HP Innovation Journal Issue 12: Summer 2019 | Page 14

DRIVING EFFICIENCY THROUGH TECHNOLOGY While megatrends help us better anticipate global shifts, resource constraints, and customer needs, it’s disruptive technol- ogies that enable us to innovate and reshape our future. Digital manufacturing, including 3D printing, for example, can help to reduce waste and energy emissions in rapidly growing urban areas, where transporting materials and waste is both costly and inefficient. It also has the potential to make a significant impact on energy consumption. Traditional manufacturing consumes about one-third of worldwide energy production. 9 If you apply digital manufacturing, including 3D printing, to the full life cycle of manufacturing—design, transportation, production, inventory, etc.—you have a chance to substan- tially reduce that energy use. The last technology trend we feel will drive efficiency is the concept of virtual machines or digital twins. When machines can learn and respond to the data they sense and capture, it becomes possible to create virtual models of the machines. If we can do our development and testing on the digital twin, and then final deployment on the actual machine, we complete the entire process faster and more efficiently. We can take this a step further and hook these virtual machines together to optimize complex physical systems and processes before any setup takes place. Compa- nies are already starting to deploy this technology and it will only become more useful as more data becomes available and models become better at simulating the real world. F See “Edge of Computing/Energy Efficiency” article in this issue to learn more. P.62 Edge computing advancements in silicon allow for data pro- cessing with AI and machine learning inference to happen locally, erasing the need for data transmission and the higher energy use that transmission requires. Those energy-efficient compute architectures are also chang- ing the nature of software development. Instead of spending hours on coding, a software engineer could curate data for a task or series of tasks, build a model, and then deploy it. For example, instead of an HP software engineer writing firm- ware for a 3D printer, they would collect the data that comes from the thousands of printer sensors and actuators—which nozzle gets fired, loader gets turned, heat sensor reading, etc. They would then send that data to the new machine learning chip. The machine learning chip learns the data model, and then it is run as an inference on the printer. All the engineer did was curate the data. This leads to fewer bugs, lighter- weight code, and a more efficient code base. The industry name for this type of development is Software 2.0. It’s what Tesla uses to deliver autonomous driving. 9. The Outlook for Energy: A View to 2040,” ExxonMobil 2018; Leendert A. et al; Runz, H. et al 12 HP Innovation Journal Issue 12