03 — AI Agents Are Coming
00:32:01 — Sam Altman seeks trillions of dollars to increase chip-building capacity
00:38:47 — Some investors are deciding not to invest in OpenAI and other AI startups
00:44:03 — A touchdown for AI Advertisements
00:47:21 — Finance worker duped by a deep fake call into $25M payout
00:50:12 — AI disruption to consulting firms like McKinsey
Summary
Meet Google Gemini
Google has announced that, from now on, Google Bard will be known simply as Google Gemini.
Google also announced Gemini Advanced, a paid subscription tier for $19.99 per month that gives you access to Google’s most powerful mode—Gemini Ultra. With Gemini Advanced, you get access to Ultra 1.0, which will also be available soon in Google Docs, Gmail, and other apps.
According to Google: “Gemini Advanced is far more capable at reasoning, following instructions, coding, and creative inspiration.”
Early tests indicate that this is true. Warton professor and AI expert Ethan Mollick, who has had beta access to Gemini Advanced for a month, says “Gemini Advanced is clearly a GPT-4 class model.”
He notes that, while Gemini Advanced does not obviously blow away GPT-4, it is roughly equivalent, though it has its own strengths and weaknesses.
Right now, you can get two months free when you sign up for Gemini Advanced. After your free trial, if you don’t want to spend $20 a month, you can still use a weaker version of the Gemini family of models, Gemini Pro, for free simply by going to gemini.google.com.
AI Agents Are Coming
It is becoming more apparent every day that AI agents are becoming a reality—fast. These AI agents are autonomous AI systems that can take actions for you on your behalf.
According to The Information, OpenAI is “developing a form of agent software to automate complex tasks by effectively taking over a customer’s device.” The agent would perform clicks, cursor movements, typing, and other actions required by humans to operate different apps, a person with knowledge of the effort told The Information.
There are also plenty of buzz-worthy startups trying to build AI agents.
Our friends at HyperWrite have been working on their Personal Assistant AI agent for some time now, and they just released Agent Trainer and Agent Studio. These give you the ability to train AI agents just by recording what you do on your screen.
Says Matt Shumer at HyperWrite on X:
“This represents a major step forward in usability and accessibility for AI Agents. No more prompting. Soon, HyperWrite users will be able to customize their Personal Assistant and teach it to automate repetitive tasks in a fraction of the time.”
Other startups are also trying to build completely autonomous AI assistants. Some notable ones include the new buzzworthy startup MultiOn and Adept (which has raised hundreds of millions of dollars).
Sam Altman Seeks Trillions of Dollars to Reshape AI
Sam Altman, CEO of OpenAI, is in talks to raise as much as $5-$7 trillion to dramatically increase the chip-building capacity needed to power the AI revolution.
Investors could include the United Arab Emirates government, according to reporting from the Wall Street Journal.
This comes after Altman was reportedly trying, late last year, to get OpenAI into the chip business—an effort that caused some controversy at OpenAI and is suspected, in part, to have motivated the failed coup against Altman.
The money would fund a project to boost the world’s chip-building capacity—and it would dwarf the current size of the global chip industry, which was $27B last year and expected to rise to $1T annually by 2030 according to the Journal.
Links Referenced in the Show
- Meet Google Gemini
- Bard becomes Gemini: Try Ultra 1.0 and a new mobile app today
- Google Gemini
- Release updates
- Google's Gemini Advanced: Tasting Notes and Implications
- Google’s Gemini is now in everything. Here’s how you can try it out.
- Introducing Gemini: our largest and most capable AI model
- AI Agents Are Coming
- OpenAI Shifts AI Battleground to Software That Operates Devices, Automates Tasks
- Agent Trainer and Agent Studio
- HyperWrite Personal Assistant
- Adept
- World of Bits, and What It Means to Marketing and Business
- Episode 35
- Sam Altman Seeks Trillions of Dollars to Reshape AI
- Sam Altman Seeks Trillions of Dollars to Reshape Business of Chips and AI
- Thoughts on Altman Raising Trillions
- Some OpenAI Investors Sit Out Share Sale, Reflecting AI Divide
- Gameday AI Ads
- Finance worker pays out $25 million after video call with deepfake ‘chief financial officer’
- AI disruption to consulting and market research
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: Large language models were always just the foundation for what comes next.
[00:00:04] Paul Roetzer: And so, when we talk about this exponential growth curve in technology and capabilities, AI agents are going to be at forefront of that. And 2024, I think we're going to start to see more and more applications coming to consumers and to businesses in this space.
[00:00:22] Welcome to the Artificial Intelligence Show, the podcast that helps your business grow smarter by making AI approachable and actionable. My name is Paul Roetzer. I'm the founder and CEO of Marketing AI Institute, and I'm your host. Each week, I'm joined by my co host, and Marketing AI Institute Chief Content Officer Mike Kaput as we break down all the AI news that matters and give you insights and perspectives that you can use to advance your company and your career. Join us as we accelerate AI literacy for all.
[00:00:59] Paul Roetzer: [00:01:00] Welcome to episode 83 of the Artificial Intelligence Show. I'm your host, Paul Roetzer, along with my co host, Mike Kaput. What's happening, Mike?
[00:01:09] Mike Kaput: Not much.
[00:01:10] Mike Kaput: you doing, Paul?
[00:01:11] Paul Roetzer: Good, you stay up for the Superbowl last night?
[00:01:14] Mike Kaput: I stayed up for some of it.
[00:01:17] Paul Roetzer: ended up around the whole I missed a bunch of ads. We'll about ads and AI ads in a minutes, but I missed a bunch of them.
[00:01:24] Paul Roetzer: I caught few, all right, So you may have noticed in the introduction here already slight change to name of the show. So, the show was created back 2018 19, whenever first started doing it.
[00:01:41] Paul Roetzer: was the Marketing AI show and, as many you listen may you may not be a marketer.
[00:01:50] Paul Roetzer: have, the show has and expanded, our audience has become way more diverse and reaching than just the marketing [00:02:00] industry. So we have CEOs, entrepreneurs, government leaders, educators, administrators. listening to the show. And so in part of our kind of evolution here, we talked about, in last episodes on, know, a bit shift to the mission of AI literacy for all, we, dropped marketing from the name. So just Artificial Intelligence Show kind of simplify things and be more aligned with. current audience and where things are going. Now, the key is nothing is changing it's still going to be the same show me and
[00:02:38] Paul Roetzer: same thing. main topics, rapid fire items. it's I've actually seen lot of our, audience mention like, I just don't even marketing when I say, you gotta listen to this show because people are referring. people the who aren't marketers. So Mike and will provide the same perspective on marketing, sales, service, business we always [00:03:00] have. We're just the marketing from name to better align, as I with the people who. Listen to this on a regular
[00:03:08] Paul Roetzer: So appreciate all of you listening. appreciate all marketers that with us from the beginning all you non marketers that, know, the practitioners business leaders and educational leaders government leaders out there who, find value in the show every week. we're grateful you being a part of it.
