​用AI赋能产出“AI应用的未来指南”
Empowering AI to Produce a 'Future Guide to AI Applications'

At present, we are all paying attention to the application of Lora model and Comfyui workflow. In my personal opinion, I would like to explain whether it is necessary for designers to learn model training and workflow. With the internal competition and improvement of AI tools (the competition between Chinese and American AI), we will see AI generation tools accelerating iteration at the speed of rocket upgrades. Today, we have just achieved font recognition and artistic generation, and tomorrow we will see sora2 video models that are close to live streaming. And today we finally see that AI can do banner design like an e-commerce graphic designer. Don't doubt that one day AI will replace designers in H5, detail page ultra long size and other work, because as long as there is a gap, both domestic and foreign algorithm teams will actively strive to fill it first.

Why do I think that in the future, universal models and large language models will be very powerful, and there is no need to add model training specifically? Because with the optimization and self-learning ability of the model, the logical precision and intelligence of the model are becoming higher and higher. As long as it can be inferred by humans, the model can only surpass it. The model infers at the pixel level, and the granularity of inference far exceeds human senses. All work related to logic will be replaced by AI in the future. So the current Lora model and Comfyui workflow are only supplements to the shortcomings of the current model stage, and they do not mean that they can continue to be applicable in the future!