The agent assembled a comprehensive explanation of Arkane Capital by working around an inaccessible website. The arkane.capital homepage itself was heavily JavaScript-rendered with no extractable text, and key pages returned 404 errors. The agent instead relied on external sources—a VCSheet podcast interview and a Substack article—to answer all core questions about what Arkane does, who it's for, check sizes, and differentiators. The site itself did not surface the information needed to fulfill the task.

• ›Steps [16] and [17] (the VCSheet and Substack sources) were the only content-bearing sources; they contained all substantive information about Arkane's investment thesis, check sizes ($300K–$500K), focus areas (neuro, energy, logistics), founder background, and AI-driven discovery methodology.
• ›The arkane.capital homepage (steps [1] and [9]) returned HTML that was heavily JavaScript-dependent with truncated, pre-rendered Next.js markup—no meaningful text content about the fund's strategy, team, or process could be extracted.
• ›Attempts to navigate to structured pages (/about, /portfolio) failed with 404 errors (steps [2], [3]), indicating the site is a minimal branding placeholder rather than an informational resource. The agent had to compensate with eight web searches to find external coverage.
• ›The agent was explicit about what it could not find on arkane.capital itself: team details beyond Arkady Kulik, portfolio companies, application process, and a clear differentiators statement. All answers came from third-party sources, not from the domain being evaluated.
• ›The site's machine-readability is very low—no structured data, no plaintext content, no navigable information architecture. An agent cannot extract what the fund does, who to contact, or how to apply from arkane.capital alone.