Business Pill · What breaks when the model under you changes
Model deprecation is a provider retiring a model version, so everything built on it has to move. The new model may be better on average and still worse at your task.
A short explainer video, under a minute. The bakery and its ordering assistant are an illustration.
At a glance
What Model Deprecation Is
That is called model deprecation. It means a provider retiring a model version, so everything built on it has to move.
What Can Go Wrong
The new model may be better on average, and still worse at your task.
Why It Matters
The video turns this into one question: what breaks when the model under us changes?
Two versions of one model family are being compared side by side. In our report on ThinkingBox, the post says Claude Opus 5.5 and Claude Opus 5 each pass 241 of 507 tasks on all 20 attempts, at estimated costs of $7.80 and $13.30 per dependable task.
From our news coverage
The costs are the authors’ estimates at list rates, and our report verified none of the figures.
The Business Engineer Lens
Where this idea sits in the Business Engineer library:
The Business Engineer framework maps where each part of the AI stack sits and who holds the leverage. These pills are short lessons from the same library.
Model Deprecation constrains Fine-Tuning. When a new base model arrives you have to tune again, and deprecation forces that move on the provider’s schedule.
Model Deprecation builds on AI Evaluation as Acceptance Test. The defense against deprecation is keeping your own test cases and rerunning them before every switch, an acceptance evaluation.
Model Deprecation builds on Switching Costs. Deprecation forces a model switch, and the switching-costs lesson says the prompts, tests and connections around the model are the costly part.
Model Deprecation is an example of One-Sided Terms(lesson unlocks soon). The provider alone decides to retire the model, a rule only one side can change.
Model Deprecation is an example of Embedded Dependency(lesson unlocks soon). A product built on a rented model relies on a supplier inside it, and retirement exposes that hidden layer.
Gennaro Cuofano is a CRO and tech executive who has worked in AI since late 2015, starting with bringing NLP, natural-language generation, voice and chatbot products to the marketing industry. His background is in law and finance: he holds a Master's degree in Law and an International MBA with an emphasis on corporate finance (LUISS Business School and the University of San Diego, 2012), and worked as a financial analyst at a real-estate investment firm in San Diego and as an assistant controller. His work focuses on business model strategy, business engineering and, more broadly, structural analysis: how companies make money, read from their own filings. He created FourWeekMBA and leads research there and at The Business Engineer, his newsletter on AI and business strategy, with over 95,000 subscribers and more than 1,000 published analyses. His writing has also appeared on Entrepreneur (for example, How to Digitize a Family Business in Europe) and HackerNoon. Find him on LinkedIn and Substack. See how we source.