[00:03:24] Paul Roetzer: like I said, nothing is changing. Content's going to be same, going to be every Tuesday, going to drop. every episode, weekly. So stay tuned. going to be some other things we're going to be introducing as kind of this mission evolves this ad literacy all movement. stay tuned on that.
[00:03:42] Paul Roetzer: nothing more announce at the moment. right. So.
[00:03:47] Paul Roetzer: The first episode of artificial intelligence show, episode six or 83 overall, today's episode is to by Marketing AI Institute's Writers Suimmit which is presented by Jasper happening [00:04:00] virtually on Wednesday, March 6th, from 12 to PM Eastern time, following the success our inaugural writers in March, 2023, which had more than 4, 000 writers, editors,
[00:04:12] Paul Roetzer: content marketers were to present the second edition of Writers Summit.
[00:04:19] Paul Roetzer: agenda includes of AI in writing that I'm going to be presenting, AI tools platforms you know, which Mike is presenting. We're going to have AI insights from an attorney. you understand landscape, implications of copyright and both the the models the use of those models.
[00:04:37] Paul Roetzer: writing in the enterprise. going to through a panel of opportunities obstacles adopting AI for content creation, then we're going close it out with an AI in action rapid fire session showing bunch of tech demos for writers and Those are non sponsored So this is legitimate, like applications that we use internally.
[00:04:56] Paul Roetzer: Tech we're excited about. going to show bunch of demos there [00:05:00] then there'll be an Ask Me Anything session at end of that. best part, thanks Jasper, it is free attend. there's option for attending live event and there's also a on demand option and a registration as well. Go to AIWriterSummit.com for more
[00:05:18] Paul Roetzer: You can also it under the Marketing AI Institute site events. And episode is also brought to by the new Piloting AI 2024 course series.
[00:05:31] Paul Roetzer: This series is collection 18 demand courses as step by learning path beginners at all levels interns to CMOs.
[00:05:40] Paul Roetzer: Mike take you a that includes an overview of basics of aI and it works. Deep into frameworks get started,
[00:05:49] Paul Roetzer: A brand new generative AI 101 course, and dozens sample use cases and AI tools that will accelerate adoption and success. The series includes about 9 [00:06:00] hours of complete on demand short courses, quizzes, downloadable resources, a final exam, a professional certificate upon completion.Mike and I recorded
[00:06:11] Paul Roetzer: of courses in January 2024, so you will get the latest information. Go to pilotingai.com to learn more register today.
[00:06:21] Paul Roetzer: All right, Mike, got a flight to San Diego lets get this going. It is february 12th. we're, 10, 20 AM Eastern time again, timestamps, because a lot chatter starting to emerge that might be hearing OpenAI point in the near on like GPT kind of news.
[00:06:40] Paul Roetzer: So nothing official. Again, like last week, we were hearing increasing murmurs GeminiUltra that ended up happening last week. So usually you can of to get a sense some things are in the works. So, anything happens to break this week, are recording Monday, February [00:07:00] 12th at 10 AM.
[00:07:01] Paul Roetzer: All let's go.
Meet Google Gemini
[00:07:03] Mike Kaput: All right. So our first big topic today, Paul, is the one you just alluded to, which is Google has announced that from now on, Google Bard will be known simply as Google Gemini.
[00:07:15] Mike Kaput: Now in tandem with this renaming, Google announced Gemini Advanced, which is a paid subscription tier for 19. 99 per month that gives you access to Google's most powerful model, Gemini Ultra. Now with Gemini Advanced, you get access to what they're calling Ultra 1. 0, which will also be available soon in Google Docs.
[00:07:39] Mike Kaput: Gmail, and other apps. Now, according to Google, quote, Gemini Advanced is far more capable at reasoning, following instructions, coding, and creative inspiration. Now, early tests seem to indicate that that's pretty true. Wharton professor and AI expert Ethan Mollick, who has had beta [00:08:00] access to Gemini Advanced for about a month, he wrote a post in which he said that Gemini Advanced is clearly a GPT 4 class model, and he notes that while Gemini Advanced doesn't obviously blow away GPT 4, it is roughly equivalent, though it has its own strengths and weaknesses.
[00:08:20] Mike Kaput: Now, right now, you can get two months free when you sign up for a Gemini Advanced subscription, but if you don't want to pony up 20 bucks a month, you can still use a weaker version of the Gemini family of models, Gemini Pro. simply by going to Gemini. Google. com. You can use that totally for free. So Paul, first up is Gemini Advanced/Gemini Ultra. the GPT 4 competitor that we've been waiting for?
[00:08:50] Paul Roetzer: Does seem to be, you know, this, so we did our podcast last week on Monday, which would have been, I think we recorded on the 5th. doing my math right. And then we had [00:09:00] an intro to AI class last week on Thursday, I want to say, and we had like 1800 people
[00:09:05] Paul Roetzer: And that morning at like 8 AM, I was at the cafe. I was at Joe's deli, one of our favorite spots around town and drinking a coffee and I was getting ready for the intro presentation and I see the tweet alert, like Gemini, Altra, Nalvela. I was like, ah, man. I still gotta build this deck. And so I kind of raced in and created an account, got the subscription on my Gmail, you can't get through your corporate account yet.
[00:09:31] Paul Roetzer: And I started kind of playing around with it. So, yeah, I just kind of threw my morning for a nice wrinkle that day, but I wanted to kind of take a step back because honestly, this is so confusing. Like. Which Gemini models do we have? What's it called? did they come
[00:09:48] Paul Roetzer: out? Which ones are available?
[00:09:49] Paul Roetzer: Which ones aren't? So I actually went back to our notes and again, we've been talking about this. I feel like I actually went back to the last week's podcast brief that Mike and I have before we go [00:10:00] on. And I was like, didn't we talk about this last week? And then I realized like, Oh no, it came out after the fact, but we had alluded it might.
[00:10:06] Paul Roetzer: So December 6, we'll rewind so like two months ago and six days, two months and six days ago, Google announced
[00:10:15] Paul Roetzer: Gemini, so on December 6th, 2023, a blog post from Sundar Pichai, the CEO of Alphabet and Google, and Demis Hassabis, the lead of Google DeepMind, and kind of the Head of AI at Google.
