Since it's come up, I'll clarify my position and why I won't be following the rule as perhaps some people envisage it. There's an unavoidable pressure to present the best interpretation of some work, whether it's in descriptions by the researchers/inventors/whatever themselves, the university/company PR people doing press releases, the general media, etc. Indeed, after several stages of this, what is initially a perfectly reasonable "piece of work", with its uncertainties and "these are all preliminary results" has become an iron-clad important breakthrough. To me, saying that things are more uncertain, or figures used are at the optimistic end of the range of "possible" values, or that more work will be needed to increase the support for the conclusions, etc, is simply attempting to counteract the inflation effect and get a more "middle of the spectrum" value that can be used when thinking about how this effects other things. In such cases, I don't consider this a criticism as such, and don't feel obliged to offer alternatives.

In cases where I am definitely stating that I think there's something actually wrong/misguided, rather than just overstated, I do try and stick with Eric's suggested rule.