I don't think that's exactly his point. Based on available data, the game doesn't even come close to taxing the entire I/O subsystem+SSD on the PS5 or on the PC. If the game maxed out both and the PS5 had a higher theoretical max, then you'd be correct. But if say the PS5 maxes out at 20GB/s and the PC at 10 but that the game on the PS5 runs at 2GB/s and 1.5GB/s on the PC, you have to wonder why. After all, neither is anywhere near the theoretical max so they should be hitting the same speed. Severe bottlenecks that separate them should only come into play when both are near capacity which is why he claims it's the code but again, how can he prove this? That's an assumption, a logical one but an assumption nonetheless, without any data to support him.
There clearly is a bottleneck elsewhere on the PC but what is it? A line of code that needs an additional number and we're good? Or is it completely unrelated and there is a stall somewhere else in the pipeline that cannot be easily resolved?
I take issue with DF just as much as I do with Gaffers when they make an unsubstantiated claim hence why I said no one can really prove it and as you duly noted, it could be hardware, software, or something else.