we have a year to fix security everywhere

The blog post asserts the urgency to address security given that consumer-grade hardware can now run large language models (LLMs) capable of causing significant harm. It emphasizes the need for immediate action to mitigate potential risks.