Transcription
So, this is a talk I've been wanting to do for a couple years. This, it's sort of based on this documentary, um, called AlphaGo. So, if you haven't watched the documentary, I highly recommend it. I'm going to give you a few excerpts from it as we go through today. But, this documentary sort of changed my life in a, um, somewhat profound way. And so I want to, I want to share that experience with you today and kind of like walk you through the moment we find ourselves in by looking back nine years ago to what happened in South Korea.
So in 2016, Google's DeepMind AlphaGo system took on 18-time World Go champion Lee Sedol. Now, if you're not familiar with the game of Go, it's been played for thousands of years, originating in China. It is infinitely more complex than chess. So, in what became known as move 37, everything fundamentally changed about technology, about the progress of AI, and in my view, about the future of humanity.
So, beating a professional Go player had been considered a grand challenge in AI for decades, and most leaders thought we were at least a decade away from it happening. Move 37 is often talked about, and you may have heard about it as a technological breakthrough. I actually had people, when they saw this session title, say, you know, they talked to me about the technology piece. What I want you to do is to focus on the human element of move 37. And so again, what I'm going to do is take you through moments from this documentary and provide context to you as to what this means to all of us as marketers, as business leaders.
So, in this scene from AlphaGo, the documentary, Sedol has already lost the first of five games. During game two, he starts to battle his emotions, coming to grips with the fact that AlphaGo appears to be superhuman at the game of Go. So, this was an exhibition that Sedol had agreed to because he, quote, "thought it would be fun and thought he would win five to zero." So, when Sedol would get anxious, he would actually leave the room. They had a special area created outside for him to go have a cigarette. So, that's what he would do when he just, you know, felt the nerves. While he was away, though, AlphaGo would just continue to play. So, there's a human player, Aja, who is one of the Google DeepMind developers, who sits across from Sedol, but Aja isn't doing anything other than putting the chips where AlphaGo tells him to. So, he interfaces with the computer monitor. It tells him to put a chip somewhere. Aja places the chip, and then he just sits there.
I want to see Lee Sedol when he sees this move.
[Music]
He's back. Lee's back.
[Music]
Fore!
[Music]
[Music]
Normally, I have to think about one, two minutes, no more. But this time, I think more than 12 minutes.
[Music]
The more I see this move, I feel something changed. Maybe for humans, we think it's bad, but for AlphaGo, why not?
So, for me, the move 37 moment, the whole premise of this talk, is it's the moment when you realize AI is better than you at the thing that you do. So, for many of us, that moment is coming sooner than we think or may want to believe. It may start as individual tasks. It's, it's better than you at little things, but those tasks will add up, and it will start to affect entire jobs.
In January of 2025, Noam Brown, who you've heard of if you listen to the podcast, we talk about Noam quite a bit. Uh, he came from the Meta team and now he's on the OpenAI team. He said, "It can be hard to feel the AGI, meaning artificial general intelligence, which I'll add some context to in a few moments if you're not familiar with the concept, until you see an AI surpass top humans in a domain you care deeply about. Competitive coders will feel it within a couple years. Paul Schrader, who I'll explain in a moment, is early, but I think writers will feel it too. Everyone will have their Lee Sedol moment at a different time."
So, Noam was actually responding to and retweeting a post from Paul Schrader that he had put on Facebook. Paul Schrader is an American screenwriter, film director, and film critic. First became known for writing the screenplay to Scorsese's Taxi Driver, and then later continued his collaboration with Scorsese, writing, uh, Raging Bull.
So, Schrader on his personal Facebook page, which is public, you can go look at these posts if you want. He says, "This is January 16th. I've come to realize AI is smarter than I am, has better ideas, has more efficient ways to execute them. This is an existential moment akin to what Kasparov felt in '97 when he realized Deep Blue was going to beat him at chess."
So, someone replied and said, "What brought you to this conclusion?" He said, "I asked it for Paul Schrader's script ideas. It had better ones than mine." Later that day, he posted, "I just sent ChatGPT a script I'd written some years ago and asked for improvements. In 5 seconds, it responded with notes as good or better than I've ever received from a film executive." The next day, "I'm stunned. I just asked ChatGPT for an idea for Paul Schrader film. Then Paul Thomas Anderson. Then Quentin Tarantino, Harmony Korine, Ingmar Bergman, Rossellini, Lang, Scorsese, Mno, Capra, Ford, Spielberg, Lynch. Every idea Chat GPT came up with in a few seconds was good and original and fleshed out. Why should writers sit around for months searching for a good idea when AI can provide one in seconds?"
