Actuarial Modelling at a Crossroads: GPUs, AI, Open Source and the Future of Code

This session will look at some of the important choices facing actuarial modelling today, from a vendor's point of view, explored in four themes:  GPUs vs CPUs: the search for X thousand times faster. We will outlining the technical differences between the two, how these affect model expressions and performance, and the commercial implications, before sharing key lessons learned from our own experience.  AI for actuarial modelling: from PDFs to Models: We will explore how artificial intelligence can transform actuarial work, with some practical examples, outlining benefits and risks. Open Source: to buy or to build? We will weigh up the advantages and drawbacks of proprietary versus open-source solutions, while highlighting the wider costs of software development and maintenance that are often hidden.  Low Code vs Verbose Code: Is less more or is more more? We will revisit the history of actuarial coding from mainframes to low code and modern vibe-coding, and considering the pros, cons and preferences that shape how actuaries build and maintain models today.