Possible wording mistake in the paper

[ltsabadz]

Hi, I was reading your paper on arxiv and noticed what might be a contradiction in the description of noise vs sample prediction difficulty on SNR levels (page 8).

While sample prediction is easy at high signal-to-noise ratios (SNR) and difficult at low SNR, noise prediction is the exact opposite:
easy at high SNR and hard at low SNR.

I think you meant hard at high SNR and easy at low SNR for noise prediction. It is a minor thing but if you are going to have a v2 revision you can fix this also.

Thanks for the work!