But when you have a dozen of these battery packs doesn't the current get compounded? If the motor is pulling 48amps isn't it pulling only 4 amps from each battery?
Yes. The classic picture of spot-welded packs shows a "ladder shape" of pure nickel sheet that is laser-cut and rolled up to make it easy to ship around the world. Home builders use this because it is what the "big battery companies use" so, it must be the best, right?
At low current, pure nickel is not horrible. It works. It can be spot-welded easily on affordable welding equipment by expensive robots. Plus it is so stiff that it actually adds some rigidity to the pack.
On this style of bus, the paralleled connection "bridges" (between each cell) are the same width and thickness as the series connection bridges. If you are only looking at the electrical needs of the pack, the paralleled wires only carry a current of less than 1A, whether charging or discharging, even at very high amps...even 24ga wire would work. However, to ship batteries internationally, the style of pack must pass vibration and drop testing (to avoid shorts and fires on cargo-planes).
Also, when you go to high amps, nickel is bad. No matter how many stacks of nickel ribbon you weld on top of each other...even when you've added enough thickness to reduce max amp heat in the bus-strips...nickel is more of a resistor than a conductor at high amps. Makita uses copper bus-plates that have a thin nickel plating. It helps corrosion-resistance in harsh humid salty weather, and it makes the copper easy to spot-weld. Pure copper is hard to spot-weld...and even if you buy a very expensive plasma welder that easily welds copper buses, what do you do for corrosion resistance? Pure nickel has a price, but a thin nickel coat isn't very expensive when it's done on an industrial scale.
If you use a stout structure to hold the individual cells into the pack shape, then you don't need the buses to handle any of the packs structural rigidity requirements. Especially if its just for yourself, and you are not selling and shipping them.
For the paralleled connections, I wouldn't go smaller than 24ga, and...use whatever is convenient. Of course using large wire won't hurt, but...don't kid yourself about a fat wire helping a 1A balancing current. The series connections from one P-row to the next are where you need a certain minimum to avoid performance limits and heat.
The nickel bus in the pic above has more metal mass connecting the horizontal P-group than necessary, and the four bridges that are the
series connections (four vertical pathways) might be adequate for modest amps. If I had a pack like this and I feared the series connections were a bottleneck that was limiting the packs performance and causing a lot of waste-heat? I would overlay the existing series connections with 0.30mm thick copper foil. Its cheap, and any thicker would require shears instead of just scissors, and...as much as I don't like to see builders soldering buses onto cell ends, I don't think its a problem soldering copper foil over spot-welded nickel (as far as the heat damaging the cell electrolyte in the interior)