Event Replay: Jobs in the Intelligence Age

I'm Mark Murray, the Editorial Director at OpenAI, filling in for my colleague, Natalie Cone. Today, we're bringing you another installment of our Future of Work series with OpenAI's Chief Economist, Ronnie Chatterji. Ronnie, welcome back.

Great to be here, Mark. I feel like I'm like Alec Baldwin, who hosts Saturday Night Live a bunch of times. I feel like I call on forum a bunch of times. I really enjoy it. It's been exciting to kind of share ideas here. And the cool thing is the economy is moving so fast, or AI and the economy is moving so fast, there's always new stuff to talk about. So great to join you.

Well, it's great to have you, Ronnie. And I think, yes, you're at least a member of the five-timers, seven-timers club here.

In today's conversation, we're going to explore one of the biggest questions on people's minds. And that's, how can we prepare for the jobs of the future as AI adoption grows?

And whenever I talk to people about AI, the constant question I get is, what does it mean for my job? Or how will it affect my company? These are exactly the kinds of questions that we're going to explore here today.

And at OpenAI, we believe that AI has the potential to unlock more opportunities for more people than any other technology in history. But it's also going to be disruptive. Jobs will look different. Companies will need to adapt. And we'll have to learn new ways of working.

We're going to talk about how jobs are changing, the skills workers will need to remain competitive, the role employers play.

play in adopting AI responsibly? And what guidance should be giving students and educators as they prepare for careers that don't even exist yet?

Ronnie, thank you for being here. And before we jump in, wanna remind our audience to send us your questions. Just please go into the chat and we'll try to answer as many as possible.

But Ronnie, my first question to you is actually about the big news today coming from OpenAI that's right hot off the presses, that we have a brand new OpenAI jobs platform as well as OpenAI certifications. It's all about building AI skills and trying to match them to employees and employers. And the question I have for you is like, what's the significance of this and what kind of problems is this trying to address?

Well, yeah.

Yeah, it's an exciting day. I'm particularly sort of pleased that we're putting forth some potential solutions in this space. One of the things I do a lot of my job is the research, just trying to figure out how it's affecting the job market. I read other people's research on these topics. And there's been some amazing papers that have come out recently on these topics. And we'll talk about some of those later on.

But the big question, right, that we need to answer is, regardless of how jobs are changing or which shops are changing, we know that AI is gonna cause a lot of issues in the economy where there can be opportunities, but also challenges. The question is, how do we prepare people to sort of seize those opportunities? How do we prepare people to overcome those challenges and maybe find ways to use AI at work? And I think that's what today's announcement is about.

You have this idea that there's gonna be a whole generation of people who wanna learn AI skills. And I can tell you, someone who's spent a lot of time in open AI learning how to use chat to be better from a lot younger.

more experienced folks than I, there are skills to learn. There are ways to do prompts. There are ways to get more out of chat GPT, for example, not to mention the API and other ways to use artificial intelligence. So if there's this demand for people who want to learn these skills, the next question is how do you teach people this stuff? How do you get people the skills they need to push forward in their careers?

And then the second thing is where are the jobs going to be that require those skills? Am I bringing together employers and employees to figure out what those skills are going to be and what jobs are going to be there on the other end? That's the match we need to make.

And in that way, OpenAI is going to play a really important role as a market maker to really make sure that employees who are getting those skills are finding amazing opportunities. And that's what's really exciting about the announcement tonight. OpenAI Certified is going to be, I think, really popular, probably even more popular than anyone can imagine given how much excitement there is on AI.

And I'm also looking forward to seeing what the employers come up with.

in terms of their ability to design roles that use AI. That's gonna be something I'm sure my team will be studying.

Yeah, and you know, Ronnie, there's a lot of talk in when it comes to the economics of AI and everything that you're kind of doing on, this is gonna mean job losses, this is gonna change my job, but also there's a lot of creation too that goes along with this. So there've been with every technological change and thinking about like a lot of industrial assembly lines, thinking about the internet, thinking about computers, so many other new jobs that we currently have have been created after those technological revolutions.

In fact, there's one study that shows that a majority of current occupations didn't even exist, today didn't even exist back before World War II. So kind of what are some of these like new jobs that we could actually end up seeing and that.

in the next five, 10 years? Oh, I think it's definitely very likely that we'll see the same kind of job creation from AI in ways that we can't expect. Just like, I mean, pre-internet, can you imagine all the things that people are doing now in terms of jobs and occupations? I often tell this joke about like, if your kid told you in 1990, they wanted to be an influencer, you wouldn't even know what that means. And now it's like one of the most popular things that kids wanna be. And I think that when you think about AI, there's gonna be so many things that we cannot predict about those jobs and what they're gonna be. And I think when you think about, though, the skills that we require, and that's what, you know, with today's announcement on Certify is all about, I think there are gonna be jobs created where you're gonna have to leverage technology to achieve outcomes in ways that will scale human effort. So if you think about this, it's like agents are gonna be a really powerful part of the workforce in the next five to 10 years, that timeframe you mentioned, maybe sooner.

going to use agents, almost be agent bosses in some way, to manage all these different agents to accomplish different tasks. So let's suppose you have a job as a controller inside a company, right? But now you're able to use an agent to do so many of the tasks that were ahead of you before, it clears your place to do a lot of other things. So I think a lot of the jobs of the future are going to be like the jobs of the past, except with AI, right?

