In November, a paper by Wayment-Steele and coworkers showed that clustering the multiple sequence alignment (MSA) used by AF2 enabled the program to generate multiple relevant protein conformations, with a method they called AF-cluster.
Since then, a couple of different preprints have challenged the results from this paper: Porter and colleagues suggest that you can generate multiple protein conformations from single sequences, while more recently Schafer and coworkers (Porter and colleagues, again) assert that the Wayment-Steele paper lacks some essential controls needed to assess AF-cluster's reliability.
I see this as the scientific process working - the combination of the research from the three papers and the inevitable back-and-forth will likely result in a clearer picture of what we can do with AlphaFold.