Daniel Tan comments on Inference-Time-Compute: More Faithful? A Research Note