Bay Area Tech Economics Seminar

Intended audience: Graduate-trained economists and other quantitative social scientists from academia, policy, or the private sector who are interested in the tech economy and methods for analyzing it, and in making connections with others in this broadly-defined space.

Upcoming Seminars

Thursday, March 27, 6:30-8:30 PM

University of San Francisco

The Economics of Virtual Economies: Roblox

Speaker: Steve McBride, Head of Economy and Ads Science at Roblox

Roblox is a leading virtual economy designed to connect game players, developers, shoppers, and UGC creators across community-generated experiences (games). Roblox advances this mission by architecting and enabling a diverse, competitive, and participatory virtual economy using its virtual currency Roblox. In this talk you will learn how Roblox uses economic science to achieve these goals in addition to economic efficiency across initiatives in four product areas: 1) a virtual cosmetic goods marketplace; 2) game economy optimization; 3) ads measurement in 3D immersive worlds.; and 4) in-game currency design The emphasis is on the use of science, economic theory, and experimental methods to advance knowledge and influence product development.

Thursday, April 24, 6:30-8:30 PM

University of San Francisco

Hate the Game: Economics Cheat Codes for Life, Love, and Work

Speaker: Daryl Fairweather, Chief Economist at Redfin

Thursday, May 8, 6:30-8:30

Stanford

Speaker: Nathan Kallus, Cornell (Associate Professor) and Netflix (Research Director, Machine Learning and Inference Research)

Signup link coming soon.

Ground Rules

Seminars will operate in an open economics-style format, with comments and discussion from the audience welcomed throughout.
To facilitate open discussion, comments from anyone other than the presenter should be treated as “not for attribution.” That is, you are free to recount what was discussed, but you should not publicly identify any commenter or their organizational affiliation without permission.
Events will not be live-streamed or recorded.

Doors will open at 6PM, with the talk beginning at 7PM. Before the start of the presentation, there will be a networking reception with light refreshments to facilitate ongoing conversations. Location details for each session will be provided to registered participants.

Organizers

Stacy Carlson, Tom Cunningham, Andrew Hobbs, Guido Imbens, Peter Lorentzen, Rose Tan, Sean Taylor.

If you or your organization might like to sponsor or host this event in the future, please get in touch with any of the organizers.

Our Sponsors

Related groups and event series to follow

Open All

Groups to join

Remote/online events

Stanford's Online Causal Inference Seminar Tuesdays 8:30 AM.
SF Fed Emerging Tech Economics Research Network

In-person events

USF Economics Department Seminar (Thursdays 3PM, all are welcome)

USF Data Science Seminar series (Fridays 12:30 PM, 101 Howard Street)

Past Events

2024

Tuesday, March 18, 6:30-8:30 PM

University of San Francisco

America’s Legal Gambit to Curb China’s Technological Rise

Speaker: Angela Zhang, Professor of Law, University of Southern California

Sunday, January 16, 6:30-8:30 PM

Stanford University

Reflections of an Economist Who Manages People

Speaker: Jonathan Hall, Uber

Thursday, December 5, 6:30-8:30 PM

University of San Francisco

Does Sponsored Search Advertising Augment Organic Search? Evidence from an E-commerce Platform.

Speaker: Sarah Moshary, UC Berkeley

Thursday, November 7, 6:30-8:30 PM

Stanford University

GDP-B: Accounting for the Value of New and Free Goods in the Digital Economy

Speaker: Erik Brynjolfsson

Thursday May 9, 6:30-8:30 PM

Stanford University

The Interplay of User Behavior and Algorithm Design in Digital Ecosystems

Speaker: Hannah Li, Columbia University

The interactions between user behavior and algorithms employed on online platforms present challenges that conventional data science tools like A/B testing and recommender systems often overlook. For example, recommender systems can create feedback loops that affect both user and content creator retention. Moreover, users may strategically engage with or avoid certain content to shape their future recommendations. This talk highlights several ways these dynamics can influence platform learning and outcomes and suggests modifications to mitigate these challenges.

Thursday March 28, 6:30-8:30 PM,

University of San Francisco

The Impact of Generative AI on Jobs and Skills

Speaker: Karin Kimbrough, Chief Economist, LinkedIn

Generative AI (GAI) is beginning to shape the world of work. Our data of 1B members worldwide allows us to track this transformative journey of enthusiasm for these technologies, through adoption of GAI skills by professionals and a 70% increase in demand by companies for specific GAI skills. With skills at the forefront, we examine our data of 1B members for where progress is being made and where challenges are still evident.

Thursday February 8, 6:30-8:30 PM, Stanford University

Is Generative AI Disrupting the Digital Economy? Early Evidence from Card Spending Data Suggests Not

Speaker: Kenneth Wilbur, Professor of Marketing and Analytics, UC San Diego

Generative artificial intelligence raises concern about human jobs, but what about other products and services? If customers “hire” products to “do jobs,” is generative AI threatening the services that perform those jobs? We investigate a large card spending panel to understand how early ChatGPT-4 adopters changed their spending on other digital services. We use later cohorts of ChatGPT-4 adopters to predict early adopters’ counterfactual spending, and apply a triple-difference identification strategy with Coarsened Exact Matching. We find that ChatGPT-4 adoption increased consumer spending on other AI products. Our estimates rule out market share changes of 1% for the large majority of brands, with a few exceptions. We will present more results during the talk.