This is David Prell. This is also from earlier this year. He is a writer and writing coach. A writer? Yeah, writer and writing coach. 450,000 followers on X, uh, or Twitter, if you're not familiar. So, this is February. So, "This AI boom has set off an existential crisis for me. I've decided to stop teaching. It has only been four months since I shut down my business, but I can no longer imagine teaching writing in a way that resembles anything close to the way I taught in the past. The reason is simple. The world of non-fiction writing has fundamentally changed, and many of the skills I've developed and built in my career are becoming increasingly irrelevant. What if the work that defines you and brings you fulfillment changes? What if AI moves further into the strategic and creative realms much sooner than expected?" I wrote that in May of 2022, right after Dolly 2 came out, six months before ChatGPT emerged. I keep coming back to these same questions, though. They seem to be taking on greater meaning and importance as AI continues to advance.
For Sedol, the full weight of his move 37 moment hit the next day at the end of the third game, which he lost as well. During the press conference following that third game, Sedol shared his profound sadness and disappointment as he sat next to Demis Hassabis, co-founder and CEO of Google DeepMind, and the brains behind the building of AlphaGo.
Fore!
Foreign!
Foreign!
You have to keep in mind, Go is the national game in South Korea. 200 million people were watching this live, and he felt he had let humanity down.
So, in that same post in May of 2022, I wrote, "Something is lost, and something is gained." And so, when we have this moment, it presents opportunities as well. So, with two games to go in the match, Sedol regrouped. He consulted with other Go experts to try and figure out how is Go, AlphaGo, doing this, and he returned for game four. Unfortunately, game four started out much like the three before it, and AlphaGo appeared to be in the lead. Then, Sedol made an unexpected play at move 78 that seemed to confuse AlphaGo and its creators. They raced to the back room, and they're trying to figure out what happened to AlphaGo. Its play started becoming erratic. The AlphaGo team would determine, after the fact, that Sedol had exposed a weakness in the system. He had made a move that AlphaGo, the system, defined a 1 in 10,000 chance of a human player making. The same probability, by the way, that was assigned to move 37. It's why Sedol was so thrown off. The chance of a human player, a human expert, making move 37 was 1 in 10,000. Meaning, Sedol had never seen a human player make a move like that. And yet, the AlphaGo team assigned that same probability. When they asked Sedol afterwards, "How did you make move 78?" He said, "It was the only move on the board I saw." So, literally, AlphaGo looked at 10,000 other possible moves and assigned higher probabilities than the one Sedol makes. The Google DeepMind team called it a "god move," and it led to the unraveling of AlphaGo, at least for one game.
[Music]
Alpha resigned.
Looks like AlphaGo has resigned.
Wow. The most amazing game. Um, I'm almost going to tear up. Was, was, uh, was game four, um, where he comes back and wins. Right.
[Music]
[Applause]
[Music]
Babychech.
[Music]
[Music]
So, why now? Why, why is it important that we're here together talking about move 37, nine years after it happened? Why do we need to prepare for our own move 37 moments? Knowledge workers, all of us, think and create for a living. That includes marketers, educators, podcasters, designers, business leaders, everyone in this room. AI is rapidly advancing its capabilities and moving into our domains and areas of expertise. It's being driven by AI progress that is out of our control. We do not get to decide what the labs build.
So, AI is changing everything. I talked about it recently, as, uh, the underlying operating system to society and business. Literally, every piece of software, everything we do in society, is going to be built on top of intelligence that none of us control the building of. It is on a near-term path to surpass the average human at most cognitive tasks. This is what we would consider general intelligence. Many researchers believe it is on a very near-term path to become expert level at all cognitive tasks. So, take the best in your industry, the best in your company, and it's at that level. This is what we would consider super intelligence.
