The change is to fix a layout problem: The connections are the same, but the effect of trace resistance is different.
There is quite some ripple current to charge the filter cap. This can easily reach the 5 A range, possible 10 A at full load. How much current depends on the transformer and fuse used.
Though the traces are a little wider this gives a drop, especially in the short path from the diodes and the filter cap. The voltage at the shunt is relatively low, so that 1 mV ripply would leads to some 2 mA of current ripple. I have not measured it myself, but the problem came up in a German forum: the original board showed quite some ripple (just like the current) and the changes essentially removed it.
Adding an extra copper wire to thicken the current path would be a first fix. It is not the worst point, but one that is relatively easy to fix - and for a kit that is sold for some 20 years now, it's a shame not to fix the layout, if it is so easy.
The worst point is having to much voltage at the OPs. Even if it may initially work, chances are the TL081 may not like 40 V very lond and fail. I don't know the failure mode, but failing with to high an output voltage is possible.
This circuit has been discussed over an over gain