2023

Thursday October 12, 6:30-8:30 PM, University of San Francisco

Treatment Effects in Market Equilibrium

Speaker: Stefan Wager, Stanford University

When running randomized trials in a marketplace where prices equilibrate supply and demand, one needs to account for spillovers due to price effects. I'll show how to capture -- and correct for -- such spillovers within the Neyman-Rubin potential outcomes model for causal inference. I'll also discuss methods for spillover-aware optimal targeting.

Thursday September 28, 6:30-8:30 PM,

University of San Francisco

The U.S. 2020 Facebook/Instagram Election Study

Speaker: Matthew Gentzkow

Overview: This project is a novel academic-private sector collaboration designed to study the impact of Facebook and Instagram on key political attitudes and behaviors during the U.S. 2020 elections. Key outcomes include (1) dis/mis/information, knowledge, and (mis)perceptions; (2) political polarization; (3) political participation, both online and offline, including vote choice and turnout; and (4) attitudes and beliefs about democratic norms and the legitimacy of democratic institutions. We will discuss results from recently published papers as well as not-yet-published findings.

Thursday June 1, 7-8:15 PM

Stanford University

Economics of Generative AI

Panelists: Erik Brynjolfsson, Tyna Eloundou, Katya Klinova

Moderator: Noah Smith

A 2 hour interactive panel discussing various economic aspects of generative AI: effects on productivity, employment, and inequality; effects on competition; effects on communication and entertainment; and appropriate norms and regulation.

Thursday May 18, 7-8:15 PM,

Stanford University

Chinese Social Media and Government Influence

Speaker: Jennifer Pan (Stanford)

With 900 million social media users, China's market for social media is larger than that of any other country, and also dramatically different from all other countries. No Facebook. No YouTube, No Instagram. China's market for social media is dominated by domestic Chinese companies. This dominance of Chinese platforms has allowed the Chinese government to influence the digital information environment to an unparalleled extent. This talk will lay out key features of the Chinese social media landscape and discuss what we know about how the Chinese government intervenes in it.

Thursday April 20, 7-8:15 PM,

Google San Francisco

Speaking on data's behalf: A Bayesian framework for business decision making

Speaker: Ignacio Martinez Staff Economist & Manager, The Chief Economist's Team at Google

Abstract: In this talk I will discuss why a Bayesian framework can be very useful to inform business decisions. First, I will discuss the types of questions that business leaders may want to answer using data. Then, I will argue that the traditional frequentist framework cannot answer these questions directly and that this failure in the traditional analytic approach is a big problem and can result in the wrong decisions being taken. Finally, I will outline a Bayesian alternative to answer these questions. I will argue that we should not shy away from using informative priors, bound ourselves to how we would present results in a frequentist way, and that we should encourage decision-makers to think in bets.

Thursday April 13, 7-8:15 PM,

University of San Francisco

Big Tech: The Promise and Peril of Regulation

Speaker: Carl Shapiro (UC Berkeley)

Over the past several years, we have witnessed a vigorous push to regulate large digital platforms, notably Amazon, Apple, Facebook and Google. In the European Union, the Digital Markets Act, a bold new regulatory regime, will soon come into force, overseen and enforced by the European Commission. In the U.K, the Competition and Markets Authority is moving ahead with plans to regulate large digital platforms. In the United States, Congress has been considering legislation to regulate these same firms, with encouragement from the Biden Administration.

Professor Shapiro will discuss the promise and peril of regulating large digital platforms. He will first comment on the track record of antitrust enforcement in this area, assessing claims that antitrust enforcement relating to digital platforms has been inadequate. He will then consider the pros and cons of sector-specific regulation, drawing lessons from our experience regulating other industries subject to rapid technological change.

Thursday March 2, 7-8:15pm,

University of San Francisco

Adaptive Experimentation At Meta

Speaker: Qing Feng (Meta)

A/B tests are the gold-standard method for internet firms to evaluate the performance of system changes, yet these experiments are generally limited to evaluating the effects of only one or two variants. In many cases, however, we are interested in evaluating the effects of thousands or a potentially infinite number of possible interventions, such as treatments parametrized by continuous variables. Adaptive experimentation (e.g., Bayesian optimization, bandit optimization) is the machine-learning guided process of iteratively exploring a large action space in order to identify optimal configurations in a resource-efficient manner. In this talk, I will first give an overview of adaptive experimentation at Meta and show how this ML-assisted process allows experimenters to explore more effectively and intelligently. Furthermore, I will discuss our recent advancements in methodology, including the handling of multiple objectives, noisy and non-stationary measurements, and data from different experimentation modalities, and explain how these developments can further improve the efficiency and effectiveness of experimentation.

Related open source libraries:

Adaptive Experimentation Platform

Bayesian Optimization in Pytorch