So, when we think to the future, and anyone who's in my innovation workshop yesterday, we talked about this idea. The challenge for us humans is we can only think in linear paths. We cannot look to tomorrow and imagine what 10x or 100x better, smarter looks and feels like. But that's what we're in. We're in this exponential growth curve where the technology is getting 10 times, 100 times smarter, and 10 to 100 times cheaper every year. The reason this is happening is something called scaling laws. So, if you've ever wondered why Nvidia is worth over $4 trillion, this is why. Years ago, in the early days, prior to the formation of OpenAI in 2015, 2016, they found the first scaling law that was, if we give more compute, if we put more NVIDIA chips in a data center, we give it more data, and we train it, it just gets smarter. And so, we buy more NVIDIA chips, we build more data centers, we create more energy. It's why there's such a push right now to do both of those things, more data centers, there's more energy, because the first scaling law is still holding true. We know we build bigger models, they get smarter.
Somewhere along the way, the first real exposure for most of us was last fall. They found a second scaling law. This is the post-training era. This is, you take a trained model and you teach it specific things. Legal industry, accounting industry, marketing industry. You do reinforcement learning to teach it how to do very specific things, and they get smarter. The third that we have now found is test time. This is at what's called inference. This is when you and I use ChatGPT or Google Gemini or any of the other hundreds of AI systems that are out there. What they have found is if you give them time to think, so rather than an instant information retrieval answer, it takes its time. It goes through a chain of thought. It gets smarter too. So, we now have three scaling laws that are compounding together to accelerate the development of this technology.
And so, in each of the AI labs, we have a session closing today with people from Meta and Google DeepMind that are going to talk about sort of what's going on inside of the labs. What's basically happening is, imagine that the end product is something that comes out of the oven. We're baking something, and it's pictured as a model. The ingredients you put into that model. There's a bunch of different ways you can do it, a bunch of different resources you can apply to things like agentic and memory and reasoning. Each lab is experimenting with the ingredients. They're trying to figure out how to build the smarter model. They all generally know the potential ways to do it, but each lab is making slightly different bets as to what it's going to be, but the things on this slide are all components of them.
And so, what does this mean from a tangible perspective about where the labs think we're going? This is from last summer. This is the OpenAI stages of AI and the different levels. Level one is what we got in November 2022. That was chatbots. You put a text prompt in, you get text out. Level two is what we got in September of 2024. That was reasoning models that could go through that chain of thought. Level three, we're sort of on the cusp of. We have these agents that still need a lot of hand-holding by humans, still need a lot of evolvement. They're not self-driving, you know, as, as you would think like a car driving itself, but they're getting there really fast. And that leads to level four, which is innovation, which means coming up with ideas humans don't think of, don't have the time to compute ourselves. We are seeing glimpses of this now, specifically in math, coding, and science. That will probably progress pretty quickly in 2026 into every one of your industries. And then level five follows shortly thereafter, where you basically build entire organizations of these AIs.
So, what's happening now is, and again, if you listen to the podcast, you've heard me kind of like get on my pulpit about this, evaluations of these models up until about three to six months ago were IQ tests. So, they basically would give them all the, like, imagine like the bar exam or AP Chemistry, whatever it is, they would give these models these tests, and it's like, "Okay, we get it. They're smarter in basically every discipline than any human." The, the challenge I always saw was, "Talk to me about the impact on the economy, though. Like, where, what is it going to start to change jobs?" And so, we've now seen this. So, we have GDP Val from OpenAI, which just came out last month. It's Gross Domestic Product. They're, they're basically looking at the impact of AI on the production of goods in the United States, in particular, but you can do this worldwide. So, in their case, they looked at 44 occupations and they said, "Today's best frontier models are already approaching the quality of work produced by industry experts. Frontier models can complete GDP Val tasks roughly 100 times faster and 100 times cheaper than experts."
Another thing you look at is the ability for these things to do what we would call long horizon tasks. So, not like send the quick email which takes 10 seconds, but actually like conduct a research project, build a plan for the research project, conduct the research, edit the research, check the citations. That's more of like a long horizon. Something that might take a human 1 hour, 3 hours, 5 hours, 50 hours. What they're now trying to look for is how reliably can these systems do these long horizon tasks. So, there's a potential new scaling law emerging here. This is, it's still early. We can't like call it a law. But what this organization, Meter, has found, which does model evaluation and threat risk analysis, is that they seem to be doubling every seven months in the amount of time they can perform these tasks. So, this data is specific to coding, but it says AI models today, and this is actually from March. So, assume this is probably two hours now. AI models today have a 50% chance of successfully completing a task that would take an expert human, not an average human, an expert human, um, 1 hour. Seven months ago, it was 30 minutes. Seven months before that, 15. So, again, you can do the math. By the time we're here together next year, we're basically at eight hours on, on this side.
