they probably ran the calculations, did a test and saw ( on the open bench) the die temperature was over 170C without, and probably 140C with the little block, which is under the magic 150C rating. So the tiny block it is then, saved 2c per unit over the larger heatsink, plus you get 100 more out of the strip of aluminium extrusion, and still get to sell all the swarf as scrap metal again.
Remember what the engineers did on the Apollo lunar rover, where they needed to dissipate heat in the motor controllers, but did not have the luxury of mass to add a big heatsink, or any way other than radiation to get rid of it, but during testing they had overheat problems and failed transistors. So all they did was to fill the controller box with paraffin wax, and rely on it melting to absorb the heat of the power devices, and then during the stops there was enough time to have the case radiate the heat away again, and for the wax to solidify again, helped by having a small reflective shield over the box so it did not get exposed to direct sun.
Bet if you went to those rovers, put in a new battery pack, whacked all the wheels to shake the vacuum welds apart and tried them they would work, apart from the odometer, which used digital logic and which might still work.