Abstract
Generative AI is fundamentally changing how we develop software systems. As AI becomes capable of writing most, if not all, of the code, it raises a fundamental question: How should we engineer complex software systems when humans are no longer the primary authors of code?
Two years ago, we started the G5 project at MIT to explore this question and rethink software engineering from the ground up. Our work has led to three key pillars. First, software engineering is becoming knowledge engineering, shifting the focus from writing code to capturing system intent and semantics. Second, ALIGN introduces a new AI-centric software development lifecycle designed for AI-driven coding. Third, formal verification at scale ensures that generated software faithfully implements its intended behavior.
In this talk, I will present these three pillars and show how we have already used them to replicate a large commercial CRM system and uncover significant bugs in existing software, including PostgreSQL. These results point toward a future where engineering complex systems is less about writing code and more about specifying and verifying intent.
Recording
Note:
The recording is password protected. If you don't have the password, ask us.
https://mit.zoom.us/rec/share/HjLfXlhouxfnnApCLLBaJv9TNwtoUHrhuxpG6SNlJiviTAN3lZzHaix9hsYP3vl6.y_20WWqiuFQ_12l7
Bio
Tim Kraska is a professor of Electrical Engineering and Computer Science (EECS) in MIT's Computer Science and Artificial Intelligence Laboratory (CSAIL), co-director of MIT’s Generative AI Impact Consortium (MGAIC), the Data Systems and AI Lab (DSAIL@CSAIL), and the Everest@CSAIL initiative, and was a co-founder of Instancio, Einblick Analytics (both acquired), and now G5 Labs. His research focuses on using ML/GenAI for data systems. Before joining MIT, Tim was an Assistant Professor at Brown University and spent time at Google Brain. He also served as a Director of Applied Science at Amazon Web Services (AWS). Tim is a 2017 Alfred P. Sloan Research Fellow in computer science and has received several awards, including the VLDB Early Career Research Contribution Award, the Intel Outstanding Researcher Award, the VMware Systems Research Award, the university-wide Early Career Research Achievement Award at Brown University, an NSF CAREER Award, as well as several best paper and demo awards at VLDB, SIGMOD, and ICDE.