We publish commentary on this industry, and readers occasionally notice what is missing from it: we almost never write about model releases. That is deliberate, and the reasoning says something about how this technology actually reaches people.
Two clocks, running at very different speeds
The announcement cycle runs weekly. Something is released, benchmarked, argued about, and largely forgotten inside a fortnight.
The cycle we care about runs in years. A capability appears; some time later it turns up in software people already use; some time after that it shows up in a physical object somebody can buy without knowing or caring what is inside it. By the time it reaches a kitchen counter, the announcement that started it is old news to everyone except the person now standing in the kitchen.
The lag is the interesting part
Coverage of the first clock is abundant and well done by people closer to it than we are. Coverage of the second is thinner, largely because it is slower and harder to make urgent.
But the second clock is the one that determines whether any of this was worth it. A capability that never survives the trip from a demo to a product that a tired person can use on a Tuesday has not really arrived, however impressive the demo was.
So what we write about instead
When a capability shows up in something you can buy, we are interested. When it changes what a job costs or how long it takes in a small business, we are interested. When a company quietly removes a feature because it turned out not to work, we are extremely interested, because that is the part nobody announces.
It makes for a slower publishing schedule and a considerably better hit rate on things that still matter a year later.