In the race towards creating increasingly larger neural networks in the AI landscape, MIT's Computer Science and Artificial Intelligence Laboratory (CSAIL) offers a compact and efficient solution, named Liquid neural networks (LNN). These deep learning models are quite promising in areas where traditional deep learning models struggle, like robotics and autonomous vehicles. Given the limited computational power and storage space in such systems, LNNs were developed to address these inherent constraints. Inspired by the research on biological neurons present in small organisms, the CSAIL team aimed at devising neural networks that maintain accuracy while being computation efficient as well. Liquid Neural Networks greatly differ from traditional deep learning models and hence, represent a significant breakthrough in this field. The efficiency of LNNs lies in their use of dynamically adjustable differential equations, which allows them to adapt to new situations after the training phase, leveraging a capability not found in typical neural networks. LNNs also possess a unique wiring architecture that allows lateral and recurrent connections within the same layer. This compactness of LNNs enables them to run on small computers found on robots and other devices while maintaining interpretability of the network. They are also capable of handling continuous data streams, making them suitable for safety-critical and computationally constrained applications such as robotics and autonomous vehicles.