|Oculudentavis skull (after Xing et al., 2020).|
Just look at it. No antorbital fenestra, incomplete ventral bar to the laterotemporal fenestra, huge posttemporal fenestrae, teeth that extend posteriorly far under the orbit...
All of which might be coincidental, but then look at the mandible.
|Oculudentavis mandible (after Xing et al., 2020).|
That spike-like coronoid process is classic lepidosaur, plus the dentary is way too long compared to the post-dentary elements, then the description says "The tooth geometry appears to be acrodont to pleurodont; no grooves or sockets are discernable." And of course "the scleral ring is very large and is formed by elongated spoon-shaped ossicles; a morphology similar to this is otherwise known only in lizards (for example, Lacerta viridis)."
Add to this the size of this partially fused specimen being smaller than any extant bird (14 mm), and no feather remains, and why is this a theropod again? The endocast is big, but why not a clade of brainier lizards or late surviving megalancosaurs by the Cenomanian?
The authors add it to Jingmai's bird analysis where it ends in a huge polytomy closer to Aves than Archaeopteryx, but outside fake Ornithuromorpha. That's often what happens when a taxon is wrongly placed in a clade. Note the figured placement between Archaeopteryx and Jeholornis is only found using implied weights. At least add it to e.g. Nesbitt's or Ezcurra's archosauromorph analyses, or Cau's theropod analyses before assuming it's a bird.
Thanks to Ruben Molina Perez for suggesting this issue in the first place.
Reference- Xing, O'Connor, Schmitz, Chiappe, McKellar, Yi and Li, 2020. Hummingbird-sized dinosaur from the Cretaceous period of Myanmar. Nature. 579, 245-249.