Sure, and that's a legitimate conclusion to reach. Its different than a "no evidence" situation, and different reasonable people can reasonably view the contradictory claims as having different weights and not simply offsetting.
EDIT: And we actually have more than that, since we have statements from several individuals related to the events, plus the Github official statement in their press release on the investigation they commissioned, and those statements do not amount to simple direct contradictions of each other; but the point remains, there is a difference between a "no evidence" situation and a situation where there is a pile of evidence but where a particular observer doesn't find it sufficient to justify a particular conclusion or action.
Witness testimony is generally considered an important form of evidence. Obviously, whether there is corroborating (or, OTOH, conflicting) evidence effects the conclusions that one draws, but that's true of any instance of evidence, not just statements.
> If I stand up in the office and claim that I saw personA murder someone, does that suddenly constitute non-zero evidence that personA is a murderer?
Yes, although standing on its own it would be extremely weak evidence.
> If I write a program to randomly generate claims and post them to my Twitter account, is each tweet a non-zero amount of evidence?
No, because that's not an person making a claim of facts about which they assert direct knowledge. Its kind of irrelevant to the present circumstance, given that no one has challenged the authenticity of the source of Horvath's (or of the other participant's) claims here.