What I Learned Being a Workaholic

Thank you very much.

I have been working on artificial intelligence for 36 years, ever since I was two years old.

Actually, I remember the beginning vividly because I still have the application letter I wrote for my PhD program at Carnegie Mellon. I wanted to work in artificial intelligence because I believed it would become a technology capable of changing the world, something to which I could devote my life.

Once I arrived at Carnegie Mellon, I began living as a workaholic.

I loved the work and wanted to make a difference. I believed I was in the world to have the greatest impact possible, so I measured everything by efficiency.

At the time, computers were slow. I had to check my experimental results every two or three hours, so I trained myself to wake up at two and five every morning without setting an alarm. I could keep my experiments running and start the next one based on the previous results.

That habit continued when I worked at Microsoft and Google China. I woke up at two and five to respond to emails from my bosses in California and Washington. I wanted them to see that I was working hard and responding quickly.

My employees called me Iron Man because I answered every email immediately during the day and within two hours at night.

I was proud of that. It also made them feel that they had to work just as hard.

Since I was already waking up at two and five, I thought, why run to the den to use the computer? I had people build a small rack that allowed me to answer emails with a keyboard and mouse resting on my stomach while I was still in bed.

That was the life I lived.

Things became even more intense when I moved to China.

Silicon Valley is known for long hours, but the working culture in China was even more extreme. When I became a venture capitalist, I was surrounded by, and invested in, entrepreneurs who worked incredible hours.

I once persuaded one of our companies that it needed a little more work-life balance. Its new recruiting slogan became, “Come work for us. We are only 996.”

That meant working from nine in the morning until nine at night, six days a week.

Unfortunately, the company was later beaten by competitors working 997.

I worked as hard as the entrepreneurs I supported, and we made tremendous progress. I was proud of what I had accomplished at Apple, Microsoft, Google, and later at Sinovation Ventures.

We became proud investors in five artificial intelligence companies valued at more than a billion dollars each. I felt that the goals I had set for myself were being achieved and that AI had finally matured.

I will not go into the technical details of artificial intelligence today. Think of it this way: When AI is given a narrowly defined task and an enormous amount of data, it can make decisions at a superhuman level.

We have seen this with AlphaGo and IBM Watson. We see new applications every week in banking, Amazon, Facebook, Google, and many other companies.

AI is driving these applications, generating enormous amounts of money and making a significant difference.

But that created a new quandary for me.

When AI becomes so good at performing individual tasks, what happens to the people currently doing those jobs?

What happens to loan officers, customer service representatives, telemarketers, fruit pickers, agricultural workers, assembly-line employees, and dishwashers?

Many of those jobs will be displaced.

I felt proud that our work was making such a difference, but I also worried that we were destroying jobs.

Even worse than the loss of jobs and income could be the loss of meaning.

The Industrial Revolution taught many of us to believe that work is the meaning of life.

Before industrialization, a craftsperson might build an entire product. The assembly line replaced that work with many people repeatedly performing individual tasks. It was useful for industrialists to persuade workers that fulfillment came through employment.

For thousands of years, the meaning of life was not so completely dominated by work. That changed during the Industrial Revolution.

If AI displaces routine jobs, the resulting loss of meaning could be disastrous.

As you can tell from the stories I have shared, I was a willing victim of this way of thinking. I became maniacally focused on producing results. I developed terrible habits related to sleep and stress.

I was an obsessive workaholic.

That obsession ended abruptly about five years ago, when I was diagnosed with stage IV lymphoma.

My PET scan showed approximately 20 tumors lighting up like fireballs. They melted away my confidence and the pride I had placed in my work.

The diagnosis forced me to reconsider the meaning of my life.

I went through denial, anger, bargaining, depression, and eventually acceptance. I began chemotherapy, but when I finally reached acceptance, I also reflected on the way I had lived.

I realized that my priorities were completely out of order.

In my pursuit of excellence at work, I had neglected my family.

My father had died before I had the chance to tell him how much I loved him. My mother had dementia and no longer recognized me. My children had grown up.

Looking back, I understood that my priorities and workaholism had been a grave mistake.

During my illness, I read a wonderful book by Bronnie Ware about caring for people at the end of life. She had accompanied thousands of people as they were dying and documented their greatest regrets.

At the top of the list was not spending enough time with loved ones. Another was working too hard.

It felt as though the book were speaking directly to me about the errors of my life.

Fortunately, chemotherapy worked. I am able to be here with you at End Well because my life did not end, and I am well.

During my illness, one of my most important moments of awakening came through a conversation with a well-known Buddhist monk in Taiwan.

He asked me, “Why did you live your life the way you did?”