Second thing is, there's going to be some jobs that we aren't really thinking about. Think about all the lawyers who are going to have to understand in what places agents can actually complete transactions. How are agents going to be governed by new laws? What things can they do and can't they do? There'll be a whole areas of law that are created about this, and people are going to need to figure this out.

So if you think about both the work of complementing existing workers in their current roles and all these new opportunities to figure things out for challenges that don't yet exist, there's going to be a ton of opportunities in this space. I think the challenge, of course, for economists, we've

tend to look at backward looking data and try to figure out what's happening. The forecasting thing is pretty difficult. And the best way to deal with that is to give people the broad based skills they need to be flexible in the job market. And that's why I'm excited about the training components that we've been supporting. And we talk about the creations that could occur, but where do you see some of the bigger disruptions in the current labor economy right now on some of the industries that might most be impacted by AI?

I mean, I think about it all the time. And I think what the cool thing about working in open AI is like, this is a big topic of discussion everywhere. People wanna understand, figure it out and be part of the solution. I think for remote work particularly, anything can be done remotely, we're not interacting a ton with other people and the tasks are dealing with relatively structured information. And your job is to analyze that structured information and have some sort of output. And that's gonna be a lots of kinds of jobs in sort of telecommunications, retail, IT.

call centers. If it has those characteristics, right, that's gonna be easier for AI to do. Why? Because a lot of the human interaction, when you're working remotely, is not gonna be as important in that job. So if it can be done remotely, it's more amenable to being done by AI.

I feel like, you know, Mark, we see each other in the office, we have colleagues, a lot of our work is like personal, human, EQ related. But if you're in a job where you don't have as much of that, that's a job really to think about, okay, is this where AI could really change a job or make a huge difference here?

The second is, if you're dealing with a lot of structured info and your job is to summarize and analyze a report out, if those reports are relatively standardized and don't require a lot of human judgment, that's also something AI can do.

My hope is that we're gonna find ways, though, to do that kind of work with AI and let the folks who are in those jobs now either take on new tasks, move up or move into a different direction in their career.

and do new kinds of jobs, creative AI. That's the hope and the opportunity. It's what we've done since the beginning, but it doesn't happen without a concerted effort from individuals, from employers, and from government to make sure people get the skills they need to get those opportunities. We'll fail if we don't make that happen, and that's why we're trying to do this work right now.

And you talk about skills, and if you're an employee right now, and you mentioned the human skills to be able to have, the knowledge, background, the flexibility, what are some other important skills that we think if we wanna be open AI certified in this new economy?

I think definitely knowing how to be sort of comfortable with the technology, how to use prompts to get to the answer you want. You know, those things could also change as the product evolves, but we need to learn how to adjust to as we have different versions of the product. But for now, knowing how to prompt, knowing how to use AI to achieve your goals is gonna be really, really important. I tend to ask.

with the best prompts are. So maybe that's a strategy in and of itself. If you're a developer, right, you're working with the API, learning how to make that work for you, learning kind of what kinds of tasks you can do, where you need more supervision, where it can do it on its own. These are all things gonna be really, really important. I think being able to complete end-to-end workflows in particular areas, like if your job was to basically make sure that a report was approved by four or five different people for four or five different reasons, can you build those workflows into something like ChatGBT and make sure that that workflow is actually completed with the fidelity that you need to make sure that you deliver your product. These are all the kinds of things you're gonna need to learn how to do, whether you're working in merchandising, whether you're working in finance, whether you're working in retail.

And I think that's something to look for on the skill side. I will say, as an economist, I mean, I'm looking to employers to help sort of define what those kinds of skills are gonna be needed, right? I think you don't wanna be in a position where, let's say folks in academia.

or folks inside sort of only one AI lab are trying to figure out what are those jobs going to be? You need the employers who are actually going to employ folks to say, hey, here are the AI skills I need. And that's, I think, going to be part of this process and something that makes me feel really good about it. Because when you have that collaboration, you're going to get better outcomes for people.

And Ronnie, you mentioned the employers. I mean, tell me some of the conversations you're having. They're looking at this whole new landscape. They're seeing what AI can do for their businesses. What are the discussions on saying, look, this is the kind of exact person I need to be able to have to kind of make these agents work? Flexible and fluent is what I think of. They want people who are flexible, because the job descriptions are changing fast. And this is challenging for all of us. The skills required, the job. I often joke with people, I've had three jobs since I've joined OpenAI in many ways. It's changed a lot. And that kind of change can be really disconcerting. It can be difficult. It can also be exciting, depending.

depending on how you think about it. But I think for a lot of folks, employers are looking for flexibility, knowing that some of the things we're doing now at AI could be very different even in six months.

The second thing I think is fluency. I mean, I talked to people who are just in the beginning of their AI journey. They haven't used it to solve a lot of different problems. They're doing sort of very basic prompts. I've learned a lot just being at the organization. I'm sure other people can too, but I think they're looking for fluency and experience with this to go open it up and start working.

And sometimes experimentation is really good. I've done this in a lot of different areas of economics where I've been talking to chat about economics and trying to learn what I can from it. I think looking for evidence that you've experimented and explored and learned something from AI, regardless of what tool you use, that can be really important.

