People need to understand the difference between LLMs and Neural Networks.

LLM training is a massive energy hog that gives us nothing but the illusion of coherent human-made text.

Non-LLM Neural Networks are much broader in use, almost always massively less energy-intensive to train, and often incredibly accurate when finely-tuned for specific purposes.

LLMs can die in a fire, and nothing would be lost. NNs in general are incredibly useful and honestly a massive source of potential for bettering healthcare (and science research in general) globally.