[Interview] Kunle Olukotun: Pioneering the Future of AI and Semiconductor Design

Kunle Olukotun, a Stanford University professor and co-founder of SambaNova Systems, is shaping the future of artificial intelligence (AI) and chip design. Renowned for his work in chip multiprocessor architecture, Olukotun is now leading efforts to develop next-generation chips tailored to AI computing—from data centers to edge devices. Beyond technology, he’s also concerned about pressing global issues like democracy, governance, and climate change. We met with him at the Federated Computing Research Conference (ACM FCRC 2019) in Phoenix, Arizona, for a conversation on AI, semiconductors, and the future of humanity.

Q: The Rise of Semiconductors for AI/ML

A few years ago, the semiconductor space for AI and machine learning (ML) wasn’t seeing much investment. Why the sudden shift in interest?

Kunle Olukotun: The semiconductor industry was sluggish for a while, mainly because venture capitalists struggled to see the financial returns. The long development timelines meant it took years before companies saw results. But everything changed a few years ago with the rise of Google’s Tensor Processing Units (TPUs) and Nvidia’s GPUs. These breakthroughs made substantial profits and sparked interest in alternative approaches. Looking ahead, we’re seeing Moore’s Law slow down, which has driven innovation for over 50 years. At the same time, AI and ML require immense computational power to train complex models and handle vast datasets. This demand for computation, especially for both training and inference, is fueling continued interest in domain-specific approaches for ML computing.

We’re also witnessing the emergence of companies like Graphcore and Cerebras, each taking unique approaches to AI chips.

Q: The Role of Startups in Innovation

Startups are often more agile than established companies. What advantages do they offer in the semiconductor space?

Kunle Olukotun: Innovation thrives in startups because they face less friction compared to larger companies that have to maintain existing products. In a startup, you’re free to experiment and bring fresh perspectives. It would be difficult to do that in a company with a long-standing product portfolio. Startups can take risks and try out new ideas, and in doing so, they drive the entire industry forward.

Q: Semiconductor Development in China

How do you view semiconductor advancements in China, particularly in areas like AI?

Kunle Olukotun: In some areas, such as ML for video tracking and facial recognition, China has a significant advantage due to the vast amounts of data available. However, I believe research and technological advancements are better driven in open societies. Open exchange of ideas and research is crucial for innovation.

Q: Balancing Academia and Entrepreneurship

You’re both a professor and a company co-founder. How do these roles complement each other?

Kunle Olukotun: Academia and industry are deeply interconnected. Research often sparks innovations, and companies bring those ideas to life. When you’re in academia, you can explore the theoretical side, but real-world problems that companies face lead to more applied research. Also, working with brilliant minds in both spaces is key to driving success.

Q: The Right Size for Early-Stage Semiconductor Startups

What size do you think is optimal for early-stage semiconductor startups?

Kunle Olukotun: It depends on the scope of the project. At SambaNova, for example, we’re tackling both hardware and software for new processors, which requires a larger team—over 100 people. It’s a big challenge, but that’s how you scale to meet the needs of modern computing.

Q: The Future of Fully Autonomous Driving

Is fully autonomous driving a realistic near-term goal?

Kunle Olukotun: Level 5 autonomy—completely driverless cars—is a difficult challenge. I have a Tesla myself, and while the technology is promising, human drivers still present a huge problem. If all vehicles were autonomous and obeyed the same rules, it would be easier, but human unpredictability remains a significant hurdle.

(When asked if SambaNova is developing autonomous driving chips)

No, we’re focused on developing chips for general-purpose machine learning. While our technology could be applied to areas like computer vision and natural language processing, autonomous driving is not our focus.

simplicity and cutting-edge Tech
“(ACM FCRC)Planetary cession is for giving insight. I think the tutorials( Hardware Accelerators for Training Deep Neural Networks, Using AI to Improve Architecture, Non-Traditional Workload Domains, ISCA) is a little technical, too technical.”

Q: Thoughts on the Future of AI

What’s your perspective on the future of AI and human-AI interaction?

