This is true in principle, but since I take disagreements pretty seriously I think it is normally false in practice. In other words there is actual harm and actual benefit in almost every real case.
Of course the last part of your comment is still true, namely that a mixed cause could still be better than a pure benefit case. However, this will not be true on average, and especially if I am always acting on my own opinion, since I will not always be right.
... a mixed cause could still be better than a pure benefit case. However, this will not be true on average ...
That's the question, what is the base rate of the options you are likely to notice. If visible causes come in equivalent pairs, one with harm in it and another without, all other traits similar, that would be true. Similarly if pure benefit causes tend to be stronger. But it could be the case that best pure benefit causes have less positive impact than best mixed benefit causes.
...... since I take disagreements pretty seriously I think it is nor
The article is here.
The book is by William MacAskill, founder of 80000 Hours and Giving What We Can. Excerpt: