The idea of digital twin is not new. NASA used it first in the 1960s for its space program. It got revived in the early 2000s mainly in manufacturing and automotive industry. The idea was to create a digital replica of the product and the process that was synchronised with real world as per the need. The key characteristics of the real system being twinned were well-defined, well-bounded, and controlled by physical laws. The principal objective of creating digital twin was to ensure product and process quality. Since the past 10 years, adoption of digital twins has spread even to socio-techno systems characterised by emergent behaviour, and so has the spread of use cases. Today, technology is available, though with wide variance of effectiveness, to address digital twin life cycle. Recent advance in AI and especially Gen AI has helped reduce the cognitive burden too. However, many challenges remain, e.g. robust method, effective repurposing of digital twin to a different context, or transferring the learning courtesy domain-specific purposive digital twins to other domains for the same or similar purpose. All in all, these are exciting times to be working in digital twin space.