I strongly disagree with this. Apple did make a mistake. It was badly priced and both too close and too far in capabilities from the other phones in the lineup. Being marginally thinner to get a worse battery, worse CPU/GPU than the Pro, worse camera than either, made no sense at all.
I wrote a comment about this one before. The Air never made sense. It is priced between the regular 17 and the Pro. Going with rough size equivalents (so no maxes) the Air has a 0.2" bigger display than the 17 or 17 Pro with an appropriate amount of extra pixels. It has worse battery than both. Worse camera than both (or fewer camera capabilities than both). Its CPU/GPU is closer to the Pro so that is somewhat compelling versus the regular 17.
But it is on net a worse phone, or insufficiently compelling except to someone who wants to save $100 for a size equivalent 17 Pro. The Pro is "only" $100 more but overall much better (in quotes because that can be a lot to many people, but if that's a lot, don't buy a $999 phone). And the $200 doesn't buy enough to warrant the upsell from the regular 17 considering what's lost (camera, battery life).
The right product was either no product or going back to the mini (or something closer to it). The mini at least fits a niche that isn't filled by the 17 and 17 Pro, and the camera and battery tradeoffs are things that would be acceptable in that size of a device. Retain the single lens on the back, maintain the same battery volume (to retain about the same battery life, maybe improved with the smaller screen) which means thickening the device but not substantially (should be about on par with the 17 and 17 Pro).
The best thing about the Air is getting the SOC down to such a small size, which means that future phones could become basically a battery + screen + tiny SOC whose size is determined more by the camera lenses attached than by the actual chip and storage needs. That's not nothing, and if the cost of developing that was one lousy phone for this generation that isn't too bad. Hopefully they figure out how to use that development in a phone people want next time.