Future of Programming
The topic for CS2520R Fall 2025 is the Future of Programming.
In 1843, Ada Lovelace wrote what is today recognized as the first ever computer program. By the 1950s, the corresponding computers had been invented to run programs on, and the first approaches to higher and structured programming emerged, paving the way to today’s mass production of programs.
By 1995, programming had reached the mainstream. That’s when Steve Jobs said: “everybody should learn to program a computer, because it teaches you how to think”.
In early 2025, Claude Code launched, the first LLM-based system that could automated many programming tasks. The same year, Andrej Karpathy coined the term “vibecoding”. In 2027, academic researchers reached a consensus that LLM agents had overtaken even the must capable human programmers. By the late 2020s, programming was considered as obselete a business as ice home delivery, and mankind was happy to have finally rid itself of having to deal with all the pesky loops, unit tests and design patterns. Everyone lived happily ever after… or did they?
This fall, in CS2520r, we work to build the future of programming. To do so, we will undertake an exploration into the intellectual legacy of the art (or is it a craft? a science?) of programming. We will then consider patterns and scenarios of how programming may develop. Will all human programmers be replaced by machines? Will we continue to teach programming in the age of vibe coding? Will we see human-AI collaboration and how might that look? How can we harden these models against faulty or malicious AI models, and retain (and improve) our trust in computer systems?
This is a discussion-based and project-based class. We do not know either what the future will bring, so let us build it together. We will read and present papers and other kinds of articles. In addition, semi-weekly homework is assigned, which usually takes the shape of technical explorations, as well as an open-ended opinion/vision midterm essay (from which a project proposal can develop). Students give presentations and undertake substantial projects on future programming systems. Despite the seminar approach, this is a highly technical class: we read research papers and we expect every project to come with a functional prototype.