(The Wall Street Journal, Holman W. Jenkins, Jr., April 15, 2026)
Anthropic chose not to make its new Mythos model public to avoid risks of cybercrime and lawsuits, instead reporting thousands of unknown vulnerabilities in the most widely used software to governments and responsible private actors, thus turning a potential threat into a net benefit for society.
The same capitalist forces that pushed Anthropic to act responsibly are allowing the United States to maintain its advantage in artificial intelligence over China, demonstrating that market incentives work better than any statist or “non-profit” approach.
Mythos has already paid for itself for society because it made visible thousands of security flaws before they were exploited by criminals or hostile governments, paving the way for a new billion-dollar business in fixing such vulnerabilities and overall improving the quality of human decisions.
1. Responsible choice by Anthropic
“Anthropic did not release Mythos to the public. Instead, the leading artificial intelligence company reported to governments and responsible private actors that its new AI model showed an extraordinary ability to discover thousands of unsuspected security flaws in the most widely used software products.”
2. Mythos does not create, but discovers vulnerabilities
“Mythos did not create the vulnerabilities. It found them. Identifying these risks is a net gain for the world, not a disaster, not a reason to pull the emergency brake on artificial intelligence.”
3. Capitalist incentives at work
“How could it have been in Anthropic’s interest to release Mythos knowing it could have triggered a global wave of cybercrime? Instead, Anthropic made a huge global commercial advertisement for its business.”
4. Solution to cyber threats
“Anthropic’s new product is not a cyber threat but a solution to cyber threats. Now it can be expected to create a billion-dollar business repairing the software flaws discovered by Mythos.”
5. Real opportunity of artificial intelligence
“This is the real opportunity of artificial intelligence: the opportunity to improve all forms of public and private decision-making processes. Artificial intelligence makes better thinking cheaper.”
(Excerpt from the newsletter by Giuseppe Liturri)
