Cyber defense is dominant, risks from open coding models manageable
_sholtodouglas · x · 2026-08-15
Regarding the security risks posed by open-source AI models, Sholto Douglas argues that cybersecurity is inherently defense-dominant. If financial institutions and critical infrastructure providers actively conduct white-hat hacking over the next two years, he has zero concern about the security of any open coding model. He suggests a trade-off: accepting substantially increased cyber attack risks to preserve freedom, provided the government actively makes this choice after evaluating the risks. Conversely, he notes that biology is offense-dominant, and fixing vulnerabilities in this domain will likely extend into the 2030s, implying we will face significant risks for a long time.
Related event: Debate on AI Risks and Open Source Model Safety(3 posts)→
More from Safety
- Polymarket: 69% Chance a US State Enacts Data Center Moratorium by End of 2026 — Polymarket · 2026-08-16
- Strong AI Governance Becomes Key Competitive Advantage Over Raw Compute — noahsolomon · 2026-08-16
- Scaling laws are predictable, but risk is jagged and concentrates unexpectedly — chrisrohlf · 2026-08-16
- Google experimented with text watermarking back in 2011 — yoavgo · 2026-08-16
- Who is accountable when AI agents make bad decisions? — KKevinjad · 2026-08-16
- AI watermarking detection relies on original logits and prompts — binarybits · 2026-08-16