[00:10:29] Paul Roetzer: So they did a joint. blog post announcing that Gemini was coming.
[00:10:34] Paul Roetzer: So this, and I think Gemini Pro maybe roughly available that day. This is again it gets really confusing, but, I thought it would be helpful take a step back real quick and get context of what exactly is going on with Gemini. Why is this a big So in that blog post, which we will put in the show notes, if you, if you didn't read it, Um, there is a, it starts with perspective from Sundar and then it has perspective from Demis.
[00:10:59] Paul Roetzer: And [00:11:00] so Sundar says, every technology shift is an opportunity to advance scientific discovery, accelerate human progress, and improve lives. I believe the transition we are seeing right now with AI will be the profound in our lifetimes, far bigger than the shift to mobile or to the web before it. AI
[00:11:18] Paul Roetzer: has the potential to create opportunities from the everyday to the extraordinary for people everywhere.
[00:11:25] Paul Roetzer: It will bring new waves of innovation and economic progress and drive knowledge, learning, creativity, and productivity on a scale we haven't seen before. He then goes on to say, Now, we're taking the next step on our journey with Gemini, our most capable and general model yet, with state of the art performance across many leading benchmarks.
[00:11:47] Paul Roetzer: Our first version, Gemini 1. 0, is optimized for different sizes, Ultra, Pro, and Nano.
[00:11:55] Paul Roetzer: These are the first models of the Gemini era and the first realization of the [00:12:00] vision we had when we formed Google DeepMind earlier this year. He's referring to when Google Brain and DeepMind merged into a single research lab at Google earlier in 2023. This new era of models represents one of the biggest science and engineering efforts we've undertaken as a company. I'm genuinely excited what's ahead and for the opportunities AI I unlock for people everywhere. Okay, so keep the macro level. from Sundar in your mind and I'll come back to why that matters in a moment.
[00:12:31] Paul Roetzer: Demis in that post said, for a long time we've wanted to build generation of AI models inspired by the way people understand and interact with the world. AI that feels less like smart piece of software and more like something useful and intuitive, an expert helper or assistant. Today, again, this is December 2023. 2023.
[00:12:52] Paul Roetzer: We are a step closer to this vision as we introduced Gemini, the most capable and general model we've ever built. Gemini is the result [00:13:00] of large scale collaborative efforts by teams across Google, including our colleagues at Google Research. It was built from the ground up be multimodal, which means it can generalize and seamlessly understand, operate, operate across and combine different types of information, including text, code, audio, image, and video.
[00:13:18] Paul Roetzer: Okay, so that's kind of like the macro level. At that moment in time, Bard was still Bard, and now, so they've now kind of taken the complexity of the, you got gemini, you got Bard, you got Duet AI, which is the workspace version of all of this, and then you have Vertex AI, which is like, the access to the APIs, I think, basically.
[00:13:42] Paul Roetzer: So if you're a developer, you're working with Vertex AI. So it gets really kind complicated to follow. So I like the fact that they're rebranding BARD to Gemini. I think that makes sense. I actually tweeted at Jack Krawczyk, who is the product
[00:13:58] Paul Roetzer: lead for BARD slash [00:14:00] Gemini from Google. He's very active on Twitter.
[00:14:03] Paul Roetzer: I would highly recommend following him. It's just at Jack K. And he, he's very engaged with users and he's constantly asking for feedback on Barge last Gemini. So he's a good guy to follow, but I actually asked him, I I liked the rebrand. is there any plans to, when I say I like the rebrand Gemini now, can we change Duet AI for Workspace to Gemini for Work and clear up all the confusion?
[00:14:30] Paul Roetzer: And he said, yes, part of AI Premium will be access to Gemini in Gmail and Docs, formerly known as Duet AI. AI.
[00:14:37] Paul Roetzer: So. As of this moment, Duet AI is still the Workspace brand, as far as I know, at least when I looked this morning. But it sounds like they're going to drop that as well and just go with
[00:14:47] Paul Roetzer: Gemini brand.
[00:14:48] Paul Roetzer: So, I did subscribe, through my personal Gmail, so I have access to Gemini Advanced now, which again is using their Ultra 1. 0 model.
[00:14:58] Paul Roetzer: It also is [00:15:00] available on my iPhone. So I have the Google app and can now just toggle between regular Google search and Gemini advanced, which is kind of
[00:15:07] Paul Roetzer: So I like that. And I know on Android you can get the, an app dedicated to Gemini. So you can definitely get in there and start doing it. They're giving you two months free. They're trying to get user trials here. So it's 19. 99 a month, but you get two months free. It seems like initially. the main differentiator is that you can connect to google flights, Google hotels, Google maps, Google workspace, and YouTube.
[00:15:32] Paul Roetzer: So there's what they call extensions and you can turn these on.
[00:15:36] Paul Roetzer: Um, now will say when you turn them on, they are very clear that they are going to use what you give them. So you actually, their privacy statement says. Please don't enter confidential information in conversations or any data you wouldn't want a reviewer to see or Google to use to improve our product services and machine technologies.
[00:15:59] Paul Roetzer: And there's [00:16:00] literally a warning that pops up the top of your Google advanced screen that says, Your conversations are processed by human reviewers to improve technologies powering Gemini apps. don't enter anything you wouldn't want a review, want reviewed or used. So just kind of some things to keep in mind.
[00:16:18] Paul Roetzer: My experience so far, what I've been finding myself doing, is when I have prompt, I will actually go through and I have Gemini Advanced Open, ChatGPT, and Perplexity. Those are kind of my three Main ones that I'm exploring. Anthropic Claw, it's, you know, still great and everything. But those are the three I'm playing with and so I will often actually give the same prompt to all three of them and I will kind of see on an ongoing basis how they perform against each other in certain scenarios.
[00:16:48] Paul Roetzer: So I don't have enough deep research yet. Not like Mollick's article.
[00:16:54] Paul Roetzer: But it's definitely more powerful. Like, you can really see it. [00:17:00] And so, like, my final thought here is, when we go back to that context of what Sundar and Demis wrote on December 6th when they introduced Gemini, it really sticks in my mind because At the of IthaMollick's post, if you hung around for the close, he said that the question of why it doesn't clearly beat GPT 4 is really interesting and perhaps consequential.
