Quick tip for anyone new to Stability Matrix, Never update anything unless you are 100% sure of it.

A Reddit post from user FluidEngine369 is gaining traction among AI image generation beginners, offering a critical warning: never update any component within Stability Matrix unless you have a full backup and are absolutely certain of the change. Stability Matrix is a popular launcher for managing Stable Diffusion and other generative AI packages, but its update system can be a trap for the unwary. The user cautions specifically against updating core dependencies like PyTorch (torch versions), attention optimizers such as xformers, flash attention, and sage attention. These low-level libraries directly affect model performance and compatibility.

Even seemingly harmless update prompts that appear on bootup or periodic update buttons within the interface can trigger a cascade of broken dependencies. The result is often an irreversible setup that fails to generate images, requires a clean reinstall, or loses carefully configured settings and custom models. The advice is simple: if your workflow is stable and producing good results, leave everything alone. Ignore the warnings, and always back up your entire Stability Matrix folder (including models, embeddings, and settings) before touching any update. For professionals relying on consistent output, this is a critical operational best practice.