A Chinese research team has claimed to simulate a dynamic similar to that of a human brain in a machine. This simulation does not recreate an actual living brain that changes over time with experiences and emotions, but rather captures a fixed structure. The researchers suggest that their approach challenges traditional views of how brain simulations are understood. Unlike the nematode worm, whose entire neural connections (called a connectome) can be precisely reconstructed, the human brain remains too complex and data-intensive to model with the same accuracy. To address this challenge, the team used a combination of electron microscopy, mesoscopic statistics, and MRI scans to create hybrid models that work across different scales. However, these models do not support continuous updates, interaction with the environment, or physical embodiment. The concept of a digital twin of the brain (DTB) is based on the idea of recreating the structure and function of a specific human brain in a computer. Unlike generic models, the DTB aims to reflect the unique characteristics of an individual brain. However, current models fall short in this regard. They capture general brain functions well but fail to represent the uniqueness of a real person’s brain. More importantly, these models remain static once created, unlike living brains that are constantly changing. It's like taking a single photograph of a person and claiming it captures their entire life and future choices, which is an incomplete and static representation. One might assume that modeling more neurons makes the task harder, but in reality, the challenge lies in the resolution and availability of data. For example, the same method used to map the nervous system of a nematode worm, which has only a few hundred neurons, yields highly accurate results. However, when applied to the human brain, it only provides a rough sketch. The reason is that mapping the entire human connectome — all the connections between neurons — is not feasible with current technology due to the brain’s complexity and density. To overcome this limitation, the researchers proposed an innovative approach using hybrid multi-scale scaffolds. This method combines three distinct data sources: electron microscopy for precise details on small brain regions, mesoscopic statistics for generalizing repetitive structures, and MRI for an overall view of the brain. This strategy avoids the need for an exhaustive reconstruction and instead focuses on structural-function relationships. It also shifts the focus from creating an autonomous digital brain to studying the co-evolution between the real brain and its digital copy in a controlled environment. Despite these advances, the digital twins of the brain are still partial simulations. They can reconstruct a brain's structure and simulate its dynamic behavior, which is a significant achievement. However, three major goals remain unattained in the near future: the ability to continuously update the model as the brain evolves, interaction with the environment in a closed-loop system, and physical embodiment — the ability to act in the real world. These challenges are both technical and scientific, involving issues like identifying what makes a brain unique, validating models, and managing personal neural data. The potential applications are vast, including healthcare, neuroscience, and artificial intelligence. While the progress is encouraging, the research highlights that the human brain still holds many secrets. Machines are getting better at drawing its contours, but they have yet to capture its constant, living evolution. Whether this natural advantage will last remains to be seen.