{The article of Marie and Loos describes a regularized $GW$ approach inspired by the similarity renormalization group second-order perturbative analysis to the linear $GW$ eigenvalue equations. The article is well-organized and the presentation is clear. I think this article can be accepted as is. Nonetheless, I do have a few minor suggestions.}
{In Eq. (45), the authors mention a reverse approach where, if I understand correctly, the $\omega$-dependent self-energy is directly modified using the SRG regularizer. How does this approach perform on $GW$50 and compare to qs$GW$ and SRG-qs$GW$?}
Yet, we decided to include this discussion in the supporting information in order not to bring confusion to the take-home message of the manuscript which is the new SRG-qs$GW$ static form.}
{I am a bit surprised that the SRG-qs$GW$ converges all molecules for $s =1000$ but not for $s =5000$. The energy cutoff window is very narrow here: $0.032$-$0.014$ Ha. Moreover, from Figs. 3, 4, and 6, the IPs are roughly converged in the order of $s =50$ to a few $100$. I think an analysis of the denominators $\Delta^{\nu}_{pr}$ for the typical molecules would be very informative. In particular, what are the several smallest denominators at the beginning and how do they change along the self-consistency procedure?}
\alert{We thank the reviewer for this suggestion. Unfortunately, we have not been able to extract useful information out of the enormous number of denominators. Indeed, the number of small denominators is large and their associated screened integrals can be very different as well, hence it is difficult to explain quantitatively why the convergence is not reached for a given value of the energy cutoff. Note that it is possible one would be able to converge the calculations for $s=5000$ by changing, for example, the DIIS parameters.}
{In Eq. (18), I think $H^{\text{od}}$ is generally not a square matrix and it is better to say $H^{\text{od}}(s)^\dagger H^{\text{od}}(s)$ instead of $H^{\text{od}}(s)^2$.}
{I think the y axis (counts in each bin) should be presented in Figs. 5 and 7. Or at least the limit of y axis should be fixed for all subplots in Fig. 5 or Fig. 7.}
{This is an excellent manuscript, which I very much enjoyed reading. In particular, it includes a comprehensive overview of the literature in the field, which I find very valuable (ref. 119 should be updated). The final result is an expression with a slightly different regularization as before, but it works well, is well-founded, and is easy to implement. I don't see arguments against it.}
The authors used the "dagger" symbol in eq. (21) and further, although they use real-valued spin-orbitals. In that case, also the matrices $W$ are real. It seems more consistent to either allow for complex-valued spin-orbitals (e.g. in eq. (8)) or only use the matrix transpose.}
%\alert{Indeed, this is inconsistent. Therefore, we have changed the definition in Eq.~(8) in order to allow for complex-valued spin-orbitals. We thank the reviewer for pointing this out.}
\alert{Although we agree with the reviewer, we would prefer to stick with the more general "dagger" notation. Allowing complex-valued orbitals would modify equations in several places in a non-trivial way. This would definitely alter the readability and accessibility of the paper which we would like to avoid.}
\textcolor{red}{\textit{''Note that a positive EA indicates a bounded anion state, which can be accurately described by the methods considered in this study. However, a negative EA suggests a resonance state, which is beyond the scope of the methods used in this study, including the $\Delta$CCSD(T) reference. As such, it is not advisable to assign a physical interpretation to these values. Nonetheless, it is possible to compare $GW$-based and $\Delta$CCSD(T) values in such cases, provided that the comparison is limited to a given basis set.''}}
