Jason Clinton, deputy CISO, Anthropic. Interviewed by Shannon Yavorsky. Clinton leads security at Anthropic and focuses on AI safety, cybersecurity and infrastructure.
Key takeaways
- AI capability growth remains structurally predictable
- Compute used in training models has grown exponentially for decades, with ~4x year-on-year increases expected to continue in the near term.
- This underpins high confidence that models will become materially more capable over the next two years.
- Implication: software economics may collapse—but value persists elsewhere
- Clinton suggests the marginal cost of software could approach zero, driven by AI-generated code and automation.
- However, value will shift to branding, convenience and especially data moats, which remain hard to replicate (e.g. security platforms with large deployed datasets).
- AI companies are evolving into ‘neo-hyperscalers’
- Frontier model providers are positioning themselves as infrastructure platforms, offering “intelligence as a service”, analogous to cloud computing.
- The long-term opportunity lies less in consumer features and more in enterprise APIs and backend integration.
- Cybersecurity is entering an AI-driven arms race
- AI models are now capable of offensive techniques previously limited to nation-state actors, lowering the barrier for attackers.
- Defensive strategies must therefore layer multiple controls (“Swiss cheese model”) augmented with AI at each layer.
- Agents are already transforming enterprise workflows
- Internal agents can autonomously diagnose outages, generate fixes and coordinate teams via existing tools (e.g. Slack), reducing resolution times to minutes.
- Use cases extend across legal, product, partnerships and data analysis, often replacing traditional data science workflows.
- Human oversight remains central—but is shifting
- Systems increasingly operate with “humans on the loop” rather than “in the loop”, with humans acting as final approvers rather than active operators.
- Legal frameworks (e.g. copyright) still hinge on human involvement, leaving open questions as autonomy increases.
- Mission alignment is a competitive advantage
- Strong organisational alignment around principles (e.g. restrictions on harmful use cases) drives better decision-making and execution across teams.
Takeaway for CVCs
The investable edge is shifting away from standalone software and towards infrastructure, data moats and security capabilities in an AI-accelerated environment where both opportunity and risk are compounding rapidly.
This is an AI-generated summary, which has been lightly edited by GCV staff.


