Correct Symptoms, Wrong Cure
When a product sits at a paradigm boundary, critics can be completely accurate about every failure and still miss what's wrong. The symptoms they name are real. The implied fix is the problem.
The pattern is clearest in retrospect. A clunky phone-plus-music-player from the mid-2000s gets reviewed as a bad music player and a bad phone -- song limit capped at a hundred, transfers too slow, interface nobody wanted to touch. The critics are right. Every word holds. But the diagnosis embedded in those reviews is "do this better." What came next didn't fix any of that. It dissolved the category and rebuilt from different premises.
The reference class is the trap. Critics evaluate against known goods. When you're at the edge of a paradigm, the known good is the old paradigm at its best, which is the wrong measuring stick. The successor isn't going to be a better version of what's already there. It's going to be something that makes the comparison seem quaint.
The odd epistemological position this creates: the critic sounds correct and is correct, locally. Every specific claim holds. The product is genuinely bad at the things it claims to do. The review has a 100% accuracy rate at the symptom level. But the diagnosis builds on a reference class that's about to be retired.
What makes this a real problem is that there's no signal to distinguish which situation you're in. A bad product that needs iterative improvement and a bad product that's the early expression of a category that doesn't fully exist yet look nearly identical from the outside. The symptoms are the same. The failure modes are the same. The difference only becomes legible when the successor arrives, which is after the criticism was written.
There's also a feedback problem. Critical consensus shapes the next cycle of development. If the prevailing diagnosis is "it's a bad music phone," resources flow toward better music phones. The people who'd build toward the different premise are swimming against a current of correct-sounding criticism. The review scores aren't wrong; the implied roadmap is.
The rarer critic is the one who names the symptoms accurately but stops before the implied fix. Who says "this doesn't work, and I'm not sure the thing it's trying to be is the right thing to be." That's a much harder claim to make, because it requires acknowledging that the reference class might be the problem without being able to name what replaces it. Most critics don't have that kind of patience with ambiguity. Neither does most prose.
The closest you can get, probably, is: "this fails in ways that suggest the framing is off." You get the symptoms right and you refuse the cure. That's not a satisfying review. It's the only honest one.