This year's bill IIRC doesn't propose any significant changes to marginal rates (other than No Tax on Tips or Overtime, neither of which will move the needle much in collections).

Yes, it does. It has significant changes to the marginal rates.

Remember, the CBO score isn't comparing what's going to happen over the next ten years with the bill to what's happened last year. It's comparing what they project will happen over the next ten years compared to what would happen if the bill weren't adopted. It's not whether 2026 will be different than 2025 - it's whether the bill will change the government's revenues and spending in 2026 compared to what would happen if the bill weren't passed in 2026. That's why they score temporary measures the way they do.

To illustrate, suppose the revenues for a given tax are currently at 100 per year (an index). You propose changing that tax permanently to 90 starting in 2026. The effect on revenue over the ten year window is -100 (-10 per year for 10 years).

Suppose instead, though, you provide that the tax change will only last three years. The effect on revenue over the ten year window is now only -30, instead of -100. That's because the bill you passed only has three years of 90, and the remaining seven years will be the same 100 that would have existed without the bill.

Now, if you get out to 2028 and near the expiration of the "temporary" tax cuts, and make them permanent, suddenly you're having a greater impact to the budget than was previously scored. Your original score only showed you reducing revenues by -30 because there were only three years of reduced revenue, but now you're reducing revenue for every year. So your second bill is changing what's going to happen in years 2029, 2030, 2031, etc. from what was originally scored. That additional ongoing reduction of revenue has to be scored in the impact of the second bill, because it wasn't part of the first bill.

And when you step back, that makes total sense. If I adopt a permanent tax cut, it has a long-term reduction in revenue. If I adopt a one-year tax cut, and then change that to a permanent tax cut after one year, it also has a long-term reduction in revenue - so you would get an irrational analysis if you only scored it as being a one-year reduction followed by zero change at all.