Light-responsive molecular motors and switches are molecules that can perform many useful functions in nanotechnology by absorbing light energy and converting it to other forms of energy. In the last decade, we have initiated a line of research unique to Swedish academia in which more useful and potent light-responsive molecular motors and switches are designed through computational studies in theoretical chemistry. Thereby, we have discovered a number of ways to improve the functions of such systems. In particular, in the most recent years, our research in this field has focused on understanding how different forms of friction (from solvent molecules or metallic surfaces) affect the dynamics of the photochemical reactions through which (a) molecular motors convert light energy into rotary motion, and (b) molecular switches store solar energy. From a modeling perspective, these challenging tasks have required accurate simulation of reaction dynamics in electronically excited states, combined with a simultaneous description of the solvent or surface environments in which the motors and switches are operated. In our work, the strategy to meet this challenge has been to couple semi-classical, non-adiabatic molecular dynamics (NAMD) simulations of the motors and switches based on quantum chemical methods, to classical MD simulations of the solvent or surface environments based on empirical force fields. In this way, we have been able to model the photochemical reaction dynamics of the motors and switches as they experience different forms of friction, and thereby gain very valuable insights into how their performance can be optimized. As part of this work, we have developed so-called "wrapper" codes that integrate our own in-house codes for NAMD simulations with commercial electronic-structure software packages. In this continuation project, we will now use these codes to address another outstanding challenge in the field of artificial molecular machines, which is to design so-called molecular photogears that can achieve through-space transmission of the double-bond rotary motion of light-driven molecular motors onto a remote single-bond axis. Indeed, molecular photogears with this capability are just as important for the future usefulness of artificial molecular machines as "real" gearboxes are for the functions of macroscopic machines in our everyday lifes. In this context, it is noteworthy that, as part of the previous SNIC Medium Compute project upon which the present proposal is based (SNIC 2022/5-566), our group has recently developed and published the first ever functional design of a molecular photogear of this kind (see Chem. Eur. J. 2023, https://doi.org/10.1002/chem.202303191). However, while functional, this design has ample room for improvement, both in terms of its intrinsic gearing properties (e.g., its gearing fidelity) and in terms of its performance in the presence of solvent friction. Against this background, the overall goal of the present continuation project is to improve our initial photogear design in these respects and test our corresponding predictions using the aforementioned NAMD codes as key evaluation tools.