They're not the only ones looking at this, though. So, Anthropic, some of you may be fans of like the Claude model. They just released their third economic index report where they're trying to look at geographic distribution and usage of these tools. So, the report says, "Existing capabilities of Claude and other frontier systems are already poised to transform economic activity given how broadly applicable the technology is. If more advanced models simply expand the set of automated tasks, then the risks increase that workers performing such tasks will be displaced." What the chart is showing is intent of the prompt. And what they're seeing is a dramatic shift where the prompts people put into Claude are intended to perform tasks. It's not information retrieval now, it's "do this thing for me now." And so, you see it was the first time where it crossed over. And you can see the trend line to the point where you can imagine 70% basically by this time next year is intended for it to do the thing for you.
So, if you've heard me give talks, if you've taken any of our courses, you've probably seen this chart. This is the AI timeline that I presented on the podcast first last year and updated earlier this year. This is generally where I think we are. So, for the last couple years, we have had these language models continually, continually advance, have these new capabilities added. That is continuing. Multimodal means the models can now not only do text, they can do image, video, audio, code. Agents are becoming more reliable. They are not autonomous yet, for the most part. So, there's a lot of hype around these. Assume they're probably not as advanced as you hear, but also assume they're probably going to get there faster than you expect them to. And then that leads to robotics, which I'm not going to get into today, but humanoid robots is a thing. The, the sci-fi movies are going to be coming to life by the end of this decade. And then AGI from there.
So, this is the definition of AGI. I mentioned AGI earlier. This is my definition. There are many definitions. This does not mean it is the right one. This is after 13 years, 14 years, however long I was studying this space. This is the best I've got at the moment. There's a very important distinguishing factor between my definition and the definition from the AI lab leaders, and that is, it just has to outperform the average human. Let's be honest, a lot of your co-workers are average at their jobs, right? All that matters, in my opinion, is, can it do their job? It does not have to be PhD level at everything. It does not have to be the top 1% marketer. Economically speaking, the question is, can we replace people with it? Is it better than the average or below average people?
And so, there's this new economic Turing test. You probably heard of the Turing test back in like the 1950s, Alan Turing, like, would you know when you're talking to an AI or not? Roughly is, is the concept. We've passed that. Like, that was passed probably a couple years ago. So, there's this new concept of an economic Turing test, which is, would I hire an agent or a collection of agents working together instead of a human, and not just to do a task, but to actually perform a job? And so, this is why I try not to get too caught up in this whole AGI thing, because one, the labs keep changing what they think it is. They just keep moving the goalposts as to like, whether we're there or not. And my argument is, it just doesn't matter. What matters is drilling into your job, your team, your company, looking at the reality of where the technology is and saying, "Is it going to change our work?" That's really the only thing that matters.
So, the other reason why we have to think about this now is because we live in a capitalistic society. People will pursue money. If you look at the software industry, it is roughly estimated to do $300 to $500 billion a year in sales. So, if you are an AI lab, if you're an AI startup company, and you're trying to figure out, "What should we build that we can make some money using this AI technology on?" One place you look is the software industry. Salesforce, who has their Dreamforce conference this week, I believe, $38 billion a year. Adobe, $32 billion. ServiceNow, $11 billion. Workday, $8 billion. HubSpot, $2.5 billion. So, one chance is, well, let's just build an AI-native version of those companies and let's go get that market share. That's a big market. Not nearly as big as the estimated 100 million knowledge workers in the United States.
So, there's give or take about 136 million full-time employees in the United States. Estimates vary somewhere between 70 and 100 million of them think and create for a living. They're, they're us. So, what I did is I went to the US Bureau of Labor Statistics and I pulled the most recent data I could get from May of this year on total wages by profession. So, there's about $11 trillion in wages per year. So, if you wanted to build an AI-native company, you could go after the existing software industry. Or you could say, "Hold on a second. There's 13.3 million sales and related occupation jobs with total wages annually of $722 billion, double the smallest estimate of the software market, just for that one." You could go to customer service representatives, 2.7 million employees with $123 billion. You could do market research and, and marketing specialists, 860,000, $74 billion. Median communications workers. Now, mind you, you can export this database yourself, and there's 800 or so occupations, I think it was.
