An analysis of AI model lifespans reveals that median durations are heavily influenced by vendor reporting practices rather than inherent model longevity. The author found that the median lifespan of AI models, when calculated based on models with announced shutdown dates, shifted significantly over a short period due to changes in the sample size and vendor inclusion. Google's frequent release of preview models with set end dates skewed its median lifespan lower, while the longest observed lifespan was for GPT-4, spanning over 1,200 days based on disparate arrival and shutdown date records from different vendors. AI
IMPACT Vendor reporting practices significantly impact perceived AI model lifespans, complicating planning for users and developers.
RANK_REASON The item is an analysis and opinion piece on AI model lifespans and vendor reporting practices, not a direct release or product announcement.
AI-generated summary · Google Gemini · from 1 sources. How we write summaries →