The article discusses experiments using Opus 4.5 and GPT-5.2 to generate exploits for a zero-day vulnerability in QuickJS. It concludes that the future of offensive cybersecurity may rely on token throughput rather than the number of human hackers, as LLMs prove effective in exploit development.

Generative AI models, such as OpenAI's GPT-4, are enabling rapid development of exploit code from vulnerability disclosures, reducing the time from flaw announcement to proof-of-concept to mere hours. Security experts have observed a significant increase in the speed at which vulnerabilities are exploited, necessitating quicker responses from defenders in the cybersecurity landscape. This shift underscores the need for enterprises to be prepared for immediate action upon the release of new vulnerabilities.