I agree with your comment in principle but the specific example of Roundup is rather weak. Yes, glyphosate is usually studied in isolation rather than as a commercial formulation including surfactants etc., but we have decades’ worth of epidemiological studies showing that proper use of Roundup has, at most, a marginal effect on tumour incidence. Conversely, the most-discussed studies that purport to show Roundup’s carcinogenicity have well-known, glaring methodological flaws (including some that formed the basis of the IARC report). The case of Roundup is made more complicated by the fact that Monsanto/Bayer has been caught red-handed skewing the publication record without disclosure of conflict of interest, and lobbying scientific journals. But the same is true for the opposition: for instance, the now-retracted 2012 Séralini study also failed to disclose the authors’ conflict of interest. And beyond improper publication practices (which, yes, is serious), there’s no evidence that Monsanto/Bayer actually falsified information.

In sum, I’d rank the risk of Roundup being carcinogenic on roughly the same level as that of 5G: possible but unlikely, given the best available evidence.