Two years ago, I tweeted about this: AI replaces either humans or existing software, and the balance between cost and efficiency of these humans or software has already been optimized to the extreme in today's world. Therefore, the tipping point of AI's value essentially lies in AI surpassing the societal intelligence cost threshold. Once it crosses that point, AI's value indeed rises non-linearly, so if AI scores less than 60, it's basically a zero. Two years later, large models are getting smarter and have clearly crossed the societal intelligence cost threshold. AI agents are starting to take over more scenarios in work and life. The 'AI disrupts everything' narrative is already reflected in U.S. software stocks, and scenarios where humans are replaced are gradually happening. The more people doubted it before, the more anxious and panicked they are now. There's no need to fud the decline of the old era; trends are irreversible and bound to happen. For individuals, the more important thing is figuring out how to find your value in the future—to become someone who benefits from AI rather than someone who can be replaced.