Kunle Olukotun: The future of AI hinges on how humans and AI interact. Stanford recently launched the Human-Centered AI initiative, which aims to ensure that AI serves humanity. I believe AI is currently just a sophisticated tool. It doesn’t have agency or a soul like humans do. We’re still far from the level of creativity or reasoning that defines human intelligence. I think it will take at least 50 years before we can even begin to think about AI with human-like agency.

(*According to Stanford HAI, Fei-Fei is considered one of the pioneering researchers in the field of artificial intelligence. She previously served as the Director of the Stanford AI Lab (SAIL) from 2013 to 2018. For her Ph.D. dissertation research, Fei-Fei looked into a combination of cognitive neuroscience and computer vision…She dedicates her research to human-centered AI and is a national leading voice for advocating diversity in STEM and AI.)

Q: The Role of Semiconductor Giants

Where do you see the semiconductor giants like Intel and Samsung in the future?

Kunle Olukotun: Companies like Intel are certainly leading the way, but the underlying technology is starting to slow down. That means innovation is more important than ever. Intel has made impressive strides, and as long as they keep innovating, they’ll remain relevant. Samsung, on the other hand, faces limitations with software development. For a company to thrive in this space, a strong software and application ecosystem is just as critical as hardware.

Samsung is known for its limitations on SW.
“Certainly, the software has to be huge stuff developing their SW and application capability. But I want to know. In Korea, in terms of when people say to go to engineering school, which is it more prestigious to be a software developer or an electrical engineer? It’s might be SW engineer. I mean I think that’s right. You need more doing software.”

Q: Google and IBM have opened AI research centers in Africa, while Samsung has expanded in places like Cambridge and Toronto. What’s your take on this trend of spreading AI research globally?

Kunle Olukotun: I think it’s a great idea to develop AI talent outside of traditional hubs like the Bay Area. Personally, I’m from Nigeria, so I see the value in growing research and innovation in regions that might not be the typical focus. Google, with its deep pockets, is in a strong position to invest in these areas, and it’s good to see them supporting initiatives in Africa, like the research center in Ghana, to address pressing issues like disease and food security.

As for Samsung’s expansion in cities like Cambridge and Toronto, I believe Google has already made its mark in those regions. They’ve established strong research teams in the US and EU, so it’s natural for them to further expand their reach into new territories where they can tap into more diverse talent and challenges. The competition to push the boundaries of AI research is definitely global now, and it’s exciting to see where it will go next.

Q: Humans have created books and libraries to store information. With AI now taking over some human tasks, do you think it could ever become truly human-like?

Kunle Olukotun: Right now, AI is still just a very advanced tool. It doesn’t have the ability to make its own decisions or have its own goals like humans do. To be human, you need to have agency, a purpose, and a soul. AI hasn’t reached that level yet. Maybe in the far future, it’s possible, but it’s still a long way off—probably around 50 years.

Q: The Impact of AI on Jobs and Society

What happens to jobs and society when AI takes over more roles?

Kunle Olukotun: AI’s potential to replace jobs is one of the biggest societal challenges. Just like how the industrial revolution reduced the need for physical labor in factories, AI and automation are making certain jobs obsolete. This raises crucial questions about the future of work and how society will adapt. What will humans do when robots can perform tasks like driving, cleaning, or even teaching? It’s a problem that requires long-term thinking, especially since automation leads to greater wealth concentration in the hands of a few.

I think it depends on the point where we get to. Does it include the democratization of capability and wealth which causes concentration? As it, what happened in the industrialization? Use to, the case is that if you were in America you could graduate high school and could go into a factory for physical labor. And make a decent wage and live a decent middle-class life.

Now that’s not possible because of automation and reduction in the physical labor and then the number of jobs is drastically reduced. And so what you get is a reduction in the number of people who can actually get into the middle class by working factory-like jobs. Then you could argue that automation makes overall society more productive. But the question is did it’s an improvement for more people or just a concentration of wealth. So I think the question on robotics and AI is ‘what’s the impact’. For instance, if I don’t need lawyers anymore because all these skills are automated, that means humans no longer need to do the job. And the computer can do it more cheaply and doesn’t ever get tired. So then the question is what humans do.