Those are the two things I hear. And the last piece I'll say, Mark, which you didn't explicitly ask about, but I think it's important, is like, we're still looking for the human behind that because at the end of the day, you're gonna lead a team. You're gonna have to work with people.

cross-functionally, you're going to have to not just produce the content, but explain it to other people. And those are all things that they're looking for humans to still be able to do with AI. And I think that's going to be a key part of the job force, our workforce going forward.

Yeah, and Ronnie, with this new announcement that we had on open AI certifications in our job platform, we ended up having just a new poll that open AI kind of distributed through our sub stack, the prompt that's out there today. It shows that three quarters of small businesses, according to a poll that we ended up commissioning, saying that they believe that AI skills are instrumental and important for their businesses' future. Like that they end up seeing this, even the smallest of businesses that are employing less than 10 people say, you know, three quarters, this is what I kind of need. I mean, to me, it's like, it's interesting in whether you're kind of getting in some of your conversations.

on the AI has has changed so rapidly in such a small amount of time but a lot of employers and a lot of businesses seem to be getting the message. It's so interesting Mark you bring it up because you're right like when you talk to large companies and I do some of that and you talk to startups and small businesses they're all equally interested in AI.

You know sometimes we talk to things you'll say okay this is a big company thing a fortune 500 type of thing then other times you're saying oh this is only a Silicon Valley conversation you know startups or Research Triangle Park where I'm from and other times there's like a small business issue let's say I'm working capital or bank loans and you're hearing it more from them that you would be from others.

What to do about AI is something no matter what kind of business you have people are talking about. I think that is really interesting and speaks to both the potential but also sort of the uncertainty that people have.

I think people know this is gonna be a big deal for a business right. If you're a business leader and you're thinking not thinking it's gonna be part of your future

you're probably missing something. But how exactly to do it, how to learn, how to implement it, how to learn from those experiments, how to make sure that you're staying ahead of your competitors with this, that's really, really the key in terms of how they're thinking through it.

Small businesses are the engine of the economy. A lot of my research, as you noted, some economics and entrepreneurship and small business, we know that in the US and large parts of the world, small business is the driver. Most people are working for small businesses and the job growth is coming from sometimes new businesses that are often small. And so it's really important to focus on the sector and think about technology adoption.

Sometimes small businesses that have been around for a long time, well-established, not growing as fast and not as willing to adopt new technologies. And I think that's gonna be the challenge.

We have to make it really easy for small businesses to use AI, to experiment with it, to make sure they don't fall behind bigger organizations that have more sort of capital to invest in these projects. And AI can be...

really easy to use, but the implementations and making sure your employees are trained up, that can take some time and resources. And we have to make sure that small businesses are equipped to do that if we want to keep the dynamic pieces of our economy going.

Well, Ronnie, we've talked about employees, we've talked about employers, what about students? Talking about a lot of students who are in college, in business school, in graduate school, maybe high school or below. What are, you know, when we talk about these AI skills, what needs to be accomplished from the student's end?

Well, first, I just want to say having kids myself and talking to so many students at all levels of school, it's like, please don't stress out too much. You know, I think about my own kids when I say that, which is the world is really dynamic and exciting. You know, I will talk to so many students who are going to start like an AI-based startup and they're going to, you know, move out to California or New York or wherever it is and try to raise venture capital.

or they're gonna start a small business in their town using AI. And it's really, really exciting. And that like vibe has been like really notable as far as AI, right? So that's exciting.

Other students who had more maybe traditional career paths in mind, they say, oh my gosh, I'm about to make this huge investment to go to law school or medical school or do an MBA. And I don't know exactly what's on the other end. And that worries me. That's really natural. And the same dynamism that creates opportunity economy creates uncertainty. And so it's really, you know, there's no single answer. You can't figure it out. Don't stress out.

I have a lot of advice for students on that angle. I do think, right, being fluent with the tools in your area of interest is really important. So if you're interested in medicine, thinking about sort of AI applications of healthcare, really, really important. And your way to kind of get exposure to that is use it in the application area that you're really passionate about. You don't have to turn into a coder. And in fact, you know, I asked Chachi Pateen for a lot of coding advice. So you don't need to turn into a coder.

but you can apply it to solve the problems that you're interested in. I think that's the best way to proceed because I think in most of these fields, there's gonna be a tremendous, probably growing role for humans, but some humans who are using technology to do the work and more. And so for students, maybe thinking about their career as AI, as a tool, as a copilot, as a coworker, is probably a better way to think through this.

I also think not losing that element of the human connections, building the networks with your friends, your professional contacts. Yes, AI is really important, but in many, many ways, you can see that the human connection can get more important in these coming years. And so I think not forgetting that piece really matters, especially at a time that so many of us are on our phones and trying to figure things out and glued to a screen. It's really, really important to be able to connect with other people in all kinds of work you're doing.

So those are two things I think are really important. Trying to predict which fields are gonna rise and fall is a lot harder, and you can definitely read the news and pay attention.

and see which areas are growing, but you also have to follow what you're interested in. And I think that's the combination, trying to get what you're interested in aligned with some of these technological changes would be the way I would think about it for students.

