The Quiet Gift of 2025: Three Models That Changed Everything
Three frontier AI models from Anthropic, Google, and OpenAI arrived in December 2025. The gap between engineers using AI to reason through problems and those using it for autocomplete is widening every month.
Capability is the only durable AI moat. Tooling is rented; capability compounds.
Prioritize internal capability development over external tooling acquisition.
Organizational capability is defined by the skills and processes an entity can execute, not solely by the tools it possesses. True advantage stems from embedding expert knowledge directly into operational systems.
The utility of advanced AI models is maximized when organizations translate tacit knowledge and established practices into machine-executable forms, such as agent skills. This encoding process builds durable, proprietary intelligence.
Investment in AI must shift from evaluating vendor models to cultivating internal teams who can deeply integrate and adapt AI, turning generic models into specific, high-value organizational assets.
Rapid iteration with new AI capabilities is essential; the pace of innovation outstrips traditional adoption frameworks, making direct engagement more valuable than theoretical evaluation.
The first question for any AI program: is this building an internal capability, or is it merely consuming a service?
Capability is the only durable AI moat. Tooling is rented; capability compounds.
Prioritize internal capability development over external tooling acquisition.
Organizational capability is defined by the skills and processes an entity can execute, not solely by the tools it possesses. True advantage stems from embedding expert knowledge directly into operational systems.
The utility of advanced AI models is maximized when organizations translate tacit knowledge and established practices into machine-executable forms, such as agent skills. This encoding process builds durable, proprietary intelligence.
Investment in AI must shift from evaluating vendor models to cultivating internal teams who can deeply integrate and adapt AI, turning generic models into specific, high-value organizational assets.
Rapid iteration with new AI capabilities is essential; the pace of innovation outstrips traditional adoption frameworks, making direct engagement more valuable than theoretical evaluation.
The first question for any AI program: is this building an internal capability, or is it merely consuming a service?
After 20 years in software development, Norman is both a hands-on leader and defining the new age of AI SDLC for some of the biggest brands in the world — and exploring it with the builders. He writes here about things he is hearing and seeing. All posts are his personal points of view and do not reflect any employer or any customer he has ever had contact with.
The views and opinions expressed in this article are the author’s own and do not represent the positions of any employer, client, or affiliated organization.
