In terms of the Consumer Reports numbers, as I noted, the problem areas for the Model 3 are "body hardware" (the worst of the three), "paint and trim," and "in-car electronics." OTA updates won't affect either of the first two categories, either positively or negatively.
I don't recall the exact phrasing, but CR asks respondents to note issues that have caused significant problems in terms of downtime or cost, so I'd expect that a new bug that's fixed with the next software release would not show up in the survey unless it was really bad -- say, if an update caused the center console to stay black and not respond to a reset, thus necessitating a tow to a service center. "In-car electronics" is a broad enough category that it could include both hardware and software, and so we can't know how many of those problems fall into each of those two sub-categories.
All that said, the CR results come from a survey and so are somewhat subjective; there's no way to prevent somebody from reporting the "dancing cars" in the display as a serious electronics problem. Survey research is also subject to a host of difficulties and limitations, but those are well-known by people who do such research. There are statistical tools that can help spot insincere responses, but to some extent, survey research is protected simply by large numbers of responses. CR has a lower limit of a certain number of responses before they'll include a model in their survey results. (I don't recall what that number is, offhand.)