Ronnie, you mentioned kind of the uncertainty or the worry on trying to prepare for jobs that don't even exist right now, but might exist five or 10 years from now. Is there any historical parallel? You know, obviously this is happening so fast, faster than when I was a college student getting email for the first time and seeing the internet develop. I don't think that there were any high-level discussions the way that you and I are having about the economics of the internet age or the economics of personal home computers kind of put our kind of situation and the labor economy into historical parallel.

I think, you know, when I think about the internet and I think about the jobs that got created, let's say e-commerce that didn't exist before,

you always wonder, okay, how do people get into those businesses? And this is gonna be the same thing for students in AI. There's gonna be all these new kinds of businesses and business models where there's not gonna be a thing on the resume that say, oh, you did that before because it didn't exist before. So how do you put yourself in position to get those first of kind jobs, those first of kind companies? I think there's two different ways to think about it.

Some people are gonna be fortunate enough to be in position to get hired by an amazing organization that's really on the cutting edge or work for an amazing manager who is really leveraging technology, doing new things. If you think that we're an inflection point in the economy and the skills you gain, let's say, in your 20s are gonna be really important to your success afterwards, which I think is true, trying to work for the most innovative company that you can get access to and the most innovative managers, ones who are taking risks, probably a good way to expose yourself to some of those tools of the future. It might not be the biggest brand name, but working for a place where you can actually learn kind of what's new and interesting could be really, really helpful.

important. And it's a different way to think about a job than just going for the employer that pays the most or the employer that has the best, you know, the highest stature. So if you have an opportunity to be choosy, and I know that's hard, working for a manager or an organization that's really experimenting with new tools and new business models, that's probably a way to position yourself for what's next.

For other folks, right, it might be that, look, they're working in the more traditional industry where that change really hasn't happened, or they're in a job where it's more static. And that's fine too. I think on the side, though, there's lots of micro-credentials and courses and experimentation you can do.

I always tell people, like, when you're younger, you have a lot more time to invest in human capital in so many ways. This is the time to kind of gain those skills on the side and get prepared. That way, when someone says, hey, I have a job for something that's never been done before, are you qualified? You'll be able to say, well, here's the work I've put in to prepare. Not really knowing what it was going to be, but I've done all these different things. I think that can carry it a long way. Because remember, if it doesn't exist yet, no one really has.

the perfect resume to do it and anyone who does probably did it on accident so if you can develop the flexibility uh in those skills and do some of that stuff on the side I think that would put you in a good position.

Well Ronnie we're already starting to have some questions trickle in so you want to start tackling some of the questions that I that I see are getting pinged to us right now?

Yeah sure I'll let you tee them up and uh you know any hard question Mark Murray's going to answer and I'll take all the easy ones. No I'm kidding just uh however it goes.

This is teed up for an economist like you Ronnie and this is straight up the first question we ended up getting saying hey it's great to see supply side innovation coming from open AI on today's announcement how do you see the demand side receptivity and change?

Oh that's a great question and I like the the framing too thinking about it as like a supply side solution right how do we increase the number of people in the workforce with these skills to use AI and like

I know there's a lot of folks, and I'm part of this, who are analyzing the numbers, debating what's gonna happen, worrying about what's gonna happen, really, really important. But at the same time, we can start to be part of the solution, which I think is fun. And that's why I like working in an organization like this. I can do a little bit of both, right?

So when it comes to increasing the supply side, you're gonna say, okay, what skills, who wants to sign up to learn what kinds of skills to get OpenHAZ certified? That's exciting. I think that's gonna lead to a lot of interest.

On the demand side, there's two pieces of it, I guess. One is how many workers, students, are gonna be interested in getting OpenHAZ certified? I have a sense that you're seeing a strong interest in this.

Look at the majors in computer science increasing around all universities. Look at the number of people who are interested in learning about AI or joining a forum, like this group here. Look at people using ChatGPT and other products. There's a ton of interest to get smart on AI.

And I think, man, I'm gonna do these classes probably on the side when no one's listening.

because I think this is gonna be really interesting. I'm sure I'm gonna learn a ton from doing this. I think there'll be a ton of interest if I'm evidenced by all the interest in AI in the economy. On the employer side, I also think there'll be a lot of interest because there's gonna be these new roles across the organization, finance, marketing, to use AI.

I think what is gonna be the challenge where we can help a lot is, okay, what are those roles? What exact qualifications do you need? And who's the right person for that role? And I think if we can create a nice sort of matching between the people who are getting training and the companies that are hiring, and this is very hard to do, and there's lots of experiences where this hasn't worked in other areas, but if we can do this right, we're gonna be able to let people who get OpenAI certified have opportunities that they wouldn't have had otherwise.

Then I think you'll see the demand and supply side start to meet. A lot of interest from employers, if you look at job listings, but I think it's also a moving target that we have to keep careful.

because the requirements, what it means to know AI, to be fluent in AI, are also changing every day. That makes our job harder, but it doesn't make the work any less worth doing. It actually makes it more work. Ronnie, I'm so glad you ended up mentioning communities like the OpenAI Forum, kind of helping to create the demand because that seems, you know, I've been a member, I've been working here at OpenAI for the last six months, seeing all the activity coming from the OpenAI Forum. And to me, like that, the excitement and all the conversations, I do think, really do help with on the demand side. So that's a great question.

