diff --git a/src/tex/ms.tex b/src/tex/ms.tex
index d342ed9..f9f296d 100644
--- a/src/tex/ms.tex
+++ b/src/tex/ms.tex
@@ -35,7 +35,7 @@ \section{Introduction}
 Aggressive scaling of these models has led to a rapid increase in their capabilities,\autocite{brown2020language,zhong2024benchmarking} with the leading models now being able to pass in some evaluations the United States Medical Licensing Examination.\autocite{kung2023performance} 
 They also have been shown to design and autonomously perform chemical reactions when augmented with external tools such as web search and synthesis planners.\autocite{openai2024gpt4, Boiko_2023, bran2023chemcrow}
 While some see \enquote{sparks of \gls{agi}} in them,\autocite{bubeck2023sparks} others consider them as \enquote{stochastic parrots}---i.e., systems that only regurgitate what they have been trained on.\autocite{bender2021dangers}
-Nevertheless, the promise of these models is that they have shown the ability to solve a wide variety of tasks they have not been explicitly trained on.\autocite{bommasani2021opportunities, anderljung2023frontier}
+Nevertheless, the promise of these models is that they have shown the ability to solve a wide variety of tasks they have not been explicitly trained on.\autocite{bommasani2021opportunities, anderljung2023frontier, ai4science2023impact}
 This has led to tremendous economic interest and investment in such generative models, with an expected market of more than \$1.3 trillion (almost \$30 billion for drug discovery applications) by 2032.\autocite{bloomberg}
 
 Chemists and materials scientists have quickly caught on to the mounting attention given to \glspl{llm}, with some voices even suggesting that \enquote{the future of chemistry is language}.\autocite{White_2023}