I am a Postdoctoral Associate at the Massachusetts Institute of Technology. I work under the supervision of Neil Thompson. I am interested in the economics of digital platforms, artificial intelligence, and social media. I completed my PhD in economics at the Stern School of Business at New York University under the supervision of Luís Cabral. Here is my CV. I describe my research below.
Birth, Life, and Death of AI Models
(with Aaron P. Kaye and Neil Thompson)
We characterize the lifecycles of open-weight AI models using a weekly panel of the 108,074 most-downloaded models on Hugging Face, which together account for 99.5 percent of the platform’s more than 53 billion downloads. We link these data to token usage from OpenRouter and performance rankings from Arena and Artificial Analysis. We establish key institutional details about the supply of and demand for AI models, and we organize our analysis into three phases: birth, life, and death. In the “birth” phase, we show that major developers often release models in batches, concurrently introducing models that are both vertically differentiated (e.g., by size, capability, and latency) and horizontally differentiated (e.g., by task specialization, hardware compatibility, and alignment). Because weights are open, developers can also build on one another’s models. We document cumulative innovation through fine-tuned, quantized, merged and adapted descendants of source models. In the “life” phase, we quantify how demand evolves after release. Even within the sample of top models, demand is concentrated, and the concentration persists: the median model loses ${\sim}68$ percent of its release-week downloads within ten weeks, while the most popular models decline gradually; as a result, the top percentile’s advantage over the median model is ${\sim}600$-fold at release, and ${\sim}900$-fold six months later. Further, we estimate spillover effects from new popular model releases on related models and find that the first popular third-party descendant is associated with a 76 to 99 percent increase in weekly downloads of its source model. In the “death” phase, we construct measures of technical obsolescence and usage obsolescence. We find that demand is slow to respond to a model falling behind the technical frontier, while signals of economic irrelevance shift the timing of demand. Our findings characterize open-weight AI as a market of complementary model portfolios, one in which third-party developers building on a source model can raise demand for the original, and in which demand is slow to respond to technical obsolescence.
Influencer Dynamics
Consumers use social media for entertainment and to discover new products. To reach potential customers, brands pay influencers to feature products in their content. Payment depends on the size of the influencer’s audience, and the effectiveness of the endorsement relies on trust. Excessive product recommendations may erode the relationship between an influencer and their followers. I develop a dynamic model in which an influencer produces sponsored posts (which include product recommendations) and organic posts (without product recommendations). Both affect the influencer’s growth and require effort to produce. A sponsored post incurs additional costs: it requires searching for and negotiating with brands, and followers are less likely to engage with it. Using 2,780,011 Instagram posts and 136,453 TikTok posts from 1,369 influencers, I quantify the engagement penalty by comparing identical posts across platforms. Analyzing the influencers’ career histories, I show that organic and sponsored posts have similar effects on follower growth. I leverage variation in the number and types of posts to estimate the unknown cost parameters in the model. An influencer with 100,000 followers optimally produces about 0.25 sponsored posts and two organic posts per week, and influencers with more followers produce more content of both types. Viral posts provide a transient boost to growth, but career progression largely depends on consistent content production. Regulating sponsored content can theoretically backfire by decreasing incentives to produce organic posts, but counterfactual simulations assuage this concern. A cost increase that reduces sponsored content by 25% only causes a 2.7% drop in organic content.
Feeding LLM Annotations to BERT Classifiers at Your Own Risk
(with Yucheng Lu)
Using LLM-generated labels to fine-tune smaller encoder-only models for text classification has gained popularity in various settings. While this approach may be justified in simple and low-stakes applications, we conduct empirical analysis to demonstrate how the perennial curse of training on synthetic data manifests itself in this specific setup. Compared to models trained on gold labels, we observe not only the expected performance degradation in accuracy and F1 score, but also increased instability across training runs and premature performance plateaus. These findings cast doubts on the reliability of such approaches in real-world applications. We contextualize the observed phenomena through the lens of error propagation and offer several practical mitigation strategies, including entropy-based filtering and ensemble techniques. Although these heuristics offer partial relief, they do not fully resolve the inherent risks of propagating non-random errors from LLM annotations to smaller classifiers, underscoring the need for caution when applying this workflow in high-stakes text classification tasks.