Another question, and we ended up getting this from Andrew Holtz from Machine Learning Tech. And the question is, given that AI will be a commodity in 10 to 15 years when today's children enter the job market, what is your vision for how their education should be changed? It's a great question.

I always tell, thank you, Andrew, for the question. I always ask my kids, like, think about AI today. And they complain a lot, actually, because they'll say, oh, it didn't answer this question right, and it can do all this other stuff right. How come it can't do the simple thing? I said, this is the worst it's ever going to be. So when you join your organization one day, your first job out of school, and for them, it's a long way off, like, AI is just going to be part of like the package. It's exactly what Andrew's saying. Like, you get a laptop on the first day, you get your phone, like, AI, in whatever way it's instantiated, will be part of that, right?

Right now, we think about chatbots a lot. In the time period Andrew's talking about, it could be instantiated in different ways. Right now, it's a chatbot. But you will have intelligence at your fingertips in a way that is intuitive for you to use with definable problems you're trying to solve. And using that in those jobs is going to be something like a table stakes, right?

But over the next 10 to 15 years, Andrew, there's going to be a lot of differences. Some firms are really going to get this.

build this into their workflows, others are going to be slower. Some firms are going to attract the kind of employees who are going to be savvy with this stuff and come up with new ideas, others less so.

Some countries are going to be ahead of the game because they have the right infrastructure, right, and they can actually serve all that compute. There's going to be a lot of variation as we go from here to there, and that's going to create a lot of opportunities for the economy, right, and for individuals.

And so I think one thing to think about is, on the education side, learning how to be part of that wave in a positive way, to join the right companies at the right time in the right role is a big part of that.

In terms of how you learn and acquire information, I think it's the same thing in schools. Some schools are going to pick this up right away, and I've been, you know, I talked to lots of schools about this. Some are going to integrate it into learning, teachers are going to be using it in the classroom.

Those schools will probably produce students that are savvier, more comfortable with AI, but they're also going to have to make sure that students are using it in the right way to foster critical thinking and learning.

And I expect there'll be a lot of experimentation on that as well. I think when we get to the stage you're talking about decades from now, like I imagine, much like my kids are using the calculator to use for more complicated math, they had to do all the numeracy stuff at the beginning. They learned how to add and subtract and divide and multiply, and they weren't allowed to use calculators during that piece.

And now they're doing higher level math and they don't have time to do all the addition stuff because if they spend all the time doing that, then they're not gonna do the harder algebra problem. So they can use a calculator in those instances and they get the higher level math. And then one day they'll use the graphing calculator, and they can use that to do more kind of calculations. They still have to learn how to draw an X and Y axis and draw a graph from scratch.

So I kind of feel like AI will evolve in that way. We'll wanna introduce these tools at the right times. It's probably not gonna be called AI. It'll be part of other kinds of products, other kinds of devices, but that intelligence will be an important part of their education.

My dream is it'll just help people learn more. I feel like, at least as an adult, I have the advantage of having been educated before, so it didn't have that conflict for me. But I use it to just learn new stuff every day. And as a person who has lived the life of the mind, it's such a tremendous opportunity. I feel like it's really expanded my knowledge. And I want my kids to have that same opportunity, but I definitely want them to learn the basics, including critical thinking and doing math before they get the chance to do all the other stuff.

Well, you mentioned learning more, critical thinking. Here is a great seg question, and this comes from Natalie Sims, who asks, what's your view on how we can use AI to augment our capability without eroding our cognitive skills by making us lazy thinkers? It's, I mean, I know this is, I'm an educator at heart, you know, I'm a professor and doing this job now, but I think all the time about the classroom. And like, if you're leading a discussion in the classroom.

As an MBA teacher, I would do that with MBA students. And like, you're making that eye contact. Everyone's engaged, right? It's not like people should be looking down at their screens and typing things in. They're looking at you.

Now maybe they did preparation beforehand with their head down. When they're in class, we're all experiencing that together. What you want is in that situation is what you wanna see from the executives at like a leading organization, which is like being quick on their feet, making a smart analytical point that builds on the things our colleagues have said.

How do you build those skills? How can AI help do that? And I think a lot of it is in the preparation, the research that you're doing. And also critical to this is like exploring counterfactual ideas. So many students, so many people, they don't think about the alternatives to what they're saying or the easy things like, well, what about this? What about that?

AI is really good at critiquing your work. And so I think actually it can be used in that way to foster more critical thinking. So I hope students use it in the preparation. I hope they do that to kind of

to kick the tires on their ideas and surface counterfactuals. And then I hope when they come to class in a sort of Socratic method that I would teach in, they're ready to engage on the human element and convince their classmates, not with the machine, but with their arguments and their brains. And I think that will really make a big difference.

I think that if we let people outsource to any kind of AI and then submit like written assignments, and there's no way to know who generated this or what it's saying, there's no way to defend it or argue about it, that's not gonna be a good model for education. So for me, it's like, it's the way we structure our classroom, the way we use pedagogy.

And as a professor, like I wanna encourage them to use it in a way that I think is gonna be conducive. And maybe different professors have different policies, different schools will do different things and we'll see kind of what is best. And I'm a big believer, obviously, in research and evaluation. So I think we should experiment, but also make sure we learn from those experiments. So that's how I'm thinking about it right now.