I answered, “To make a difference. I measure everything by how much of a difference I make.”

He responded, “Changing the world and making a difference may simply be cloaks hiding your vanity. Perhaps you are trying to make yourself famous.”

Human beings struggle to resist greed, fame, and vanity.

He told me that if I truly wanted to change the world, I should stop thinking about how to make myself great. Instead, I should ask whether I could offer a little more love.

“At the very least,” he said, “give love to the people who have loved you. And if you can do more, give love unconditionally.”

That has become the way I choose to live.

When I faced the possibility that I had only months remaining, my family gave me comfort. Their love was as important and effective as chemotherapy.

Looking back, I realized that their love had always been present. I simply had not given them the attention they deserved.

I changed my life.

I moved back to be close to my mother. My wife now travels with me wherever I go. I schedule vacations around the time my children have available, rather than expecting them to follow my schedule.

These changes have made an enormous difference in my life.

They have also changed how I think about artificial intelligence.

Is AI good or bad? What happens if it displaces all our jobs?

The answer is that it will not displace every job.

There are important things AI cannot do.

AI cannot truly create. We give it a goal, and it optimizes toward that goal.

More importantly, AI cannot love.

Despite what science fiction and movies may lead you to believe, AI possesses no love, compassion, or empathy.

In the movie Her, it was Scarlett Johansson’s voice that made us believe an artificial intelligence could love.

When AlphaGo defeated Ke Jie, we saw Ke Jie cry because he loved the game. AlphaGo felt no emotion. It experienced no joy or happiness and had no desire to embrace someone it loved.

That difference points toward the future of work.

Imagine a chart with creativity on one axis and compassion, empathy, human touch, and trust on the other.

The jobs requiring little creativity or compassion are the ones most likely to be replaced by AI.

But there are also wonderful ways for humans and artificial intelligence to work together.

Consider the physician of the future. AI might serve as the analytical engine that helps make a diagnosis. The doctor could then become the human connection to the patient, providing warmth, care, and reassurance.

Doctors could visit people in their homes and give them greater comfort and confidence throughout treatment. We know that human connection can improve how people respond to care.

That is one kind of symbiosis.

In work that requires creativity but less human connection, AI can become a powerful tool. It might help scientists invent new drugs and develop better treatments that allow people to recover and live longer.

We should also celebrate the work that requires both creativity and compassion.

The future does not have to be bleak. AI will not simply take every job. There will be opportunities to celebrate our humanity and for people and machines to work together.

Still, the next 10 or 15 years will be difficult as many routine jobs disappear.

The number of people currently doing routine work is enormous, while the number of highly creative jobs is relatively small.

But work centered on compassion can help bridge that gap.

We will need more caregivers, nurses, doctors, teachers, concierges, massage therapists, and mental health professionals.

These workers can become the human connectors in our society. With the right training, people displaced from routine jobs may be able to move into work centered on caring for others.

The United States is projected to need 2.3 million additional healthcare workers, including nurses, home-care workers, and elder-care professionals.

As people live longer, we will need more people to care for those over 80, who require far more assistance than people between 60 and 80.

Those jobs are growing.

But we face a challenge. Routine jobs such as truck driving and operating heavy machinery often pay considerably more than personal-care work.

As a society, we need to increase the compensation and social status of caregiving jobs. People facing displacement must be able to believe that retraining will lead to a dignified and sustainable future.

Companies should help provide that training. Amazon, for example, has offered substantial annual reimbursement for employees retraining for careers such as nursing.

Some people may ask, “You are an AI scientist. Don’t you think robots will eventually care for older adults?”

I will tell you one final story that explains why I do not.

An entrepreneur once showed me a robot designed to care for older people. It was a remarkable machine.

It could wake people at the appropriate time, remind them to take medication, order food, call a doctor, and provide entertainment.

But the entrepreneur came to me and said, “I have a problem. The older adults are using only one function: customer service.”

They would press the customer-service button, and a real person would appear on a video screen.

Then they would say, “Why hasn’t my daughter called today?” or “Let me show you a picture of my granddaughter.”

People do not want robots to care for them.

We need to do far more to make compassionate work possible. That is what people want, and machines do not have the capacity to provide it.

Even as robots improve, many people will still prefer human connection.

My experience of facing death, combined with my understanding of artificial intelligence, has led me to see AI as a kind of serendipity.

AI is here to remind us that work is not the meaning of life. We exist for reasons beyond our jobs.

Artificial intelligence is a tool. It can take over routine labor and potentially liberate us from it.

That freedom could allow us to do what we love, spend more time with the people we love, and discover why we exist as human beings.

Thank you.