OpenAI says its AI found 948 Linux flaws and wrote an exploit for every one

Share

OpenAI says its autonomous vulnerability-discovery pipeline analyzed the Linux kernel in mid-October 2026 and validated 1,724 flaws from 2,308 reported candidates. The system produced 948 unique privilege-escalation attacks that grant root access, along with 291 proofs-of-concept for information-disclosure vulnerabilities. The total operating cost was $1,213,430, or approximately $1,279 per exploit, and 61 vulnerabilities have already received upstream patches. The findings highlight that cryptocurrency decentralization does not protect individual systems running Linux, including nodes, validators, and key-custody infrastructure.

Source: Read the original article

Disclaimer: this content is for information purposes only and is not financial advice. Cryptocurrencies are highly volatile: you may lose all of your capital. Always do your own research. Legal notice
Telemac
Telemachttp://cryptoinfo.ch
Passionné de nouvelles technologies, j’explore l’univers de la blockchain et des cryptomonnaies pour partager l’actualité et les innovations du secteur.

Read More

Items