Update docs/paper/paper.md

Co-authored-by: Copilot <175728472+Copilot@users.noreply.github.com>

Author: nichollsh

docs/paper/paper.md

@@ -52,7 +52,7 @@ AGNI is developed with the view of being coupled into the [PROTEUS framework](ht
 * solve for an atmospheric temperature structure that conserves energy and allows for convectively stable regions,
 * operate with sufficient speed such that it may participate in the exploration of a wide parameter space.
 
-These are possible due to the method by which AGNI numerically obtains a solution for atmospheric temperature structure and energy transport [@nicholls_convection_2025]. Our model uses the Newton--Raphson method to conserve energy fluxes through each level of the column to a required tolerance. A typical runtime when applying the model standalone using its command-line interface (\autoref{fig:application}b) with a poor initial guess of the true temperature profile is 3 minutes. When providing a 'good' guess, such as when AGNI is coupled within the PROTEUS framework (\autoref{fig:application}a), an atmosphere solution will be obtained in less than 1 minute. A single radiative transfer calculation takes approximately 30 ms, performed under the correlated-k and two-stream approximations using [SOCRATES](https://github.com/FormingWorlds
+These are possible due to the method by which AGNI numerically obtains a solution for atmospheric temperature structure and energy transport [@nicholls_convection_2025]. Our model uses the Newton--Raphson method to conserve energy fluxes through each level of the column to a required tolerance. A typical runtime when applying the model standalone using its command-line interface (\autoref{fig:application}b) with a poor initial guess of the true temperature profile is 3 minutes. When providing a 'good' guess, such as when AGNI is coupled within the PROTEUS framework (\autoref{fig:application}a), an atmosphere solution will be obtained in less than 1 minute. A single radiative transfer calculation takes approximately 30 ms, performed under the correlated-k and two-stream approximations using [SOCRATES](https://github.com/FormingWorlds/SOCRATES).
 [HELIOS](https://github.com/exoclime/HELIOS) [@malik_helios_2017] is a popular atmosphere model similar to AGNI, but it depends on an Nvidia GPU in order to perform radiative transfer calculations. Whilst this makes each calculation fast, it also means that HELIOS cannot be used on platforms without an Nvidia GPU or with limited resources. GENESIS [@piette_rocky_2023] has been applied to lava planet atmospheres but is closed-source and not publicly available. [Exo_k](https://perso.astrophy.u-bordeaux.fr/~jleconte/exo_k-doc/index.html) [@selsis_cool_2023] is open source and written in pure Python, but not designed to be coupled with an interior evolution model. These codes have been used to model the atmospheres of static non-evolving planets, while AGNI stands out as being the only open source code currently integrated into a comprehensive interior-atmosphere evolution framework like PROTEUS. No other models of lava planet atmospheres implement a real-gas equation of state.
 
 Coupling with PROTEUS is one primary use-case for AGNI. However, our model can also be used standalone [as in @hammond_photometric_2024] through its command-line interface and configuration files, or through Jupyter notebooks (as in the tutorials). \autoref{fig:application} below compares these two primary use-cases driving the development of AGNI.