278 Million Reasons to Pay Attention... Andrew Ng and Greg Hart | ASU+GSV Summit 2026

This is a great way to finish our main stage presentations. We have the privilege to have the leading learning platform in the world, the founder, Andrew Ng, with us as well as the CEO of Coursera, Greg Hart. Andrew, it's almost impossible for me to believe that Coursera is 15 years old. And I think Lucy combined you and Udemy already because with the number of learners will be on the platform once the acquisition is done, but as of the last quarter, a couple hundred million on the platform. Can you, when you started Coursera, what was your kind of vision? Did you think you'd be at 200 million learners in 15 years or what was different than what you expected when you started the business?

I think when we started Coursera, frankly, I was so naive back then. I thought, how hard could it be? So part of me wonders, you know, put this on the internet, surely there'll be billions of people that use this. I think we're still on that journey to keep on working hard to serve learners every day.

So, and just to give people an update in the room, so obviously the merger was announced with Udemy. What's the status of that and when do you expect that to come together? And then maybe just very quickly, why are you excited about the combination?

So, when you do a ChatGPT query and say who's the top AI person in the world, the name that comes back is Andrew Ng, which is pretty cool. So to have Andrew here, you know, with all the things that AI is doing, it's pretty special. But Andrew, from your standpoint, how is AI, how do you think the integration or the collaboration between AI and education is going? And where are you seeing the ROI, or are you seeing the ROI?

It feels like the integration of AI and education is still in the very early phases. What has rapidly changed, as Greg was alluding, is the skills that all of us should be learning. And so companies like Coursera, DeepLearning.AI, Udemy have been shifting what we deliver to learners in terms of helping them build more relevant skills in this AI era. And then I feel like all of us have this instinct that we can accelerate the delivery of learning, or more importantly, accelerate how much individuals can learn. And I think that's been much tougher. One of the great limiters for how fast people can learn is just the biology of the human brain, right? Our brains can only process so much information so quickly. And so there will be an acceleration. Coursera and others were delivering better learning experiences. But will the human brain ever learn 10 times faster than it does today? I don't know. But certainly what we have to teach people is rapidly changing. And I've been excited about, for example, I think everyone should learn to code. I feel like all job roles should embrace AI skills. I feel like, you know, every knowledge worker should be able to use AI to write code, to prompt, to more effective workflows. So that delivery of new sets of skills is really important and that's been moving quickly.

So, and Greg, you've now been with Coursera a little over a year. What is different than what you expected when you came in? And what were you most excited about right now?

Say that again.

That's...

And you're kind of saying it, but, you know, are the LLMs a friend or are they a foe?

And so, and you look at, I mean, by the way, this has an issue in education overall. It's certainly been an issue overall for online learning. And Coursera, you know, has created amazing access, but like, again, how do you, one of the things is how do you make sure that they complete the course and are they using it to advance their lives? How is AI, what are you doing to help in creating better engagement and helping the student complete a course and keep on going?

So you both were at the HumanX conference in San Francisco, was it earlier this week or last week? I'm getting blurred. And we certainly have a theme here of this kind of collision between learning and earning and the workforce and education. Again, I think Coursera is positioned really terrifically to be a part of this wave that's taking place. What did you learn there about workforce and adopting AI and how that's applied to the jobs of the future and how that ultimately fits in with the Coursera vision?

So, you know, one thing we chat a lot about at HumanX is the question of ROI from AI. Over the last year, a lot of businesses have let a thousand flowers bloom, bottom-up innovation spread, and that's great. We should absolutely keep on doing that. But the ROI from a lot of this bottom-up innovation is just not there yet. And I have a strong point of view on why this is the case and the important role that training has to play in resolving this. It turns out that most bottom-up innovation results in point solutions. A business often does multiple things to create value. A lot of bottom-up innovation will have someone identify one step of a process and automate it and make it a little bit more efficient and save some money. But this gives you incremental efficiency gains, which is nice but not transformative. The challenge I see when I speak with a lot of C-suite from, you know, Fortune 500, is it takes a top-down motion, some of the broader scope to look at all the steps to decide, with this step optimized, I'm going to redesign this whole workflow and that drives business growth. Maybe one tiny example from edtech is years ago we started auto-grading, and if you swap out a human grader for an AI grader, that's nice, it saves a little bit of labor, we should absolutely do that. But the thing that's maybe more transformative is you can say, oh, with AI grading, I can now get back to the learner in five minutes rather than a week. That gives you more opportunities to change the learning experience. So I think we're still in the very early stages of educating across many businesses entire teams to enable more bottom-up innovation but then also delivering that executive-level education so that bottom-up innovation or top-down innovation can meet and someone can say, oh, if you can redesign this part of the workflow, let's change the way we do business to drive business growth. And I think this is one of the reasons why it feels like there's almost unlimited demand for learning these new AI tools, both from businesses and from individuals. And I think we're all eager to serve that to help drive that ROI from AI.

Anything to add?

