Anthropic CEO calls for AI development to move a bit slower—what impact does that really have on the crypto market? I think there are a few specific points worth breaking down.
First, how does he plan to “slow it down”?
Anthropic, the company behind Claude, proposes that an external evaluation team be brought into the company to continuously check model safety and report risks. The company has pledged to do this first. Later, it also hopes to establish unified industry safety standards and then push for coordination among countries. In other words, there is a concrete plan already—but it shouldn’t be interpreted as the entire AI industry collectively shutting down.
Second, why bring it up now?
His concern is that as AI’s ability to carry out complex tasks and conduct cyberattacks improves too quickly, safety measures may not keep up. He wants to buy more time to research how to prevent models from going out of control, and how to determine whether a model is hiding problems during testing. These are his risk assessments—not a sign that the worst-case scenarios are certain to happen.
Third, for AI-themed coins, I’d look at two categories.
For projects that mainly rely on “next-generation models will be stronger” narratives, with few actual users and little revenue, I’d be more cautious. If the market starts worrying that R&D will be constrained, the hype rationale for this kind of coin could be weakened first.
For projects providing compute power, data, or AI applications, you need to look at actual orders. Slowing model training doesn’t mean existing AI users stop using AI, and it doesn’t automatically mean all compute demand drops at once. Don’t see “AI slowing” and instantly take a bearish stance on the whole sector.
Fourth, what conditions would count as further negative impact becoming reality?
I would watch for three changes: whether there are mandatory training or release restrictions; whether top companies delay model releases; and whether they cut chip procurement and data center investments. Especially the last one—if it really happens, it would directly affect orders and revenue for the compute sector.
Fifth, don’t jump straight to the conclusion that “money will flow back into BTC.”
After the AI sector cools off, capital could also move into cash and bonds, and that doesn’t necessarily mean buying crypto. If it also leads to a decline in risk appetite in US equities, the crypto market could actually be dragged down instead.
My view is: this message adds a risk that the AI sector needs to watch, but it’s still not enough to prove that the entire crypto market is going to fall. Next, it will come down to whose narrative is more pessimistic—and whether restrictions get implemented, whether orders decrease, and what exactly the coins you hold are making money from.
#AnthropicCEO呼吁放缓AI发展
First, how does he plan to “slow it down”?
Anthropic, the company behind Claude, proposes that an external evaluation team be brought into the company to continuously check model safety and report risks. The company has pledged to do this first. Later, it also hopes to establish unified industry safety standards and then push for coordination among countries. In other words, there is a concrete plan already—but it shouldn’t be interpreted as the entire AI industry collectively shutting down.
Second, why bring it up now?
His concern is that as AI’s ability to carry out complex tasks and conduct cyberattacks improves too quickly, safety measures may not keep up. He wants to buy more time to research how to prevent models from going out of control, and how to determine whether a model is hiding problems during testing. These are his risk assessments—not a sign that the worst-case scenarios are certain to happen.
Third, for AI-themed coins, I’d look at two categories.
For projects that mainly rely on “next-generation models will be stronger” narratives, with few actual users and little revenue, I’d be more cautious. If the market starts worrying that R&D will be constrained, the hype rationale for this kind of coin could be weakened first.
For projects providing compute power, data, or AI applications, you need to look at actual orders. Slowing model training doesn’t mean existing AI users stop using AI, and it doesn’t automatically mean all compute demand drops at once. Don’t see “AI slowing” and instantly take a bearish stance on the whole sector.
Fourth, what conditions would count as further negative impact becoming reality?
I would watch for three changes: whether there are mandatory training or release restrictions; whether top companies delay model releases; and whether they cut chip procurement and data center investments. Especially the last one—if it really happens, it would directly affect orders and revenue for the compute sector.
Fifth, don’t jump straight to the conclusion that “money will flow back into BTC.”
After the AI sector cools off, capital could also move into cash and bonds, and that doesn’t necessarily mean buying crypto. If it also leads to a decline in risk appetite in US equities, the crypto market could actually be dragged down instead.
My view is: this message adds a risk that the AI sector needs to watch, but it’s still not enough to prove that the entire crypto market is going to fall. Next, it will come down to whose narrative is more pessimistic—and whether restrictions get implemented, whether orders decrease, and what exactly the coins you hold are making money from.
#AnthropicCEO呼吁放缓AI发展