[00:17:23] Paul Roetzer: Now, again, keep in mind, GPT 4 is Now, been in the world for almost a year, it was March 2023, and it was built, it was finished 7 months before it came out. So GPT 4 is basically like 19 or 20 months old, in terms of a technology. So the fact that Google couldn't beat it is kind of weird with as fast as this stuff advances and improves that GPT 4 is still the top model after almost 20 months in existence is really perplexing honestly like if you think about it and so [00:18:00] Ethan presented four potential scenarios as to why that is that that Gemini altar isn't clearly Beating GPT-4 yet, so he said one G.
[00:18:11] Paul Roetzer: PT four class models are about as good as AI gets using large language model technology, suggesting that the exponential change in AI capabilities is ending. He says, I think this is unlikely, but possible number two. Google needed a model to compete with GPT 4, so they trained up Gemini to that level and stopped. More advanced models are coming soon. 3. OpenAI has some special sauce that no other company can and they are the only ones who can easily achieve GPT 4 Plus abilities. attempt was
[00:18:46] Paul Roetzer: the best they could do without knowing the OpenAI secret. And number 4. It is a coincidence that this model happens to be so close in ability to GPT 4.
[00:18:56] Paul Roetzer: and we learn nothing from this. He says, I think two is most [00:19:00] likely, but I have no idea if that is true.
[00:19:02] Paul Roetzer: So if you go back to You know, this is the beginning, all the things that Demis and Sundar said, I feel quite strongly that number two is the answer. I think that what we're seeing one isn't even the most powerful version of Ultra. Like I'm almost. positive that they have far more capabilities than we're accessing through Gemini Advanced at the moment. They haven't even integrated all the modalities into this thing yet. So, I think 1. They're probably holding back some capabilities. to they likely could have continued training this, and maybe they did, and just didn't release it.
[00:19:43] Paul Roetzer: So, I do think that we're really just at the very beginning of what these things are going to be able to do. to do. I think we're going to probably see Llama3 from Meta at some point, maybe in Q1. It seems more and more like we may see ChatGPT, improve with [00:20:00] GPT 5 or some other iteration. Even they were vague tweeting last night, the ChatGPT app said something like, think step by step, like they're starting to Sort of like tweet things that imply something else is around the corner. So I feel like this is a really important thing.
[00:20:20] Paul Roetzer: I wouldn't, if you're paying 20 bucks a month for ChatGPT I guess my big takeaway here, if you're already paying 20 bucks a month for ChatGPT I don't think you're going to get some leap forward in capabilities. For another $20 a month from this. But as we talked about last week, Google has massive distribution.
[00:20:36] Paul Roetzer: You're already in Google search all day long. You're already in Chrome all day long, and I think that distribution may end up being one of the biggest factors for adoption for them. so I will continue to pay for it because I'm experimenting with it. Not because I found some application here for Gemini Advanced that gives me some amazing capabilities I don't already have [00:21:00] through ChatGPT.
[00:21:00] Paul Roetzer: It's, it's research for us.
AI Agents Are Coming
[00:21:03] Mike Kaput: So in our second big topic this week, it is becoming more apparent every single day that AI agents are becoming a reality and they're becoming a reality fast, by AI agents, we mean. Autonomous AI systems that can take actions for you on your behalf. So according to a recent report in the information, OpenAI is, quote, developing a form of agent software to automate complex tasks by effectively taking over a customer's device.
[00:21:35] Mike Kaput: This agent would perform clicks, cursor movements, typing, and other actions required by humans to operate different apps. And that's according to a person with knowledge of the effort who the information interviewed. Now, there are also plenty of buzzworthy startups that are trying to build these types of AI agents.
[00:21:55] Mike Kaput: Our friends at HyperWrite have been working on their personal assistant AI agent for [00:22:00] some time now. And they released Agent Trainer and Agent Studio, which are new features that give you the ability to train AI agents just by recording what you're doing on your screen for any given workflow or interaction with an app.
[00:22:17] Mike Kaput: Uh, Says Matt Shumer at HyperWrite on X, this represents a major step forward in usability and accessibility for AI agents. No more prompting. Soon HyperWrite users will be able to customize their personal assistant and teach it to automate repetitive tasks in a fraction of the time. They're not the only startup working on this.
[00:22:36] Mike Kaput: There are companies, like the new buzzworthy startup MultiOn and Adept, which is an existing player who's raised hundreds of millions of dollars. So, Paul, it's actually interesting timing here because it's been almost a year since you first published an article on our website called World of Bits and basically predicted this imminent [00:23:00] rise of AI agents.
[00:23:01] Mike Kaput: And in that article you said, quote, When you start connecting the dots, it appears that we are moving toward a world in which AI agents will not only retrieve and present information and answers, but have the ability to take actions in the digital world. This changes things, likely in ways for which marketers, business leaders, software entrepreneurs, and humans in general are not prepared. you walk us a bit through that post, kind of why you wrote it, why it matters today, and what's changed, if anything, since you wrote it?
[00:23:34] Paul Roetzer: Yeah, I think, so we talked about this in episode 35, so if you want to, like, go back and rewind and kind of listen to the whole thing, we'll also put link in the show notes, but for the original article. But I think it is helpful to go back and kind of give the context here. So
[00:23:53] Paul Roetzer: If we go back to that article, I'll just kind of read a few excerpts for you because it gives really good context as to what's going [00:24:00] on and kind of where this is So the article said, we are so caught up right now figuring out aI writing tools and large language models that most marketing and business leaders, as well as SAS executive software as a service executives, And investors are missing the bigger picture. This is all just the foundation for what comes next.
[00:24:18] Paul Roetzer: Let's look at email marketing as a practical example of what I mean. Imagine you want to send an email promoting an upcoming event, product launch promotion, but rather than a series of clicks and manual entries, you simply spoke or typed prompts for what you wanted the machine, AKA the AI to do. I did later in the post share the example of HubSpot.
[00:24:37] Paul Roetzer: Like if I wanted to send an email in HubSpot, it takes a minimum of 21 clicks. So this example of I can just give a voice prompt or a text prompt instead of clicking around 21 times within HubSpot. So the article went on to say, I'm not talking about simple information retrieval and natural language generation such as a chat feature that responds queries and prompts. I'm saying that the AI will [00:25:00] have
[00:25:00] Paul Roetzer: the ability to perform actions, clicks, form fills, etc. the same way as humans. Based on a collection of public AI research papers related to a concept called World of Bits, and in light of recent events and milestones in the AI industry, including legendary AI researcher Andrej Karpathy announcing his return to open AI, which was in February of 2023, it appears the capabilities for AI systems to use a keyboard and mouse are being developed in major AI research labs right now.
