Balancing Tech, Economics, and Humanity: Lessons from Amit Gandhi

Senior Economic and Technical Advisor at Airbnb and Senior Fellow at 美国宾夕法尼亚大学 - 沃顿商学院 , Amit is an applied economist renowned for groundbreaking work in industrial organization and econometrics. From shaping economic methodologies to leadership roles at major tech companies (Tech Fellow at Airbnb and Chief Economist at Microsoft), he has been a driving force in tech and education.

Kimberly: Amit, your educational journey has been impressive. What inspired you to pursue this path, and how have your academic pursuits influenced your professional trajectory?

Amit: I am an academic at heart, and somebody taught me that you can't change your type. I became an academic and specifically got a PhD in economics. I did add the MBA with the notion that one day, this would be useful to the business world. Being an academic is a dedication to a class of questions and a process and an approach that really sees things through to a longer-term set of insights, ideas, papers, and artifacts that take an intellectual form. Fundamentally, business is about results and outcomes, and the pace is much faster.?

I think for my arc, once my academic journey felt like it had reached a certain point, I started to look at what’s next: how do I have an impact with these ideas, and what can I do with them beyond teaching and scholarship? I think every academic has this existential life, midlife moment where you start to wonder, well, how do we apply it? For me, technology, business, and data are such a hand-in-glove fit to economics, but what I still find fascinating about the job is that it’s sort of marrying those two cultures. So, how do you get them to interoperate? It's very clear to me that businesses need the kind of social-scientific thinking that economics or deep business degrees would offer.

Kimberly: You've worked with diverse teams of researchers and people in the corporate setting and across various environments. How do you approach collaboration between individuals with diverse backgrounds, and are there any lessons that you could share with us?

Amit: Maybe the most important question is: how do you cross-functionally collaborate? I think the number one thing I had to learn, which was maybe the biggest step going from academia to the business world, is that in academia, you are used to listening to the sound of your own voice. You're paid to be on stage teaching with spotlights on you all the time. It’s really important as you transition to working in teams to flip that – now you are the student, and you are listening. Everyone has something big in their head that they know deeply better than any person in the world. If you can extract that nugget of gold from someone, you can extract that nugget of gold from someone else, and then we can synthesize it. Then, you're sitting on a jackpot. A lot of it is learning how to become a master listener. How to actively listen is maybe the most important soft skill an academic needs to understand in moving into the business world.?

I have also found that people in businesses are curious beyond belief. They want to understand, learn, and go deeper. Make sure your role is clear in that dynamic and that you're not trying to usurp anyone or grab their knowledge to use for your own nefarious purpose. Try to add value to something they already understand. The critical thing to teamwork is to recognize that when you walk into a team environment, people are doing things already, they have jobs, and the system is moving. You are a new exotic fruit being added to an environment and must determine how to add value.?

Kimberly: As the Senior Economic and Technical Advisor at Airbnb , you're in a unique position to drive innovation. How do you balance the technological advancements with the human-centered aspects of the platform??

Amit: Airbnb, as a product, as a business, and as an idea, never ceases to amaze me. When I moved from Microsoft to Airbnb, a big difference was that, at its core, Microsoft is a software product. It's supposed to be a great software product, and economics certainly play a role and help make sense of what the software is doing. However, at some level, Microsoft is a product even without economics. Airbnb, and more broadly generalized marketplace businesses that are digital platforms, are bringing various different sides of a market together to create an exchange. There is a lot of technology involved. While there's immense data and sophistication in the technology, at its core, it's not a computer science problem – it's a human problem. It's a human, social, economic, anthropological problem.?

How do you blend that data, AI, and machine learning with those very human constructs? Understanding humans, trusting each other, building businesses as entrepreneurs, being explorers of the world, and engaging in new experiences. I think Airbnb is in a unique position to look at those things side by side. Travel and experiences on Airbnb are an important part of someone's life. What we're going to have to start teaching in MBA school soon is how to merge AI and the motivations and behaviors of people.?

Kimberly: What do you envision as the next frontier in AI and data science? How do you think it will impact the industry?

Amit: 2023 was a year I don't think anyone could have expected. I remember when ChatGPT dropped. It's interesting because it wasn't a fundamental change in the sense that it had been around, and this class of technology and models, these so-called transformer models, had existed, and there were chatbots, but there was this excitement. I'm torn about whether this is really AI. This is clearly powerful technology. It's almost like we’ve been working with numbers for so long in data science, and we put numbers in models, and models spit out other numbers. But a lot of what's happening here basically says, what if we just replace numbers with natural language? If you replace numbers with natural language, then it's the same sort of game.?

It's like suddenly we can use voice, text, and images, and these are the things that we as humans consume all day. Therefore, our relationship is sort of bringing in that sense that it feels AI-ish because our relationship to the way humans experience the world is now similar to how these models we build are experiencing the world.?

On the other hand, I'm not sure that this is AI, but I'm happy to be proven wrong in that the grand hypothesis here is that if you can predict text well enough, you must understand what you're talking about. It’s a bit of a philosophical question, and I think the problem is if the technology doesn't understand what it's saying, then you can never really control the hallucinations. This is an incredibly useful tool that will change jobs and the way we work. I'm happy to call that AI a placeholder for this exciting new technology. Can the hallucination risks of it be managed? I don't know. That's the part where we just need to kind of wait and see.

Listen to the full episode on Apple Podcasts, Spotify or YouTube.

Subscribe to The Executive Room newsletter, and don’t miss out on future editions.