If you're a security researcher, there's a very clear arbitrage opportunity right now: quickly apply for the Claude Verification Program and OpenAI Daybreak Trusted Access for Cyber. This arbitrage window won't last long. Don't complain about regular models frequently triggering cyber restrictions, and don't think it's too difficult. It's precisely because most people find it troublesome that this arbitrage opportunity exists. Spend some time, and you'll get it done.