[00:25:30] Paul Roetzer: This has been attempted in the past, but recent advancements in language AI appear to be bringing this closer to reality. If AI develops these abilities at scale, the user experience of every software company will have to be reimagined, and it will have profound impacts on productivity, the economy, and human labor.
[00:25:51] Paul Roetzer: Later in the article, I went on to say, In a 2017 research paper called World of Bits, an open domain platform for web based agents, [00:26:00] Karpathy and other authors explored the potential of AI agents to complete tasks such as booking flights and completing forms through simulated usage of a keyboard and mouse.
[00:26:10] Paul Roetzer: They made progress, but obstacles remain. In conclusion of that paper, it said, In this paper, we introduced World of Bits, a platform that allows agents to complete web tasks with keyboard and mouse actions. Unlike most existing reinforcement learning platforms, World of Bits offers the opportunity to tackle realistic, realistic tasks at scale.
[00:26:32] Paul Roetzer: We showed that while standard supervised and reinforcement learning techniques can be applied to achieve adequate across these
[00:26:38] Paul Roetzer: environments, the gap between agents and humans remains large and welcomes additional modeling advances. It was around that time that Karpathy left OpenAI and went to head up AI and computer vision at Tesla.
[00:26:54] Paul Roetzer: Then, in October 2022, so a few months before we're talking about in [00:27:00] February, he did an interview on the Lex Fridman podcast, which Mike and I are both huge fans of. And in that interview, he implied that those barriers that they talked about in the conclusion to the 2017 paper were coming down. So, Fridman said, you briefly worked on a project called World of Bits, training a reinforcement learning system take actions on the internet versus just consuming the internet like we talked about.
[00:27:24] Paul Roetzer: He went on to say, do you think there's a future for that kind of system interacting with the internet to help learning? Karpathy said, yes, I think that's probably the final frontier for a lot of these models. When I was at OpenAI, I was working on World of Bits, and basically it was the idea of giving them access the keyboard and mouse.
[00:27:40] Paul Roetzer: And the idea is basically you perceive the input of the screen and you're able to take actions. He went on to say, now to your question as to what I learned, it's interesting because World of Bits was basically too early at open AI. This is around 2015 when they were working He did then go on to say, it is time to revisit that.
[00:27:59] Paul Roetzer: And open AI [00:28:00] is interested in this. Companies like Adept are interested in this and so on. And the idea is coming back because the interface is very powerful. But now you're not training an agent from scratch. You're, you are taking the GPT as an initialization. This is the real key part as to kind of where we are today.
[00:28:17] Paul Roetzer: So GPT is pre trained on all of the text. And it understands what a booking is, it understands what a submit is, it understands quite a bit more. And so it already has those representations. They are very powerful and that makes all the training significantly more efficient and makes the problem tractable. And if you remember a few episodes back, we talked about Karpathy's YouTube video, where he talked about future of AI agents. And in that, when he said, what are kind of the next frontier of large language models, he said it can use the existing software infrastructure, like a calculator, a mouse, a keyboard.
[00:28:55] Paul Roetzer: Then a couple quick notes. The one thing I found interesting as was, you know, kind of follow along here, a guy [00:29:00] named Ben Newhouse, who works on agents OpenAI, tweeted on January 25th, that he was, I'm hiring at OpenAI. We're building what I think could be an industry defining zero to one product that leverages the latest and greatest from our upcoming models. you'd like product, deep technological challenges and writing the future, my DMs are open. So he was recruiting. But the thing that caught my attention was a guy named Peter Wellender, who's VP of product at OpenAI. retweeted that again on January 25th and said this product will change everything.
[00:29:39] Paul Roetzer: So OpenAI's VPs don't usually tweet things like this product will change everything unless they actually believe it will. And so I believe they're referring to their progress on AI agents. And so that before we even saw the information article last week was sort of my tip off that they are making significant progress.[00:30:00]
[00:30:00] Paul Roetzer: And so I think the key takeaway for everyone here is large language models were always just the foundation for what comes next. They have unlocked the ability to build AI agents. They will unlock the next iteration of robotics because the robots can embody the language models, the intelligence.
[00:30:20] Paul Roetzer: And so, when we talk about this exponential growth curve in technology and capabilities, AI agents are going to be at forefront of that.
[00:30:28] Paul Roetzer: And 2024, I think we're going to start to see more and more applications coming to consumers and to businesses in this space.
[00:30:38] Mike Kaput: So to kind of wrap this up, if I'm a marketer, marketing leader, business leader, like how can I even start to prepare for the near term implications of how much this will change?
[00:30:51] Paul Roetzer: Well, if
[00:30:52] Paul Roetzer: you're willing to give access to everything you do on your computer, you can probably start. testing them [00:31:00] with something like a hyperwrite or multi on.
[00:31:03] Paul Roetzer: I personally am
[00:31:04] Paul Roetzer: not. So I follow along in the space and listen closely to what's happening and what's going on in the research labs and watch the demos, but I'm not at a point where I would personally trust these companies to have access to that.
[00:31:19] Paul Roetzer: But like I said, in the past episodes, like once Google does this. Or Apple, and they already have all my data, I'm far more likely to say, all right, I'll give it a go.
[00:31:31]
[00:31:31] Paul Roetzer: but for me personally, I'm not willing to give up that privacy and data to startups at the moment. so I think again, it's kind of one of those, the law of uneven AI distribution that I wrote, you know, a year plus ago, like we're all going to have the ability to have the benefits, but you have to.
[00:31:50] Paul Roetzer: be willing to accept what you have to give up to get the And so in the case of AI agents, personally, I'm not willing to give up the privacy, that it take to, [00:32:00] to start testing these
Sam Altman seeks trillions of dollars to increase chip-building capacity
[00:32:01] Mike Kaput: So in our third, big topic this week, another OpenAI, themed topic here, because Sam OpenAI, is actually reported to be in talks
[00:32:14] Mike Kaput: to raise as much as 5 to 7 trillion dollars to dramatically increase the chip building capacity needed to power the AI revolution. investors included in raising this money could include the United Arab Emirates government, according to reporting from the Wall Street Journal that outlined the beginnings of this project.
[00:32:37] Mike Kaput: Now, this comes after Altman was reportedly trying late last year to get OpenAI into the chip business, which is an effort that caused some controversy at OpenAI and is suspected, at least in part, to have motivated all the drama and the failed coup against Altman. Now, details are pretty scant at the moment, but this money [00:33:00] supposedly would fund a project to boost the world's chip building capacity, and what's crazy about this is The numbers being talked about with the 5 to 7 trillion being raised would just dwarf the current size of the global chip industry.
