I think you're putting words into his mouth, and perhaps intentionally misrepresenting what he wrote.

Swift was announced just a few hours ago. While we can make assumptions about its safety based on its feature set, or we can assume that the very vague performance details Apple provided are valid in a real-world setting, none of this has really been tested or verified independently yet.

What cratermoon wrote is perfectly legitimate. Let's try Swift out in the real world, perhaps for a year or two. Let's get some real data. Then let's analyze that actual data, rather than engaging in unsubstantiated speculation, or trusting some vague performance graphs shown in a conference presentation.