In the future of work, how AI interacts with a human is a really important question. What happens if I have automated setup robots, and it fulfills all my needs, I don’t need to hire housekeepers or teachers. And all these jobs will potentially go way. I don’t need a driver. And then the question is ‘what people do?’ How will they become a productive member of society? I think that’s the question to ask. When we close to the singularity, before then, comes massive disruption. What do people do?”I think it depends on the point where we get to. Does it include the democratization of capability and wealth which causes concentration? As it, what happened in the industrialization? Use to, the case is that if you were in America you could graduate high school and could go into a factory for physical labor. And make a decent wage and live a decent middle-class life.

Now that’s not possible because of automation and reduction in the physical labor and then the number of jobs is drastically reduced. And so what you get is a reduction in the number of people who can actually get into the middle class by working factory-like jobs. Then you could argue that automation makes overall society more productive. But the question is did it’s an improvement for more people or just a concentration of wealth. So I think the question on robotics and AI is ‘what’s the impact’. For instance, if I don’t need lawyers anymore because all these skills are automated, that means humans no longer need to do the job. And the computer can do it more cheaply and doesn’t ever get tired. So then the question is what humans do.

In the future of work, how AI interacts with a human is a really important question. What happens if I have automated setup robots, and it fulfills all my needs, I don’t need to hire housekeepers or teachers. And all these jobs will potentially go way. I don’t need a driver. And then the question is ‘what people do?’ How will they become a productive member of society? I think that’s the question to ask. When we close to the singularity, before then, comes massive disruption. What do people do?”

Q. Maybe this trend is already shown in booming Youtube and entertainment.

Kunle Olukotun: Yes, but the point is you just consuming the entertainment. There has to be produced. Who’s the production? Who is paying for the product? If I say that you are all consumers, I’m generate content and charging for it, I’m making money. I got the process that giving economic value, you are just consumers. I don’t think it works in general. Youtube? User generating content but who’s making the money?

Q. Problem is governance.

Kunle Olukotun: Yes, Exactly. Any of these problems is difficult. Humans are very short-sighted, especially, even an argument could be made that this is a long-term problem, trying to solve them today, going to impact long term benefits. I think climate change is a key example of this. I can’t believe what’s happening with climate change. Politically America is not taken on this long term existential challenge to the subsistence of the human race.

Q: Addressing Climate Change

What’s your take on how governance and leadership are responding to climate change?

Kunle Olukotun: I’m deeply concerned about the political response to climate change. The failure to address it is one of the most pressing issues facing humanity today. Politicians often avoid making difficult decisions, choosing short-term gains over long-term solutions. I think we need leaders who will make the tough choices, like former President Obama tried to do, but unfortunately, this is a challenge that’s far from being solved.

Q: The Future of Governance in the Age of AI

Do you see AI disrupting the way we govern ourselves?

Kunle Olukotun: The idea of AI replacing politicians is far-fetched, but AI can certainly influence how politics work. AI algorithms have already played a role in shaping political outcomes, as we saw with social media manipulation in the 2016 election. The key question moving forward is how we use AI in governance, ensuring it serves the public good without compromising democratic values.

=================================================================

Kunle Olukotun’s work in semiconductors and AI offers a glimpse into a future driven by powerful computing technologies. Yet, as he warns, the disruptive potential of these technologies also demands careful thought about their impact on jobs, governance, and society. As we look to the future, the challenge will be harnessing AI’s capabilities while addressing the broader implications for humanity.

*Prof. Kunle Olukotun

He is Cadence Design Professor of Electrical Engineering and Computer Science at Stanford University. Pioneer of Chip Multiprocessor Designs, He is a Director of the Stanford Pervasive Parallelism Lab, developing high- level language for programming accelerators and Co-leader of the Data Analytics for What’s Next (DAWN) research program.

He co-founded SambaNova Systems in November 2017. This startup has raised more than $200 million in it’s Series A and B rounds. Samsung also invested. In April 2019, the company announced a $150M Series B funding round. He had founded Afara Websystems, acquired by Sun in 2002.