[00:33:14] Mike Kaput: Um, it is expected to be worth 1 trillion dollars annually by 2030 according to estimates from the Wall Street Journal.
[00:33:23] Mike Kaput: So Paul, why, first up, are we even talking about numbers of this size when it comes to chipmaking? Like, why is chipmaking so important to AI innovation and progress?
[00:33:35] Paul Roetzer: So the, all the generative AI capabilities that we're all experiencing today are powered by largely NVIDIA chips which are Made in Taiwan. so there's been know, kind of a, certainly an effort in the United States. probably, you know, more broadly from a global perspective to diversify and expand the chip industry.
[00:33:58] Paul Roetzer: So it's not centered in a [00:34:00] single country
[00:34:01] Paul Roetzer: more, more, I guess, volume,
[00:34:05] Paul Roetzer: more innovation happening outside of fabrication plants in Taiwan. So there's a much bigger story there, but like this kind of money, it's not just chips. Like they're looking at,
[00:34:18] Paul Roetzer: basically what Sam is doing is they're looking out ahead and saying, okay, in seven to 10 years to do what we're envisioning at OpenAI, whatever it is that their long term vision believes AI is to be capable of They're obviously envisioning a world where we need. way more, ability to build chips. We need significantly greater clean energy and we need a lot more data and in addition to everything else, in addition to talent that's able to build and enable these advancements. So he tweeted. The afternoon before the Wall Street Journal came out, so obviously he was aware this article
[00:34:57] Paul Roetzer: coming, I'm sure they, you know, checked them [00:35:00] verify information Sohe sort of re tweet the article and we believe the world needs more AI infrastructure, fab capacity, energy, data centers, etc. then people are currently planning to build. Building massive scale AI infrastructure and a resilient supply chain is crucial to economic competitiveness.
[00:35:21] Paul Roetzer: OpenAI will try to help. so again, we don't know much. They, they haven't, you know, confirmed these kinds details or these numbers. the but the, three things that I observed when I read this was one, I would assume these efforts, uh, may have something to do with his temporary ouster sure.
[00:35:44] Paul Roetzer: If this was all happening under the OpenAI nonprofit umbrella.
[00:35:49] Paul Roetzer: So if Sam was actively going out and talking to people about this kind of raise to this rapid acceleration, and he wasn't disclosing those [00:36:00] conversations to the board, that could certainly be. perceived as a pretty significant deviation from their original nonprofit charter and So again, not a hundred percent verified anywhere, but you can kind of connect some dots here that that is likely at least has some, some part in The second is, as we have said many, many times, the AI you see and use today is the least capable we will ever have in human history. OpenAI appears to believe, as do many others, that we have not hit a ceiling. Going back to Ethan Mollick's, you know, possible reasons why Gemini Advanced isn't more powerful than
[00:36:38] Paul Roetzer: 4, we have not hit a seating ceiling for what these models can do. They appear to still be following scaling laws that If you throw more data and more energy capacity and more chips or computing power, maybe with a scientific breakthrough too that we can certainly enter the realm of general intelligence super intelligence.
[00:36:59] Paul Roetzer: That [00:37:00] if we just keep powering through this and grinding and putting more compute power, more energy, that we're going to build much more intelligent aI.
[00:37:08] Paul Roetzer: And then the third is just the importance of understanding the chip industry. Again, if you've invested in or followed NVIDIA for the last, you know, few years, you have certainly seen a, a rapid acceleration in their market cap and value of that stock.
[00:37:25] Paul Roetzer: This is why it is powering the entire, you know, AI aI age that we are experiencing. So I would suggest there to read Chip War by Chris miller. It's a phenomenal book. Once I read it, I had a totally different perspective on the geopolitical climate. As well as the complexity of doing this, like, it's like, well, let's just build some fabrication plants in the U.
[00:37:46] Paul Roetzer: S. How is this so hard? And you realize how much engineering and kind of advanced talent is needed to build and maintain these fabrication facilities. Um, it's [00:38:00] truly some of the most advanced science. that we've achieved in human history. So it's not like you can just pick a plot of land and build one So, yeah, it's you know, it's a really important topic. And I think we'll probably talk more about it moving forward. But the main point here is you need to be paying attention to this. Whether or not OpenAI raises five to seven trillion, who knows? I mean, that's, such an absurd thing to even consider being reality.
[00:38:28] Paul Roetzer: I wouldn't put anything past Sam, but, it's more of the idea that they're thinking this big and they're thinking this long term that I believe worth paying attention to
Some investors are deciding not to invest in OpenAI and other AI startups
[00:38:43] Mike Kaput: All right, we've got a ton of rapid fire topics this week to cover. first up is that some investors appear to be deciding not to invest in OpenAI and some other AI startups, according to reports from The Information. according to the report, quote, for past [00:39:00] year, Owning a stake in OpenAI was akin to possessing Silicon Valley gold, signaling that the investor had ticket to the next great tech transformation.
[00:39:09] Mike Kaput: But as the valuation of the startup has tripled, some investors have refrained from buying more, and a number of marquee venture capital firms have shied away from investing in dozens of artificial intelligence startups out of fears that the sector is too competitive and prices are too high. Now this includes VC firms like Founders Fund, Sequoia, and Kozla Ventures, all very well known VC firms, and some appear to have concerns that OpenAI is not worth investing in at its new 86 billion valuation, which it arrived at late last year when it sold off some employee and investor shares.
[00:39:48] Mike Kaput: Others, however, see OpenAI's lead as so significant that even AI startups twice as efficient aren't enough, according to one investor at Founders Fund, and yet [00:40:00] others worry that the sector is too competitive. With Microsoft, Amazon, and Google, they predict ultimately winning out because they have the money to take on all comers.
[00:40:10] Mike Kaput: Now, Paul, going on here? Is investor demand for AI cooling off?
[00:40:17] Paul Roetzer: I mean, there's certainly plenty of money went into companies 2023 and more is in 2024. one article one angle consider. At a high level, don't disagree on the competitive side. talked about this last episode. think that it's hard for me see Microsoft, Amazon, Google. Losing, like
[00:40:44] Paul Roetzer: they're, there going to a lot of these software companies are going raise hundreds millions, billions of dollars that will not in three five
[00:40:54] Paul Roetzer: there's going to be a bunch of them that just didn't have a strong enough moat, couldn't keep up.
[00:40:58] Paul Roetzer: And once, know, [00:41:00] big models start from these big companies becomes very hard to differentiate. I understand, um, this challenge. I think, I don't know if I've shared this before. Before show, like I at, starting, an AI venture fund, like, know, as recently probably six to eight months ago, were looking at, know, our own studio and raising a significant amount of money and starting invest in and build AI software.