Ronnie, let's go back. We got a.

a question about our new news today about the OpenAI certifications and the OpenAI job platform. And the question is, what feedback or demand were we hearing from your users and employers that prompted the certifications and the job platform?

I think, I mean, there are others who are closer to this one, but I would say, what are some solutions? Like, how can we be part of solutions? And I think it's such an interesting time with so much innovation happening in the space we're in, and it creates a lot of these open questions. Some people are excited about these questions. They're gonna try to solve puzzles we've never been able to solve with AI.

Other people are concerned about some of the changes it's gonna bring. And I think, as we continue to work our way through all those and study them and learn the best way we can, as a researcher, that's the only way I know how, there's also a role for proposing solutions.

And I think that was the thing we heard, and it's a very solution-oriented group of people at OpenAI, I'd say.

There's a lot of people who do want to analyze and research. There's also people who say, hey, can we try to be part of the solution? Can we try to fix this? Can we try to help people? I think that's where this comes from. And what I like about this is there's not this overconfidence that we have to figure it out. I mean, there's a lot of humility about, well, we've got to figure it out together.

But there's a sense of, well, we can't pontificate and procrastinate on this. We have to move forward and try to propose something to help people get the skills they need. That's kind of internally what I think was really a big driver, at least from my vantage point. There might be others who have different views closer to that.

Yeah, and Ronnie, we're getting just like ping with a lot of questions about the OpenAI certifications, as well as the OpenAI JAP platform. Hey, where can I get this? When is it going to happen? Trust me, we're going to have a lot more information coming soon. It will get to everybody.

Ronnie, you and I are actually, you and Natalie might have a separate.

conversation about this on how you get that OpenAI certified check for you. All of that information will be coming soon. So please wait, and we'll get that. Oh, my gosh. Yeah. But on a related question, and this is more the theoretical, Ronnie, one question we got is, how can someone self-assess their current readiness for an AI augmented role? Like, what are things before you end up taking our classes and getting the certification can you start doing right now?

Well, I think as the folks who are going to design the effort and figure out what the training is going to be, they'll have a lot more to say about it. As an economist, I think about this following. When I think about my own work, I try to take a task that is a big part of my job today, and I try to run it through lots of different tools and figure out, when is it achieving the performance that's consistent with what I would like to achieve?

want as a human, let's say, who is doing the role.

I think having fluency in a couple different tools and knowing which ones are better for what. You might want to use one tool for coding and another tool for chatting about your career plan or doing some financial analysis. So I think being able to say, hey, am I fluent with a bunch of different tools? Can I use them? Does that work?

I think the second thing is like, can I solve problems and complicated workflows, not just ask a question and get an answer, using this a multi-step processing, for example, more of the agentic work. Some of the longer horizon tasks like we use through deep research. These are some things that just start in terms of doing that.

To my knowledge, and I could be wrong, there's not like an assessment on a one to 10 scale of how AI savvy you are. And if you know about it, let me know or send a comment on it. But I think even an assessment like that will probably change pretty quickly. I even think the way we're prompting now might be a little bit different in a year or two. It wasn't just historically, it wasn't obvious that the...

chatbot interface would be like the product market fit that really brought AI to hundreds of millions of people. That's relatively new and I don't think everyone expected that. And I think it's not clear necessarily how you'll be using it in the future. But right now, being able to be familiar, prompting different models, getting answers, assessing it against the objectives that you have. That's a pretty good way to start. And I hope that as we develop this program, people who are experts in pedagogy and technology, I think those two intersections will help us design sort of assessments and benchmarks. I think that's really important. It's a good question. That's what I have to say about it right now.

And Ronnie, the flip side is like, we were focusing on what employees can try to do now. What can employers, like how would employers like start to try to assess AI skills? I think, and this is a hard thing inside the organization, just like it's hard for the individual, but like, what are the roles of the future gonna be? What is the new work? You mentioned that study going back to 19...

about all the new work that's being created and new job titles and things like that. And companies are all moving so fast. And sometimes it's hard to block out the time to do scenario planning and think about the future. And I think to the extent that organizations can do that for very senior leadership and think about what are those jobs we're going to need tomorrow and what skills do we want? The companies that do that the best, the ones that know what they're looking for are gonna have the best chance of attracting new talent. So just like employees are trying to do the assessment, I think employers have to reflect as well on both sides. I know there's a lot of bandwidth constraints and there's a lot of challenges against thinking that far ahead, but that kind of scenario planning can really matter to prepare organizations for what's next. Otherwise you might find out that your competitors get there before you and it's a much more competitive market to hire those folks to those skills than you might've imagined.

One question we ended up getting, can individuals outside the tech industry benefit from these programs?

and how, you know, I do think, Ronny, there's this perception that we're talking about all the coders in San Francisco, the researchers, and not necessarily people like me with liberal arts degrees. And so, you know, how could folks like that benefit? Open for all.

I mean, I would say, you know, this is not going to be limited to people who are already really knowledgeable about the tool. I think, if anything, it's going to be even most beneficial for people who are more like beginners. And so, no, this is going to be open and for all levels. And I hope if you're someone who's like really savvy with AI, there'd be something for you to get better or deepen a skill.

