Are the founders of two American AI giants a conscientious discovery or a prisoner's dilemma? ---Mr. Xiaolong's exclusive analysis and perspective On September 13th, Anthropic CEO Dario Amodei published a lengthy article stating that the industry must slow down the pace of improving AI model capabilities. OpenAI founder Sam Altman immediately agreed: I agree, this is the main topic we discussed at OpenAI. Meanwhile, Sam Altman also tweeted a warning: The progress of artificial intelligence may lead to two extremely bad situations. One is that we may lose control of the future and be controlled by artificial intelligence. The other is that humans may fall into a world where the power to control AI is too centralized. ” Artificial intelligence must always serve humanity. To ensure this, we need methods to ensure that alignment and security technologies always stay ahead of advances in model capabilities The two situations mentioned by Ultraman actually mean the same thing: the iterative updates and development of AI pose a huge danger to humanity! The two giants shouting 'slow down' at the same time may sound responsible, but I think they are just putting on a show. The core is driven by interests, and conscience discovery cannot exist in the genes of American giant AI. Take a look at what they're saying: 'Slowing down doesn't mean stopping model training or technological progress.' Translation: It's not about not doing it, it's about not doing it too fast. That's interesting. If you really think that AI may lose control, alignment technology cannot keep up with model iteration, and concentration of power is a dystopian risk. Why didn't you stop? Why is it 'slower'? The answer is simple: whoever stops first loses. The market sees very clearly. This statement has been interpreted as a "prisoner's dilemma": both sides know that continuing to charge carries risks, but neither dare to really stop. So the two AI giants reached a tacit understanding: shouting about safety while continuing to run, just slowing down the pace a bit and giving the market an explanation. What's even more worth pondering is another layer. While Amodei and Altman are calling for a "slowdown," the market has noticed a signal that giants may want to postpone the next round of large-scale research and development, monetize existing products first, and cool down infrastructure and capital expenditures. So is' safety 'the truth or a fig leaf for financial reports? The two AI giants have not yet gone public. If they raise funds through IPO, they will secretly work hard to develop more advanced AI models. If AI really has the potential to exterminate humanity, the correct approach is not to "slow down", but to "stop and talk clearly". But if AI only needs "a little more time to align", it is essentially an engineering problem, not a survival crisis. Two narratives correspond to two interests. Which one to choose depends on whose position you stand on. The market has already voted with its feet. AI related targets fell 3-5% ahead of schedule, and pre-market expectations continued to weaken. Investors understood the underlying message: this is not about prioritizing safety, this is about slowing down growth. The two giants shout 'slow down', their biggest fear is not AI losing control, but the market losing control first.