| Abstract |
The cloud has changed the way information technology is delivered, maintained and used. A single developer can obtain hundreds of compute nodes on demand - creating radical scalability at no capital expense. Software can improve weekly, even hourly, instead of being downloaded once a year. Delivering software as a service allows the aggregation of behavior across users driving powerful recommendation engines. These technical capabilities not only enable new products but also new business models.
The trends that fueled software as as a service are now being applied to science. Start-ups such as Transcriptic provide access to fully capable labs on-demand in the cloud, again at no capital expense. Others are providing tailored services for fully reproducible computational experiments. In this talk, I provide a framework for thinking about science as a service. I overview multiple examples of services becoming available. I then talk about the implications for the role of AI in science. |