And there'll be people who, it's very new, and maybe they haven't used it in their job before to get up to speed. So I think there's going to be room for all as we design this.

And I think that's the way you can impact the most people, by the way, too. We've set a really ambitious goal to help a lot of people, right? And I think if you're going to do that, you have to make it widely available.

And, you know, this is key to just how people think about the mission, right? I mean, Mark, we talked about this offline, like the reason we both joined in different ways and different reasons is for, or different times and different backgrounds is for the same reason, which is like the mission to benefit all humanity.

We both care about that and resonate with it from our different backgrounds. And so this is the key to making it available for all and not sort of limited to someone with a specific tech background.

Ronnie, this is a super high level question we ended up getting, but the question is what gives you hope about the future of work? And I know that's like, that's a 30,000 foot view, but like kind of talk. I mean, you, you obviously look at a lot of work and the economics of work. Like what, what, what gives you hope about the future?

Yeah, it's a great question. I would say, and it's not one that I, that I get that often. So I really appreciate it. And there's a group of people that I know, and it's.

that's maybe a smaller group than I wish that really love their job. And when I asked them about like, hey, how does AI help you in your job? Like they often are ones who are like, hey, I did this, I did that. And it's unlocking all these great things. And I think I'm excited knowing how for people who love the work they're doing that they can like give this, basically do more of it, make it easier, optimize different things.

There's a lot of people I talk to there's parts of work they don't like. And sometimes it overshadows the things they like about their job. And I always wonder, could AI take that thing that's the most annoying, the hardest task, that report that's the hardest thing that you have to do every week or every month. Could it make it easier in a way to freeze you up to do other interesting things?

You know, there are so many people maybe who are bombed down, I think by tasks that need to be done. They don't have the time to do some of the other things in their job they really enjoy. I think a lot about people who deal with clients or serve other humans in their job and they wish they had less time.

on paperwork and more time on the human element. That to me really gives me hope, Mark. And I, you know, we have to work this out in all different organizations and enterprises, but to see someone be able to spend more time with a human, whether you're a financial advisor or social worker, that gives me a lot of hope. And the last thing is I think in a job market that has a lot of change, we see opportunities for new things that we never could have imagined. And while that can be disruptive, it's also really exciting for young kids.

I look at students now and I think, gosh, like I'm gonna have a really hard time thinking what they're gonna grow to become. And it can be something totally different than what I'm imagining. And while there's some challenges in that, there's also like a lot of excitement. If the world was static and they were just gonna go into the same jobs and career tracks that their parents had, I think life would be a lot less interesting, you know? And I think there's gonna be a lot of opportunity to have new and exciting careers at the same time.

So that's the kind of stuff that gives me hope. And maybe the last piece is just, I always kind of bet on humans. Like even this.

In this machine age we live in, I think humans are really flexible and we've gone through a lot of these transitions and come out large part for the better on the other side and wealthier and happier in many cases. There's challenges to that. Sometimes the wealth isn't distributed in the right way so not everyone's benefiting and sometimes mental health and wellness is a really key part of this too. But these are things that I think are really on the radar for everybody right now. We're working collectively with folks to try to solve these problems. So I'm pretty optimistic about the human element to thrive with technology. That's kind of what gives me hope.

Got it. Ronnie, we've got a question on math from Leon Como who... Oh no, Leon, don't go down there. Here's his question. Perhaps there's a company that used to have 100 employees and now can it really hire 200 if there's like 120 humans but 80 AI agents serving the bigger pie? How does this

all make sense economically? That's a good question, Leon. I thought you were asking me like a real math question, like to invert a matrix or something on, on, on live form. No, it's a really good question. So you're seeing this.

I mean, it was publicly reported that, you know, McKinsey has like created all these agents, right. And, uh, they have fewer workers than they did during the pandemic, but not by a lot, but they're also hiring a lot and how can they be hiring at a time that they're also hiring a lot of agents, not going to be hiring humans and agents.

And there's a couple of things. One is even though the agents can do all this new work, it also creates new opportunities for, let's say a McKinsey or another consulting firm to enter the new markets and they could before, let's suppose there's a regional market where they haven't been able to have consultants there because maybe it was too costly.

Now through agentic work, they might be able to open up business relationships, prospecting, get new potential clients in a way that creates some of that demand, right? So agents aren't just necessarily going to replace human work.

And in some sense, I think it's going to augment and then create opportunities that didn't exist before. If you think about now all the jobs that you're going to be able to employ agents for, you're going to need someone to supervise the agents, you're going to need to be able to someone to coordinate between them, you're going to need someone to look at their output, make sure verify what they're doing correctly. All these things are going to create jobs around the chain. So I don't think if you think about agents as a one-to-one, I think about if there's a limited amount, if we can only write the same number of lines of code in the world and you have agents writing code and humans writing code, it would be a pure substitution.

But actually, there's not a limit on the amount of code we need in the economy. We can write more code. And in some sense, that's a way to think about the abundance that could happen when you have more agents. That can be true in every case. But I think if you look at programming as one example, to me, the limitation isn't that there's a fixed pie of code that we need to write. And in fact, if you can infuse code into lots more things in the world, you can actually make a lot of improvements and create a lot of economic value.