Now, you could sit there and think, "Yeah, nobody's going to do that, though." Oh, really? Why Combinator, which is where Sam Altman was the president of before he took over OpenAI, had a call for submissions. They're literally asking people to pitch them to automate the workforce. So, this is from their call for vertical AI agents from earlier this year. "The value prop of B2B software was to make human workers incrementally more efficient. The value prop of vertical AI agents is to automate the work entirely." And their premise, no joke, go listen to the podcast I think that dropped yesterday. We talked about Mechanize. They, this startup company, their belief is, "Listen, the workforce is going to get automated anyway. We might as well just do it. Like, we get it that it's bad for humanity and society, but somebody's going to do it. So, what the hell?" That's the mentality. They see the opportunity to go raise $50 million, $100 million, whatever it is, and take on. All you got to do is show up with a pitch deck and say, "We're going to take this part of society. It's got $400 billion a year in total wages. We think we can automate 20% of that within two years." Done. Here's your $50 million. Go do it. That's how Silicon Valley works.
Jeremiah Ayang is going to be here. He's going to give you some of the, there's Jeremiah over there. He's going to talk to you about what's happening. He's doing this. He's seeing these events. He's meeting these people. This is what's happening in Silicon Valley. And the demand is skyrocketing.
So, one way you measure usage of AI is number of tokens. So, think of a token as like a piece of a word. So, every time you go use ChatGPT, the way it's metered by the labs is how many tokens are generated? How many tokens you put in, how many are generated in the prompt response. But you know what takes more tokens than words? Videos. Sora 2 came out a week or two ago. You know what else takes a whole bunch of tokens? Reasoning models and agents. And people tell me Nvidia is hit a wall. Like, the, the economy has no concept of how significant the demand for intelligence will be. Sundar Pichai last week at Google, the co-Google, said they had processed over 1.3 quadrillion. Hard number to comprehend. So, Demis Hassabis broke it down a little bit for us. 500 million tokens a second, or 1.8 trillion tokens an hour they're generating. People are demanding of their Gemini models. That's up 30% since July.
So, we are entering an era of what I call omni-intelligence. So, omni-intelligence describes both the state of AI being omnipresent in our lives. It's literally everywhere. Every piece of software we touch, everything we do. But it also describes the models. When the O series came out from OpenAI last year, they called it an omni-model, meaning that it can understand, reason, and take action across modalities: text, image, video, audio, code. So, omni-intelligence is the idea that it's always with us. It is truly the operating system of society, and it's able to do all the things that we do.
Now, this creates unparalleled opportunities to grow and to innovate and to do creativity. But market pressure leads us to the short-term gains. So, companies that are publicly traded, venture-backed, private equity-owned, they're not thinking about innovation and growth. They're thinking about replacement of workforces and cost reduction and hitting their earnings numbers next quarter. There's tremendous pressure to do these things.
So, this is IBM Chief Executive Arvind Krishna. Said the company used AI to replace the work of a couple hundred people in human resources. No big deal, right? They have like 80,000 employees. Um, Andy Jassy, this is a memo we talked quite a bit about earlier this year on the podcast. He sent an internal memo to their employees. I'll just zoom to the end. "As we roll out more generative AI and agents, it should change the way our work is done. We will need fewer people doing some of the jobs that are being done today and more people doing other types of jobs. It's hard to know exactly where this nets out over time, but the next few years, we expect this will reduce our total corporate workforce as we get efficiency gains from using AI extensively." There's a whole bunch of buzzwords in there that probably equals a one and a half point bump in their stock that day. Wall Street loves to see these kinds of things from CEOs.
Chief Commercial Officer Judson Althoff at Microsoft said AI tools are boosting productivity in everything from sales and customer service to software engineering. He said that AI saved Microsoft more than $500 million last year in its call centers alone and increased both employee and customer satisfaction. Mark Benioff, who is very bullish on this stuff. "AI is doing 30 to 50% of the work at Salesforce now." Now, when you dug into this quote, he contextually was speaking more about like code, but basically the play, the play Salesforce is making right now is redistributing talent. They're basically stopping hiring and they're trying to take their existing staff and just move them over here. So, it won't show up in the unemployment numbers. It won't show up in the reduction of workforce because they're just moving numbers side by side.