[00:41:26] Paul Roetzer: I stepped back and I don't know what bet like had a really hard time. Envisioning what software companies were going to be defensible.
[00:41:36] Paul Roetzer: And so it wasn't right time me to start in and making big bets on this stuff. Because as I am the space, just who was going to win and who wasn't.
[00:41:50] Paul Roetzer: And I kind of feel the same If I was investing in software companies, I would extremely Uh, hesitant unless felt [00:42:00] very, confident that had distribution model proprietary data that was going to differentiate or
[00:42:06] Paul Roetzer: such a clear advantage in some vertical that going make it. So even the general models became capable doing what they were doing, wouldn't have the audience within vertical. So almost if you a media company within a specific 100, 000 potential and you offer some specific capability to that vertical.So, that was, you know, how I looked at it as an investor and, know, someone who was potentially going to start things.
[00:42:32] Paul Roetzer: I and still feel similar. I a lot these software companies going to be obsoleted at some point. I think a lot the big tech companies, they're, they're going to win one way the other. being said, the timing is fascinating So that information came on February 6th that people like. hesitant to invest in aI
[00:42:54] Paul Roetzer: this hundred billion dollar valuation, whatever it is. And then two days later, wall street has article [00:43:00] saying open AI is going raise like five to 7 trillion as someone who spent some time in the industry, it just of seems little interesting.
[00:43:10] Paul Roetzer: And, basically like, oh, you don't think worth hundred billion we're five to 7. trillion. So get in now or don't, don't get in kind of thing again. don't know, like, usually there's, there's something to the timing these things. So I don't ton of weight necessarily in this.
[00:43:30] Paul Roetzer: I don't think open AI going any problem finding investors their 86 to hundred billion dollar valuation. And
[00:43:37] Paul Roetzer: I would imagine that the vision for what they're trying to next decade with AI infrastructure under the open AI umbrella, um, would enable them to justify whatever valuation they choose to raise money.
[00:43:49] Paul Roetzer: So wouldn't be crying for OpenAI
[00:43:53] Paul Roetzer: if you think they're going to get valuation or enough investors in at that valuation. It's, [00:44:00] don't think are going to have a problem.
A touchdown for AI advertisements
[00:44:02] Mike Kaput: AI was not the star of the big game, but it did play a supporting role in this year's crop of ads during the big game. So we actually saw two AI focused ads from major AI players. that aired during the event. One was for Microsoft CoPilot, which is of course the company's AI assistant that works in Microsoft apps.
[00:44:26] Mike Kaput: The other showed off some AI features in Google's Pixel 8 phone, specifically some accessibility features. A handful of other ads promoted AI features in certain products. Etsy ran an ad about its new Gift Mode feature, which is powered by AI and helps shoppers select gifts. Cybersecurity firm CrowdStrike touted its ability to use AI to protect against cyber attacks.
[00:44:51] Mike Kaput: And one ad, which was a movie trailer for Despicable Me 4, even poked fun at AI, showing the [00:45:00] franchises dim witted minion characters flooding the internet with bad AI generated images, like people with too many arms and fingers and stuff like that. So Paul, during this, did anything jump out to you about, you know, AI's role in this year's ads?
[00:45:15] Paul Roetzer: Yeah, I think we'd kind of joke like the over under on, you know, how many ads would have AI in it. and I think 10, maybe. So I was close. I thought, so Marty Swant, who's from Digiday, he had been tweeting some of this stuff and so he had an article about this. But had, let's see, Microsoft, Google, Etsy.
[00:45:34] Paul Roetzer: So that's three. And then Samsung, CrowdStrike, four, five. you mentioned the Minions one, six, Galaxy. Oh, that's Samsung.
[00:45:45] Paul Roetzer: Oh, then the one that you and I couldn't find, like, yeah, so I was, I was watching the game, but I was working and then I was getting my
[00:45:51] Paul Roetzer: kids ready for bed. So I missed some of these, but I saw a bunch of like AI people like, Oh, the anthropic ad.
[00:45:57] Paul Roetzer: I can't believe they spent money on this. It was like, there's an [00:46:00] anthropic ad. And I've searched for it. I
[00:46:01] Paul Roetzer: cannot find the video of the ad. So I don't know, maybe saw it.
[00:46:06] Paul Roetzer: I did not. Okay. So apparently there was philanthropic. Yeah, but
[00:46:10] Paul Roetzer: it was like five seconds long. So if it, the 30 second ads were 7 million, I believe this year. So I don't know what Anthropic spent, but I don't know. I guess they.
[00:46:20] Paul Roetzer: I'll be interested to see what the angle So, I like the Microsoft one. I thought it was good. I thought it was interesting. They didn't go after the enterprise audience. They were going after like the individual entrepreneur creator.
[00:46:30]
[00:46:30] Paul Roetzer: so yeah, I did, you know, I guess none of them, I saw the CrowdStrike one. I didn't even make the connection. Like, I was really confused by that one. It was like robots wererolling in in the wild west. Yeah. CrowdStrike lady went out and destroyed them by taking their robot clothes off or something. And then they, I don't know.
[00:46:49] Mike Kaput: there was
[00:46:49] Mike Kaput: a, yeah, that ad was pretty stylized, but yeah, reading further about it seems like it was in fact promoting some of their AI.
[00:46:58] Mike Kaput: focused cyber security [00:47:00] features. But yeah, good luck knowing that exactly from the ad.
[00:47:03] Paul Roetzer: So, yeah, I mean, we definitely expected, you know, the AI to to play a role and it, it was it sounds like maybe like six or seven ads, but go check out Marty's story. He's got some more details on the different ads and the strategies behind them. So we'll, we'll again, we'll put the link to, Marty's, in the show notes.
Finance worker duped by a deep fake call into $25M payout
[00:47:21] Mike Kaput: Alright, so this next story is pretty wild. A finance worker at a multinational firm was tricked into paying out 25 million to fraudsters who were using deepfake technology to pose as the company's chief financial officer in a video conference call.
[00:47:40] Mike Kaput: And so this, is according to Hong Kong police, which is where this took place. So the scam actually saw this worker being duped into attending a video call with what he thought were several other members of his staff, but all of whom were in fact deepfake [00:48:00] recreations. Now, this is not the first deepfake scam we've seen.
[00:48:03] Mike Kaput: It's not going to be the last by any stretch, but it is kind of the latest, most extreme reminder that we have, that we now have deepfake video and audio that can essentially create indistinguishable fakes of real people. So, Paul, we talk pretty frequently about the fact that companies need a deepfake strategy. But, like, what kind of steps do we need to start taking here prevent stuff like this from happening?