So I don't see it as a one-to-one. That's where the Mac can help and not sort of be totally zero-sum. Thank you for that, Ronnie. Another question that ended up coming in, and this has to do with on the education and skills front.

Do you see this as a moment where people without degrees or prior experience can enter high impact roles using IAI now? Is this AI like the great equalizer out there? Some of the studies are finding, it depends on the context, that when you give AI to a set of workers that are heterogeneous in terms of their experience, the folks with less experience or less education, they're able to catch up and meet the folks with a little bit more expertise in education on a particular task. Not clear whether that's going to generalize across all kinds of settings, but it is what you're seeing in some of the studies, the great leveling that you're kind of referencing in your question.

The other thing to think about is like, to the extent that knowing how to code a particular language is a barrier to entry of a particular job.

your ability then to learn that more quickly or be able to use AI to do that might then make that job now accessible to you even though you don't have that qualification. And so I do think you're gonna see the ability for people to get into roles using AI that they maybe couldn't before. AI tools might democratize some of these skills that previously were gatekeepers into different professions and areas.

I also think, though, you have to make sure the rest of the economy is not standing still either. So you might be learning AI to do something and the person who already have those skills is also using it. So those things are gonna matter. But in general, I think it gives a chance for people to ladder up using intelligence and acquire skills maybe more quickly than they could have otherwise.

Learning the most important things about the capability but not necessarily taking the class to get a certification. Those are the things to look for. So interesting to see how it plays out. Some of the studies suggest in that direction but I'll wait to see if it's gonna work across sort of different workforces and more complex tasks than the ones that the.

studies have been focused on. Ronnie, we're going to take a couple of more questions before we wrap up. We're getting close to time here.

This is kind of a related question, but could AI end up helping people discover latent talents or strengths that they didn't even know that they had? That's a good question.

I think so. I mean, what's interesting, and it depends on the personalization features that you want to enable, but if you want to have sort of more personalization on these products, and you think of ChatGBT, if you talk to it about your skills, your career plan, you might ask it at some point, hey, based on what you know about my career path, my goals, what are some things I should be doing that I'm not?

Maybe it's just, hey, Ronnie, you should write a book. That would be a great way to extend your skills. So I think you could actually do that. I think it depends how much personalization you want, how much you want to share about what you've been doing, and that's going to be for every individual.

to choose. For me, Chat GPT knows a little bit about my career as I ask it how the best way to approach my job is. And so it may suggest some things I didn't know that I should be doing. It also might say, hey, look, you know, based on what I've written, like, are there some skills that maybe I have that are latent that I'm not even recognizing? And they might say, look, like the way you ask these questions, like you might be really good in this particular profession or engaged in this kind of discussion. I think it's definitely a cool opportunity. And I encourage you to experiment, you know, asking Chad about those kind of things. I think it's a good idea. I might do it right up the forum.

And Ronnie, this question is about continuous learning. And the question we ended up getting, could open AI certifications help re-imagine what lifelong learning looks like where education is less about a fixed path and more about empowering people to grow continuously no matter where they start? I love it. I mean, that would be my aspiration. I think, you know, we have this.

artifact of like the way schools organize and the ages people mostly like to go to school and for a long time many people like me who've been in academia for a long time people on the outside students non-traditional students administrators everyone's trying to figure out a way to make it more flexible because we realize that the learning doesn't end when you're 21 years old or 18 years old or whenever it is you need to learn throughout your life and not just for the next job but just also for a lot of other personal reasons and fulfillment and imagine if we could give people the opportunity to actually learn at different parts of their life in an efficient way at scale I think that is where AI can be really helpful because of the personalization piece we talked about like it can talk to you about what you're interested in

so if you want to learn more of a particular topic with particular kinds of sort of guardrails in place and particular kinds of readings and maybe a particular perspective let's say on economics or political science you can use AI to do that and if we could build that into education products

that bring you through a journey at any age, wow, that would be amazing. And then, of course, then you get to this point where in a situation where employers are interested in you acquiring those new skills, they can help you identify which kinds of, you know, sort of threads you might go down. And that could be really powerful, both just for, you know, sort of learning outside of work, but definitely learning on the job too.

Ronnie, we're going to end on this question. If you could leave a note to someone entering the workforce in 2035, what would it say about how to thrive in an AI-powered economy? What I would, what I would leave to them is I'd say, look, I think you're going to thrive if you're a complement to AI and can work better with it. Because I see it as being, you know, a powerful tool, but a tool for humans to achieve their objectives. And while that means that not, well, that doesn't mean every job is going to be the same as it was in 2020.

I do think for the 2020, 2035 workforce is going to be a tool to do things they never thought possible. So that's the inspiration I want to do.

Well, Ronnie, thank you for that inspiration. The OpenAI forum community, thank you guys for joining us. You're going to hear a lot more about us. You're going to hear a lot more about our OpenAI certifications and job platform.

I know you're going to hear a lot more from Ronnie in future meetings where I think we'll go from six, seven, eight, nine Ronnie Chatterley discussions here. But this has been a pleasure. I've had just a great opportunity filling in for Natalie, who will be back. But this has been a lot of fun. Ronnie, thank you so much for joining.

Mark, great to be with you. Looking forward to our next discussion.

Thank you. Thanks to the community.