David Blumberg on AI, Data Centers and Economic Opportunity

PragerU CEO Marissa Streit interviews venture capitalist David Blumberg about his political outlook, AI, data centers and education. The conversation is advocacy from participants with ideological and investment interests; forecasts are not guarantees.

Blumberg describes his interest in both intellectual traditions and entrepreneurship. He connects his support for PragerU with a desire to communicate values he considers durable.

He discusses demographic aging, public benefits and the difficulty of changing established programs. These are his political and economic interpretations, not a complete assessment of specific policies.

Blumberg recounts an early government work experience and contrasts incentives he observed in public and private organizations. His anecdote illustrates his own perspective rather than proving that all public or private institutions behave alike.

He describes research on economic relationships and diplomacy and says commercial ties can continue even when public political positions differ. The historical figures he cites are not independently established in the interview.

Streit and Blumberg discuss how he came to identify with conservative ideas and how political leaders respond to constituencies. They make clear that values and political interpretation are central to their views.

The conversation turns to AI. Blumberg gives a simplified explanation of language models as predicting the next word or token, and argues that the scale and speed of the process enable useful applications.

He distinguishes a chatbot that answers from an agent that can carry out steps using tools. His description of agents working continuously is a vision of possible workflows, not a guarantee of reliability or unrestricted capability.

Asked whether AI will eliminate jobs, Blumberg compares it with earlier technological changes that displaced some work and created other occupations. He expects new opportunities, but the timing, distribution and quality of future jobs remain uncertain.

He is optimistic about AI helping healthcare research and lowering costs. Faster analysis or candidate discovery does not by itself prove a treatment works; clinical testing, safety and equitable access remain essential.

Blumberg argues that AI and data-center services should be affordable and broadly available. As an investor in technology, he has an interest in the industry’s growth.

The interview addresses data-center electricity, water and environmental concerns. Blumberg emphasizes technology improvements and benefits; the actual effects depend on the facility, power supply, water use and local conditions and cannot be resolved by a general claim.

He speculates about future cooling, optical interconnects and even space-based data centers. These are prospective technologies and proposals, not proof that existing facilities have negligible environmental impacts.

Blumberg highlights possible local tax revenue and useful services from data centers. Communities also need to evaluate costs and distribution of benefits for a specific project.

They discuss US competition with China and possible friction for entry-level workers as routine tasks are automated. Blumberg expects workers to move to more productive activities; that outcome is a forecast, not an automatic transition.

He considers cross-border business and envisions AI helping small firms assess counterparties. The feasibility of such systems depends on reliable data, privacy, consent and validation.

The discussion shifts to immigration, entrepreneurship and demographic change. The speakers offer political preferences and economic expectations rather than a comprehensive analysis of immigration policy.

Streit asks whether AI systems can be trusted and how values enter their outputs. The speakers argue for exposing systems and users to a range of perspectives, while expressing their own preference for particular traditions.

They discuss the sources of information used in AI and the possibility of bias. A source’s availability, licensing and inclusion in a specific model’s training data are separate questions; the interview does not establish the training composition of any model.

Blumberg encourages parents to view AI as a tool while teaching children effort and judgment. He distinguishes learning the basics from copying a generated answer without understanding it.

He argues that people need enough subject knowledge to question AI output and detect mistakes. This is a practical educational principle, not evidence that any particular tool always helps or harms learning.

The interview ends with the speakers’ optimistic view of technology and their emphasis on human moral choices. PragerU also promotes its educational mission, which is part of the program’s institutional perspective.