OpenAI Co-Founder Issues Digital Hygiene Guide as Claude Mythos Raises AI Cybersecurity Threats

Anthropic’s Claude Mythos is said to have sharply lowered the barriers to cyberattacks, making it easier to create malware and exploit vulnerabilities at scale. Developers and cryptoasset holders are high-risk targets because their accounts, private keys and funds are concentrated. OpenAI co-founder Andrej Karpathy compiled a set of personal cybersecurity practices in response, warning that advancing AI capabilities are also amplifying the risk of digital-asset theft.

Karpathy has now released a 15-step “digital hygiene guide” recommending password managers, hardware security keys and biometric authentication, along with tighter access controls for critical accounts and cryptoassets. Available information does not specify the guide’s exact publication date or cite any transaction or loss amounts. The key figure is its 15 protective measures, which are intended to reduce the likelihood of credential theft and account takeovers.