geofft's comment hits the nail on the head when it comes to what effect this did vs. should've had on her tenure case, I think. A study that fails to replicate is no indication of malice or even incompetence. Some say she should've run self-replications, but how many years self-replication studies are you willing to run on $20k/yr (in fact, how many post-college years have you made 20k/yr for a full-time job)?
At some point you get the quality of science you can pay for, and throwing the grad student under the bus by denying them tenure a decade later is as shitty and spineless as Equifax throwing an IT guy under the bus for what was pretty clearly an organizational failure. IMO.
Re: TED talks, TED is not a scientific venue, and most of the things I hear in TED talks are a hell of a lot less scientific than even Cuddy's most flawed studies. Also, I don't think those sorts of things should weigh much in hiring and tenure cases either way (aside from e.g. a usually marginal "service/outreach" category).
If Harvard hired Cuddy because she gave a popular TED talk, then that was an extremely poor hiring decision by Harvard. But that's a separate issue from the one we're discussing.