So, can I add one more thing to that? So I think in terms of workforce transformation, if you look at all the professions out there, the one profession that AI is transforming the most, I think, is software engineering. So the acceleration in coding from, you know, Claude Code, OpenAI Codex, Gemini CLI, other tools is very real and this is dramatically changing the training we need to deliver to software engineers as well. So my mental model is today we have many wonderful software building blocks, front-end, back-end, database tools, also AI tools, RAG, agentic workflows, guardrails, vector neighbors, whatever, and software engineers need to learn all of these, build many of these building blocks so that when there's an application product you want to build, that we know how to very quickly assemble them to create new software. And the workforce shifts in software engineering is very real. I see a lot of large businesses that frankly don't know how to do it. And if you've not seen an AI-native team operate, it is really unlike anything I have seen, say, two years ago, where today I will routinely assign a part of, let's say, two or three engineers a task and ask them to deliver in one month what I used to ask 10 engineers to deliver in six months. So if you've not seen this upfront, the acceleration is very real. And I think there's a lot of work to do to deliver that training, deliver that practice to let many individuals and many businesses understand how to operate more like these AI-native teams. One weird thing I predicted a year and a half ago, I think that there'll be a product management bottleneck where if the bottleneck is no longer writing the software, the bottleneck is deciding what to build, not the actual act of building. So I see, for example, a lot of software engineers that have a good user empathy that know what to build able to function much more effectively than a software engineer that expects a pixel-perfect fake design and thinks their role in life is just to implement it. So I feel like software engineering is the job that's changed the most. I think this is a harbinger of what may happen to many other professions as well. And so that pressure to deliver training to bring everyone along with us is a very exciting, you know, slightly stressful but in a good way type of pressure to be feeling.

A couple of related questions, but from what you see, and I know you think AI is changing everything and what the needs are, how does that change in terms of what we need to learn? You know, what are some of the key things that you see today? And if you kind of fast forward the clock, you know, till the next, you know, five years.

What are you seeing in terms of the value of the degree versus just having this, being able to have the knowledge and be able to validate that I actually have these capabilities? I mean, are we...

Yeah. And I mean, there's a business I think you guys could create just around all the data that you have around how people are learning, what they're learning. So what are some other, you know, kind of interesting facts about, you know, the 200 million learners you have, things you're seeing. So you talk about critical thinking growing faster than AI. What other type of observations or insights do you have?

So Andrew, you please.

And I have a contrarian point of view that I'm going to state and may get me into trouble. In Silicon Valley, I know within many circles it's trendy to talk about the job apocalypse. This idea that there'll be massive unemployment ringing the streets because many jobs may all go away. I suspect that the job apocalypse may not be nearly as bad as these trendy circles are saying. And in fact, recently, a couple months ago, there was a study from, I want to say, Citadel Capital that showed that software engineering job postings are actually up. Now, to be clear, there is a lot of job insecurity. Some of our teams talk to people at many levels. We see job insecurity not just among fresh college grads but even to very senior leaders. And to be clear, I feel sympathy for anyone whose livelihood has been affected by layoffs. Many businesses over-hired during the pandemic, during the era of zero interest rates or near-zero interest rates. So there is overhang. Layoffs is tragic for the individuals and families affected. Not dismissing any of that at all. At the same time, I feel like as people become more productive, a lot of our teams, we are just trying to hire more people, but the skills are shifting rapidly. And in terms of the skills I think people should learn, again, another used to be controversial, maybe less controversial now, one of the most important skills of the future is knowing how to get a computer to do exactly what, being able to tell a computer exactly what you want it to do for you. And for the foreseeable future, the language of computers is code. So what I'm already seeing in Silicon Valley as maybe a harbinger of what I think will happen elsewhere in the country and the world as well is the marketers that know how to code are outperforming the marketers that don't. My recruiters, my best recruiters, they don't read resumes by eye, they write code to help them screen resumes. One of my teams is building, has an offsite later this week. My front desk receptionist coded, so we should build a custom website just around the offsite. So I think because AI coding is so effective, in the future, one of the skills, out of many including critical thinking and so on, is being able to, you know, use AI to write code to build custom software. And I think if we do this, there's a potential for a flourishing of individual productivity and I don't think people that navigate this well will face as bad a job apocalypse as is trendy to talk about in some circles.

So the last question for both of you, and I'm going to ask slightly different for when I give you, Greg, your turn, but Andrew, you've been on the forefront of not just AI tech, you know, innovation for your whole career, but really education innovation. Now you've got the 15 years at Coursera. If you look ahead over the next 10, 15 years, what do you think Coursera could look like? What do you hope it could be and what do you think is possible?

I think the mission feels like we've made just a small dent in what Coursera should be setting out to do. I'm proud of the work that the company has done to transform lives through learning. At the same time, the world is going through a much bigger transformation than we've seen in recent memory. And so even more lives will be transformed, you know, one way or another. And by leaning in to help more people transform their lives through learning, I feel like there's still a lot of potential, still a lot of work ahead, more work ahead than we have done so far.

So Greg, sort of the same question. You know, what do you think this could look like in, call it, 10 years, 10, 15 years, but also how do you apply, you were obviously a senior executive at Amazon before you came to Coursera. What did you learn from Amazon? How could, how do you apply that to having Coursera reach its potential?

You guys, I wish we had all afternoon to talk because there's so many things I would like to ask you, but this was fantastic. Thank you so much for being here.

Thank you.

This was great. Please give them a round of applause. Thank you.