https://www.agconnect.nl/tech-en-toekomst/artificial-intelligence/liquid-neural-networks-in-ai-is-groter-niet-altijd-beter Liquid Neural Networks (liquid i.e. the nodes in a neuronal network remain flexible and adaptable after training (different from deep learning and LL models). They are also smaller. This improves explainability of its working. This reduces energy consumption (#openvraag is the energy consumption of usage a concern or rather the training? here it reduces the usage energy)

Number of nodes reduction can be orders of magnitude. Autonomous steering example talks about 4 orders of magnitude (19 versus 100k nodes)

Mainly useful for data streams like audio/video, real time data from meteo / mobility sensors. Applications in areas with limited energy (battery usage) and real time data inputs.