Firstly your caveats are incorrect (actuation point is around 2mm, which results in about 49g pressure:
http://geekhack.org/index.php?topic=46449.0 AND you're using 1 press per second, which is practically static, whereas my calculations are based on VERY RAPID KEYPRESSES, since the OP types really fast), which results in your acceleration being incorrect for the case we are examining.
The point is that if you type VERY slowly, a heavy keycap will require less pressure to reach actuation (in fact, close to the keycap weight difference, although this will be hardly noticeable due to factors discussed below), but if you type fast it will require more due to the force required to accelerate it. The break-even point is probably around the average typing speed, which is not really all that fast.
Additionally, spring pressure reduction due to mass is not as easily felt as reluctance to accelerate due to the same incease in mass. The keys "feel" slow to move. One reason for this is
the small distance required for the spring force to increase sufficiently to cancel out the cap mass difference (0.1-0.2mm). Another is simply the way we are wired and how we detect subtle differences.
If you use a=20m/s
2 which is close to the actual use case, you arrive at a much higher force required to accelerate the cap (almost 120g for our impractical extreme case, but roughly equal to twice the cap mass in any case). The quicker you wish to depress the key, the more the increase in force, since s=ut+1/2 at
2 As t decreases, a increases parabolically and thus F increases parabolically with a.
To do this calculation properly (including the spring force increase with compression) will require calculus and I'm afraid I just don't have the time to spend doing complex calculations for a pointless argument, which should be clear to anyone with a decent education.
You could think of it this way: If 2 cars have the same engines, but one is lighter, which will accelerate quicker? Which will feel more "sluggish"?