Venture Opportunities - Hidden in Plain Sight

Five takeaways from this week’s issue:

1. Venture capital is separating into different businesses—not simply failing. Financing the scale of established AI leaders and discovering companies before consensus require different LPs, funds, checks and skills. Kyle Harrison sees strategy drift; Fred Wilson argues bigger funds can preserve investment discipline.

2. The overlooked opportunity is financing the discoverers. Specialist early-stage managers still find exceptional companies, but large LPs increasingly need somewhere to deploy enormous checks. Silent Ventures illustrates how specialist discovery and separately supplied follow-on capital can work together—not proof that small funds universally outperform. SignalRank's pro-rata as a service to smaller managers is similar.

3. AI’s next business opportunity is making capability usable. The “diffusion gap” and rise of deployment companies point toward integration, workflow redesign and accessible products. Better models alone do not produce paying customers. That creates room for startups beyond the frontier labs. There is opportunity everywhere.

4. More intelligence exposes the next bottleneck. Anthropic reports Claude uncovering the origins of cryptographic parameters; PostgreSQL reviewers are finding defects faster than maintainers can repair them. Progress can shift the constraint from discovering problems to verifying findings and acting on them—not eliminate the work.

5. Trustworthy AI needs enforceable controls, not reassuring promises. Nadella’s distinction between intelligence and authority connects with Anthropic’s independent-evaluator initiative and the robotics emphasis on bounded, reliable tasks. Permissions, audits, human intervention and independent scrutiny are foundations for deployment—not evidence that safety is solved.

This is a public episode. If you'd like to discuss this with other subscribers or get access to bonus episodes, visit www.thatwastheweek.com/subscribe