The New York Times article discusses the significant measurement problem facing artificial intelligence (AI) technologies, emphasizing that there is currently no standardized method to evaluate their performance. This lack of reliable metrics complicates users' decisions on which AI tools to use and raises safety concerns, as companies primarily rely on vague claims about their products' capabilities. To address these issues, the article suggests that both public and private efforts are necessary to develop rigorous testing programs and promote transparency in AI evaluations.