Ben Thompson’s Deep Research and Knowledge Value is an insightful read on the fundamental shifts AI is driving in knowledge work – most intriguingly, to me, the rising value of secrecy in an era of increasing transparency.
It’s a paradox, isn’t it? The more AI makes knowledge instantly accessible, the more valuable the information it can’t reach becomes. If anyone with $200 can generate academic-level research from public data, using tools such as OpenAI’s new “Deep Research”, what happens to the competitive edge that used to come from simply knowing more? Where does real differentiation come from when AI can synthesize everything that’s already out there?
One of the most compelling insights from this piece is the realization that AI is making known unknowns more accessible than ever before, while simultaneously increasing the value of unknown knowns – the information that remains deliberately hidden from public view. Historically, information asymmetry has been a source of power, but as AI accelerates the process of extracting, analyzing, and synthesizing publicly available data, the scarcity of proprietary, non-public knowledge will become an even greater competitive advantage.
Thompson illustrates this with two fascinating case studies. In one, Deep Research uncovered a potential medical insight within minutes – something that had eluded months of professional consultations. In another, it produced a seemingly comprehensive industry analysis that turned out to be completely useless because it omitted a key, obscure player that wasn’t well-documented online. The latter failure is particularly revealing: it wasn’t an error of misinformation, but an illusion of completeness – an unknown known that AI couldn’t surface because the data simply wasn’t readily available on the internet.
This raises profound economic and strategic implications. If AI-driven research can synthesize everything that’s public, then true differentiation – and true competitive advantage – will increasingly come from what isn’t public.
This leads to the conclusion that in the AI era, the ability to withhold information – whether through proprietary datasets, private market intelligence, or human-driven qualitative insights – becomes more valuable than ever. Therefore, AI’s ability to “know everything public” will push organizations toward building deliberate walls, creating new forms of scarcity through proprietary research and exclusive prediction markets.
The implications are staggering. Just as the internet devalued news by making it instantly accessible and monetized elsewhere, AI may be doing the same to research and analysis – compressing hours of work into minutes, but in the process, diminishing the economic value of human expertise. The future won’t belong to those who can process public information the fastest, but to those who have access to the data AI cannot reach.
This is exactly why we at CTERA are developing CTERA Data Intelligence – an AI tool designed to power research on the most sensitive corporate private data.
If AI democratizes synthesis, then differentiation will come not from what is known, but from who knows what first – and, more importantly, who can keep it to themselves.
CTERA Data Intelligence is all about extracting knowledge from your private data – while keeping your secrets secret. Want to learn more? Contact Us
