Theorem news
Theorem appears in 1 story we have published in October 2026, mostly in our AI coverage. It is most often reported alongside Anthropic and OpenAI.
Latest development
Anthropic's AI designs antivenom proteins, uncovers 500 security flaws, and hides messages in images
Anthropic’s AI models are pushing boundaries in scientific research, cybersecurity, and software verification. The company’s Claude Science system autonomously designed protein binders targeting deadly scorpion neurotoxins, while its AI-driven security analysis uncovered nearly 500 vulnerabilities in open-source projects. Additionally, Anthropic demonstrated AI’s ability to generate and detect hidden messages in images, highlighting both the potential and risks of advanced AI systems.
25 October 2026
