#48 — Cold Take: Broad automation trumps R&D acceleration
March 23, 2025•4 min read
Our take: The narrative surrounding AI's potential to accelete scientific progress – whether through curing disease or "solving energy" – is surprisignly popular despite not being backed by any rigourous economic arguments. R&D is generally not nearly as economically valuable as people assume.
The big picture: While startup founders increasingly recognize that AI's greatest economic value will come from widespread automation across industries rather than accelerating R&D, institutional players like governments, universities, and funding organizations continue to prioritize narrow research applications.
Why it matters: This misalignment could lead to suboptimal resource allocation and missed opportunities as institutions fail to support the AI applications with the greatest potential economic impact.
The founder-institution disconnect
Startup founders have intuitively grasped what recent economic analysis confirms: the most significant economic value from AI will stem from automating a wide range of tasks across various industries, not from narrowly accelerating research and development processes. This insight is driving their business strategies, with 86% of founders already reporting that AI positively impacts their go-to-market success.
Meanwhile, institutional players remain fixated on AI's potential to transform R&D:
- Government agencies continue to direct substantial funding toward AI research institutes rather than broad automation applications, with the National Science Foundation having awarded 25 "AI Institute" research grants since 2020.
- Universities are primarily focused on how AI affects research and academic integrity rather than preparing students for an economy transformed by broad automation.
- Funding organizations often prioritize AI applications in specialized research domains over tools that could automate routine tasks across multiple sectors.
The economic reality institutions are missing
The economic data strongly supports the founders' intuition:
- Only about 20% of U.S. labor productivity growth since 1988 has been driven by R&D spending.
- The labor elasticity of output is approximately 0.6 in the U.S. economy, which likely exceeds the "R&D elasticity of output" by around five times.
- Generative AI alone could potentially add between $2.6 trillion to $4.4 trillion annually to the global economy across 63 analyzed use cases.
- By 2030, every new dollar spent on business-related AI solutions and services will generate $4.60 into the global economy.
Where institutions should redirect their focus
To align with economic realities and follow founders' lead, institutions should:
-
Broaden funding priorities: Rather than concentrating resources on narrow research applications, government agencies should expand funding for AI solutions that automate routine tasks across multiple sectors.
-
Reimagine education: Universities should shift from viewing AI primarily as a research tool or academic integrity threat to preparing students for an economy where broad automation will transform virtually every industry.
-
Support practical applications: Accelerator programs and funding organizations should prioritize startups developing AI solutions for widespread automation rather than focusing exclusively on specialized research applications.
-
Reduce administrative overhead: Institutions themselves could benefit from AI automation, with potential to reduce administrative costs in higher education by up to 30% through automation of routine tasks.
The institutional opportunity
Some forward-thinking institutions are beginning to recognize this opportunity:
- The House Budget Committee recently held a roundtable to assess how AI could help "root out waste, rein-in the federal bureaucracy, and reignite economic growth".
- Leading research universities are increasingly looking to automation tools to help with grant applications and administration.
- Certain policymakers are calling for a broader approach to AI governance that promotes "competitive and innovative markets, drives economic productivity, and supports workers".
The bottom line: While startup founders have correctly identified broad automation as AI's primary economic driver, institutional players need to catch up. By shifting their focus from narrow research applications to supporting widespread automation across industries, governments, universities, and funding organizations can better align with economic realities and maximize AI's potential impact.
Keep reading
#49 — Gemini 2.5 Pro
Google's new "thinking model" AI shows impressive reasoning capabilities with a 40-point lead on human preference benchmarks.
#50 — Leading in uncertain times
Economic uncertainty demands a different leadership approach. In times of crisis, how leaders respond defines not just their survival, but their company's future success. The most effective leaders combine push and pull styles, focus on the "4 C's" (Communication, Conviction, Confidence, Calmness), and build trust through competence, integrity, and benevolence.
#51 — Crisis comms playbooks
The comms before the storm. Don’t get ready, stay ready. Because crisis can come when you don’t expect it — and when it does, there won’t be enough time to prepare.