The population based science penalty is not a restriction. It does not limit your play style. It is a tradeoff whereby you can get a better science progression curve by holding back somewhat on how quickly you expand your empire. This is the opposite of a restriction. It introduces meaningful choice into the game.
It isn't as meaningful as you seem to think. The tech research rate plot isn't very interesting. Here:
Population is the X axis, increasing to the right.
Time to research is the Y axis.
This exact line is 1-1000 pop, researching a 3880 cost tech using 16% of the population X working Labs of this tech's type, and 100% bonuses in research. The limit of the research time is 41 months. Formally, limit time is (Research cost * 0.02) / (this tech avg lab output * % pop working those labs * research speed multiplier).
As you can see, the line quickly approaches a limit. At 50 pop, you'll be running about twice the limit-research-time, while at 250 pop, you'll be hitting 110% -120% of the fastest time.
So, you can see that if you keep a fixed percentage of your population working labs, your research speed stops improving. If you increase the output of labs, increase the percentage of pop working those labs, or increase your research speed bonuses, the line moves up or down proportionally (a 10% bonus moves the line down 10% time; doubling the bonus halves the time, etc.)
TL;DR - For a fixed percentage of POP working labs, the difference between 50 pop and 100 pop is big, but the difference between 250 POP and 2500 POP is minor. OTOH, mineral and energy production are absolute, not proportional, so the difference between 250 POP and 2500 POP is 10x. By endgame, it's all about the bonuses - And since a small empire will have the same bonuses as a large one with the same ethos and traits, so the large empire is unconditionally better.
TL;DR's TL;DR - Staying small requires sacrificing mineral and energy production, while going big allows you to have more minerals and energy without losing research.