Well, what if we quantify it ?

Let's take a 170b rider non-professional rider. converted to a sane scale, that is 77.1kg. That rider has an FTP of 200.

So, a 19lb bike + a 170lb is 189 lb or 85.5kg. With an FTP of 200, that is power ratio of 2.34W/kg.

The difference between a 19lb and 15lb is 1.8kg.

Same rider on that 15lb bike is 83.7kg. Same power output, works out to 2.39W/kg.

So in terms of power to weight ratios, we are talking about a number of less than 1% (0.981) for a 170lb rider.

What happens when the rider weight changes? A 200lb rider + bike 19lb bike (99.3kg) has to produce 232 watts to maintain that 2.34W/kg. What happens when we take away 4lbs? the power ratio is no 2.38W/kg. Less benefit. Percentage, still less than 1% (0.983)

On the flip side, what happens when your rider is small? Say 125lbs + 19lb bike (60.8kg). Produces 142 watts for that 2.34W/kg base line, and when you remove that 4lb bike weight, you a 2.41W/kg. The percentage is less (0.97)

What does mean? from a pure power perspective, the difference is less than 1% in power output. What that doesn't quantify is how geometry, component, fit, and other factors play in. Those factors are what make the comparison of the numbers largely irrelevant, but yes, the raw numbers say you can 'buy' up to 1% of performance improvement in weight. You can probably buy another 1% in delivery to the pavement, and another 1% in aerodynamics. But at the end of the day, all the money in the world doesn't change the fact that the engine pushing the bike is still the engine pushing the bike .