Cutting into the 2-3% payload fraction is not in itself a bad thing. Cutting into the specific payload of a specific vehicle is a totally different thing.
Falcon 9 only has so much payload. It only has so much payload to the ISS and it only has so much payload to higher orbits (GTO, etc). In fact, it is a bit too small to hit the current GTO market sweet-spot (it has bad GTO/BEO performance for its size, due to its relatively low performance kerolox upper stage).
If you cut into the F9 payload capability, you hinder its ability to meet certain goals, and fulfill certain applications, that
justify its existence. A 6000 kg-to-ISS capable F9 is pretty much useless.
To maintain the same payload capability at a lower payload fraction, you now need to enlarge the vehicle. And this starts to have so many knock on effects.
You can start by stretching the stages. You might get your original payload capability back, but now you lose the potential for future growth evolutions (if further tank stretches are not feasible).
You can upgrade the engines. Oops, Merlin engines are gas generator cycle and can't be upgraded beyond a certain performance level. You'll have to shift to staged combustion cycle. Oops, Merlin isn't staged combustion. New engine. Oops, now engine testing, construction and operation is more complex and costly. Oops.
You can widen the stage to incorperate more propellant. Oops, F9 tooling is set up for a particular stage diameter. Oops, you now possibly disrupt pad handling and vehicle integration.
The vehicle gross mass is now much higher. Oops, you might need to re-engineer the launch tower to carry the higher loads. Oops, maybe now you need another engine(s). Now it isn't even an F9 anymore- rather a 'Falcon 10'. More mass, more to go wrong, more engine construction and installation costs. New thrust structure.
The issue is that SpaceX may have designed their vehicle to do a particular thing in a particular way, and that by adding this reusability strategy they will either limit performance to a point where the vehicle does not provide a useful capability, or that they will have to perform impractical modifications to the vehicle to regain this capability.
The guys at SpaceX are smart. I'm sure their plan is pretty well thought out, but it's also pretty radical. And there is a large chance that radical things do not become successful. It pays to be skeptical, but it does not pay to dismiss things just for the sake of dismissing them.
Of course, it does not pay to approach real problems with unrealistic optimism and simplicity either.
However, it is important to keep in mind that with modern materials this can probably be reduced to half this
Since when!? Where did you read that!? Where did you get the "this can be halved" figure!?
But by doing this you are making the vehicle reusable and cutting costs by a factor of 100.
My brain does this when you say "cutting costs by a factor of 100":
How many times does it have to be stated that reusability is not a universal, holy panacea of cost-savings?