[00:48:32] Paul Roetzer: Every time we talk about this, I just feel worse about where this is all going. I, so we're going to dig in next week to a main topic around some initiatives with major tech firms in the industry that are sort of coming together around some standards to
[00:48:52] Paul Roetzer: authenticate content and, and reduce disinformation and misinformation. And maybe we'll save that bigger [00:49:00] conversation for that time because the big tech companies are absolutely aware
[00:49:06] Paul Roetzer: this.
[00:49:06] Paul Roetzer: The governments are aware of this. People know that there are major problems ahead and I think we need to just take a look and maybe kind of synthesize for people some of the initiatives that are going on to give little hope because otherwise it's just me saying we're all in trouble here and it's not going to go great.
[00:49:27] Paul Roetzer: But that doesn't tell the whole picture of what is, what is happening, so I don't know, like, this is crazy, this story was nuts. I think we're going to see lots of examples of this kind of thing. If something this advanced already happened, like, it just sort of starts to make you worry about what else is to be going on.
[00:49:46] Paul Roetzer: But, yeah, let's, Let's definitely talk more about what's going on in a positive sense of maybe some of the progress being made on these issues because I definitely think we can all agree these are major [00:50:00] issues that are going affect a lot of different parts business and society and we got to
[00:50:04] Paul Roetzer: start kind of digging into what is going on and what else can the rest of us maybe do to help alleviate some of the pains that we're going to feel on this
AI disruption to consulting firms like McKinsey
[00:50:12] Mike Kaput: Alright, so in our last topic for this week, we saw an interesting experiment with ChatGPT that actually raised some bigger questions about the nature of AI's impact on employment at consulting firms, like especially big ones like McKinsey. So this all started when a popular account on X named CarDealershipGuy, who has 44, 444, 000 followers, posted to his audience, which is car dealers asking how their month was going in terms of their business.
[00:50:43] Mike Kaput: His post got, like, almost 200 responses from owners of car dealerships. Now, here's where it gets interesting. One of the guy's followers used ChatGPT on those 200 or so responses to turn them into incredibly detailed [00:51:00] market research on the current pulse of the car market. So this guy, car dealership guy, called this research Truly incredible.
[00:51:07] Mike Kaput: And here's where kind of this bigger picture idea comes in because someone else, an ex user by the name of Liz Hoffman, posted in the reply, quote, this is literally thousands of people's jobs at McKinsey, etc. Indicating that ChatGPT combined with easily obtained public social media data, was able to perform tasks typically reserved for elite, highly paid consultants.
[00:51:31] Mike Kaput: As Car Dealership Guy put it, quote, we're living in the future. Now, Paul, this may just be one scenario online, but it did catch our attention because if ChatGPT and other AI tools can do stuff like this on the fly, Does this devalue what we traditionally hire elite consultants for?
[00:51:49] Paul Roetzer: Oh, it devalues it, but it certainly accelerates the creation of and liz is no like average Twitter user. I mean, she's, writes for [00:52:00] Semaphore and a Wall Street Journal alum and author of a book called Crash Landing. So when Liz tweeted, that's what caught my attention is her saying, this is literally thousands of people's jobs
[00:52:10] Paul Roetzer: McKinsey. And I think she was like. Kind of half joking, but probably not. And I do think that as different industries start to look at the stuff that is data driven, repetitive, making predictions about outcomes and generative, those are kind of four categories of identifying a use case. When you start to break that out and you look at the things that we spend bunch of time doing and people pay consultants a bunch of money to do, you start to realize that.
[00:52:41] Paul Roetzer: Wow, like GPT 4 or Gemini Advanced, like might actually be not only will it do it in 30 seconds, it might be better than the 700 an hour I'm paying my consultant for. That's a very disruptive future. like,
[00:52:58] Paul Roetzer: And this is why we've [00:53:00] talked about that. I don't think that the economy, and I certainly don't think most industries and business leaders are prepared for how disruptive this is going to be to knowledge work in the next like 12 to 18 months across every industry.
[00:53:14] Paul Roetzer: And so I think it's just a really practical example of how, when you know what this stuff is capable of, You look at it differently. It's like car dealership guys collecting some data and that's interesting to him and probably spend hours going through it. And someone else who knows what ChatGPT is capable of is like, let me just scrape all that and I'll create this cool little summary of it in 30 seconds.
[00:53:33] Mike Kaput: Okay.
[00:53:34] Paul Roetzer: The people who, we've said this before, like the people who understand AI and can apply it, have tremendous growth opportunities So I'm a big believer that there will be disruption to jobs. The best thing you can do to position yourself to thrive in that disruption is to understand And [00:54:00] so it's not a promise that your job won't be impacted or that you won't be impacted by AI, but if you have capabilities, other people don't have, and if you have an understanding of this technology in a way, your peers
[00:54:13] Paul Roetzer: don't, you have a far greater chance of not only surviving through this disruption, but thriving in the And I think that's over and over again, one of the key messages I hope people take away from our. podcast, our newsletter, our courses, our talks that you and I give, like just take action and figure this stuff out and you're going to give yourself the best chance to succeed moving forward.
[00:54:37] Mike Kaput: Those are great words to end this week's podcast on Paul.
[00:54:42] Mike Kaput: Thank you so much for breaking this all down for our audience. I just want to, give a very quick reminder that there are tons of items that we just don't have time for in each week's podcast episode. You can find those in our This Week in AI newsletter, which you can find at marketingainstitute. com [00:55:00] forward slash newsletter.
[00:55:01] Mike Kaput: It breaks down more in depth all the topics we discussed today, plus all the ones we didn't get to. It's a really great digest for your weekly update on what's going on in AI. Paul, thanks again.
[00:55:14] Paul Roetzer: Thank you, Mike. I will talk with everyone again next week.
[00:55:18] Thanks for listening to The AI Show. Visit MarketingAIInstitute. com to continue your AI learning journey. And join more than 60, 000 professionals and business leaders who have subscribed to the weekly newsletter, downloaded the AI blueprints, attended virtual and in person events, taken our online AI courses, and engaged in the Slack community.
[00:55:41] Until next time, stay curious and explore AI.
Claire Prudhomme
Claire Prudhomme is the Marketing Manager of Media and Content at the Marketing AI Institute. With a background in content marketing, video production and a deep interest in AI public policy, Claire brings a broad skill set to her role. Claire combines her skills, passion for storytelling, and dedication to lifelong learning to drive the Marketing AI Institute's mission forward.