Here's Jensen Huang, one of the richest people in the world and runs the most valuable company in the world. There are 32,000 employees. He said, "We could get to 50,000, but every part of our company's going to have a 100 million AI assistants helping them do their jobs." So, instead of hiring a half a million or half a billion people, we're just going to have whatever, 50,000, and we'll just do it this way.
This is Walmart. September 2025. Doug McMillon said that he expected headcount to stay flat over the next three years despite growth plans. Now, we've been hearing, and I've shown you a bunch of tech CEOs and tech leaders. This is the largest private employer in the United States, who employs 1.6, 6 million people in the United States, telling you they have no intention of hiring more people.
This is Jim Farley, the CEO of Ford. "AI is going to replace literally half of all white-collar workers in the US. AI will leave a lot of white-collar people behind." This is Janet Truncale, the global chair and CEO of Ernst & Young. "I like to think we can double in size with the workforce we have today." And then Vista Equity Partners CEO Robert Smith at a conference in June of this year. "60% of the 5,500 attendees at that conference would be out of work next year. We think that next year, 40% of the people at the conference will have an AI agent, and the rest of you will be looking for jobs." How'd you like to be sitting in that audience?
Okay. So, to be honest with you, the hardest part of preparing this talk, I, I knew the general format and where I wanted to go. It was, how do I give hope? This sucks. Like, this is not a fun talk to stand up here and give you all on day one where you're all like, "I shouldn't have come. Like, I'm leaving. Like, this is terrible. I'm gonna call my kids." Um, I promise we're going to get there.
So, here's my hypothesis. In the next one to two years, model advancements, agent capabilities will force a radical transformation regardless of company. Some of your companies are going to move slower. Some of the industries will move slower. But generally speaking, the market opportunity, the pressure to do something, and the inevitability of the technology capabilities is going to force this to happen. Some companies will choose to do this in a responsible, human-centered way. Others will cut workers. The nature of work will change. Like, this is indisputable. Every business has the opportunity to do things in a disruptive way, to do it themselves. I shared the example yesterday in the innovation workshop of Google, like this company that out of nowhere, no one thought anyone could take on Google, and all of a sudden, OpenAI and Microsoft. What did Satya, Chris, making them dance? I think the quote Satya said about Sundar, like, they thought it was fun that they were making Google sort of change their business model. If someone can come at Google in the core business that they have built the most profitable engine in history, they can come at you. So, you have to do this yourself. You have to go after the changes.
I think what's going to happen once HR catches on. So, if you all have any influence on hiring in your companies, reskilling, upskilling, I get asked all the time, "What matters? What should I be focusing on? What should my kids focus on if they're in college or high school?" This is the general direction of skills and traits I think matter. Like, I've told a lot of people, "Just go get a liberal arts degree, take business classes, and focus on these skills and traits." So, you have to have AI literacy. You have to have interpersonal communications, critical thinking, curiosity, EQ, imagination, adaptability. You have to know what questions to ask and what to do with the answers you get from the AI. You have to know how to talk to, collaborate with, and learn from the AI.
When I was building the Bureau of Labor Statistics chart, I actually discovered that Gemini, now embedded into Sheets, is actually functional. It wasn't. I don't know when this, like, happened that it, it was, but I went in, I was like, "Find all occupations related to marketing." It did. Then the best part, I said, "Can you add a column that calculates total annual wages?" It did. It actually did something in it. It added a column and did a formula and all these things. And I was like, "That was incredible."
Jobs are going to change. Some jobs will go away. New career paths will emerge. There hopefully will be an explosion of entrepreneurship. So, what can we all do as we're here together? I've said this on the podcast a million times. There's more questions than answers right now. That means there's an opportunity for the people who move through the fear and anxiety to go do something because we know the labs are going to continue to progress towards AGI and beyond. And you can't wait around for someone to solve this for you.
So, you are now together with a group of people who are all in that same spot, trying to figure this out. We have to think deeply about how this is going to impact us, how it's going to change things. We have to become AI-forward. I've said this many times on the podcast, this idea of a human-centered approach to this stuff. We're all dealing with these same unknowns. AI-forward embraces that. Says, "You know what? It's okay. I get it. We don't know what's going to happen next. There is some fear, the anxiety, but everyone's feeling it. Let me be the one that figures out how to do this responsibly. I'm going to bring other people along with me."
