A New AI Model Finds Computer Problems. Some People Use It Without Permission.
Introduction
Anthropic made a new AI model. It is called Claude Mythos Preview. The model can find problems in computer software by itself. The company gave the model to only 40 groups. But some people got it without permission.
Main Body
The model found many old problems in computer systems. One problem was 27 years old. It found problems in Linux and other systems. The company gave the model to big companies like Amazon, Apple, and Google. These companies use it to make computers safe. On the same day, some people used a private website to get the model. Anthropic is looking into this. Governments in Singapore and Australia are watching. They want to keep computers safe. In China, stock prices of security companies went up. People think the model can help find problems, but also can be used for attacks. Some experts say one bad person can now do what a whole team did before.
Conclusion
This new AI model is very powerful. It can help find problems, but also can be used for bad things. People are watching it. The future of computer safety is not clear.