Thuillier et al. (Solar Phys., 2015, DOI:10.1007/s11207-015-0704-1) discuss the apparent discrepancy between the ATLAS-3 composite solar spectral irradiances (SSI) covering the ultraviolet/visible/near-infrared (NIR) spectral region with more recent SSI measurements in the NIR. Recent measurements from IRSPERAD, CAVIAR, SCIAMACHY, SOLSPEC/ISS (the SOLAR2 spectrum from 2008), and unadjusted SIM show that above about 1600 nm, SSI is lower by about 8 % with respect to ATLAS-3. A new correction is presented in Thuillier et al. (2015) that leads to SOLSPEC/ISS (SOLAR2rev) which is in better agreement with ATLAS-3. SOLSPEC/ISS SSI underwent a +10 % change from 2008 to 2010, leading to better agreement with ATLAS-3, but it remains unclear which year provided the proper radiometric level, 2008 (SOLAR2) or 2010 (SOLAR2rev) as no link to pre-launch calibration is established. Before interpreting the NIR SSI observations using our current physical understanding (constraints by the total solar irradiance and solar models), the cause for the discrepancy between the early ATLAS-3 and all recent measurements (without a-posteriori adjustments) needs to be understood considering instrument and calibration performance alone.