As an organization, you think about it under everything: people, process, technology, strategy, budgets, all of it. It has to be reimagined as an organization. What we do, we have to put growth and innovation at the center of that. So, AI-forward becomes both a mindset and a career and business imperative. It'll come as no surprise to anyone who follows us, literacy. But it's not just me saying this. LinkedIn did this in spring of this year. What are the top skills? First time they published a report, number one on their list: AI literacy. So, now it's being expected. I think I saw a stat that was like, the increase in AI literacy in job requirements was up like 400% this year. So, professionals who understand this, who embrace this, you will have superpowers. Many of you probably already feel this, that you have superpowers over the people in your company who aren't figuring this stuff out. You will be able to outperform everyone at a rate that is hard to comprehend: efficiency, productivity, creativity. You will be more creative, more innovative, and thereby, you can actually make a difference and help drive growth in the company. You will have the highest value. You will have the highest earning power.
We at Smarter X are growing. We've doubled the staff in like four months. We will probably double it again in the next four to six months. And you know what? We probably function as a team that would be three to four times bigger, which means we can pay people more because they're worth more and they're doing more than other people. And that's the vision I want from leaders. Pay these people. Give them the opportunity to do these things. Don't make them work 60-hour weeks. You know, they can do it in 35 and create three times the value they did last year. Give them time back for their families and friends. Like, that's what's possible.
So, we have to commit to innovation and growth. If Walmart's staying flat, if EY is staying flat, if everyone is staying flat, the only companies that can create jobs are the ones that grow. So, you have to get beyond. We have to think about ideas that make a big impact on our companies, that accelerate change. Optimization does better, faster, cheaper of existing things. Innovation creates new forms of value, doing new things, going after new markets, new products, new processes. Optimization is 10% thinking. Innovation is 10x thinking. And that's the mindset we all need to be in to make a difference. Innovation makes possible what was previously unthinkable. You have to change your mindset. You have to explore the frontiers of where AI is going over the next two days. Hopefully, that's what we give you is a much greater understanding of where it's going in the next 6 to 12 months so you can get there first. Ask yourself questions like, "What can I change? What can I reinvent? What can I transform in my role, my team, my department?" Whatever your influence is, do something. Challenge yourself to do something greater than you think is possible.
And the final thought, embrace your move 37 moment or moments. We can't think of this as human versus AI. It has to be human plus AI. And so, when your move 37 moment arrives, you have a choice. You can give in to the anxiety and fear, or you can choose to learn and grow and evolve. So, AI can unlock human creativity and potential, not replace it, if that's what we give it the opportunity to do.
There are so many possible application domains where creativity in a different dimension to what humans could do could be immensely valuable to us. And I just love to have more of those moments where we look back and say, "Yeah, that was just like move 37. Something beautiful occurred there."
At least in a broad sense, move 37 begat move 78, begat, um, a new attitude at Leto, a new way of seeing the game. He improved through this machine. His humanness was expanded after playing this inanimate creation. And the hope is that that machine, and in particular the technologies behind it, could have the same effect, uh, with all of us. So, for us, we're reimagining what a research firm could look like, doing research in days with deep research technology that would take traditional firms months. I created an AI teaching assistant earlier this year that helped me produce 20 new courses and certifications for Academy that saved me hundreds of hours. I have a co-CEO that helps me run the company that's built simply as a ChatGPT. The template for this, by the way, is free on our site, ungated. You can go grab it. Claire Pomeme on our team used AI and her own creativity and imagination, plus video generation and audio generation capabilities, to create an incredible film. You will all see tomorrow morning, related to the move 37 moment. My 13-year-old daughter, who despised AI in 2022 when it first came out, she now uses it to visualize characters for her stories. She's, as an aspiring storyteller and artist, she's using it to visualize what she has in her mind. PJ Ace is doing it to disrupt the ad industry. You'll hear from him tomorrow morning. He has a talk on the rise of AI filmmaker where he's going to teach you the exact formula he used to create hundreds of millions of views in ads, one for Kelsey that ran during the NBA finals. You have to have conviction about the future. You have to be willing to be bold and make difficult career choices. You have to be willing to be different. You have to be willing to take action. And you have to be, maybe most importantly, willing to lead and inspire others. There are a lot of people afraid. We have to bring them along because together, we can make the future, not only of marketing, but of society, both more intelligent and more human. Thank you.