https://arxiv.org/pdf/2301.10263v2.pdf
It's an interesting idea because the techniques used in the model allow for star formation in the early universe to be in the order of thousands of solar masses, not hundreds, as is the case now. Star formation is typically limited by the modified Eddington limit in which the radiative pressure is balanced by gravity, such that a star on the main sequence is in hydrostatic equilibrium. Too much mass and a star will produce too much energy resulting in massive radiative-driven stellar winds, upsetting the balance between gravity and radiation, resulting in a star that can't hold itself together. The limit itself is also dependent on the type of material of which a star is composed. Heavier elements mean higher temperatures, which in turn means higher radiative pressures, thus limiting the mass. See here for further explanation https://en.wikipedia.org/wiki/Eddington_luminosity.
Of course, early on in the universe, prior to stellar evolution, hydrogen was the most abundant gas and being free from heavier elements coalesced into stars of high mass, pop III stars. It is postulated that their mass would've been significantly higher than pop II and pop I stars by virtue of the purity of the gas from which they evolved. https://en.wikipedia.org/wiki/Stellar_population This paper goes one step further and postulates that super-massive stars could have evolved via a process known as cold accretion.
Cold accretion is nothing new, in itself. However, the mechanism here is that halos of gas in the order of 10^7-8 solar masses are triggered into cooling by the Lyman-alpha process. As the cooling process takes place the cloud monolithically collapses into a protostellar core. Cold accretion flows are shunted via thick gaseous filaments deep into the protostellar core. This process creates shocks over the protostellar core significantly increasing its mass and temperature resulting in super massive stars. These super massive stars can reach the order of 10^4-6 solar masses. These stars then collapse with relatively little mass loss resulting in heavier seed BHs.
The paper starts by pointing out that there are a number of quasars at z > 6 (high redshifts). A quasar is essentially an AGN with high luminosity and an
It isn't without its problems, but it is an interesting proposition.