I think we're looking at the wrong problem here. The real problem here is the belief that you have to have meaningful balance between the stats to have fun. You absolutely do not.
In my experience, we give way too much min/max attention to attribute scores. They're there to describe the characters and facilitate mechanics that support the story we tell about them. If you don't think every PC you run has to be Superman, but can instead be Star Lord or Daredevil, you'll have more fun.
I've had a great time playing a character with a 14 as their highest attribute in a party with characters that "rolled" multiple 18s, and as a DM I've had a lot of low stat PCs played in my games that were highly enjoyable for their players.
First of all, you can balance around the uneven stats. A high quality magic weapon that suits the PC with low stats, or Gauntlets of Ogre Power, or a variety of other tools can magically level a playing field if you think it needs to be leveled.
Second, you don't need it leveled. The 'pipsqueak' hero is a cliche. Harry Potter wasn't the smartest, strongest, fastest (except on a broom which was only a minor benefit), charismatic or wisest of his peers - but he was thrust into the center of the story. What would his attributes have been? Anything higher than a 14? Heck, he might not have had any stat higher than a 12! That series is the most beloved story of the modern era - and the protagonist explicitly is only special because (spoilers) dead people loved him once.
Finally, we WAY overestimate how often a little difference in ability score has any impact on the game at all mechanically. Here is a simulation to run. Three low level ranger archers that are identical except for one having an 18 dexterity, another a 16, an the third a 14. They all share the same die rolls when at a "shoot a magical bear" target range (AC 11, HP 34). Using their Hunter's Mark, the first one attacks at +6 for (d8+d6+4). The second at +5 for (d8+d6+3), the last +4 for (d8+d6+2). If they all use the same die rolls, see how often there is a difference when the bear dies - not how much damage it takes - but when it dies. There is some, but not as much as you'd think based upon the huge deal we make about these differences in stats. Now, consider that a one round of difference in when the bear dies in a real combat also only matters if the bear lands an attack during that extra round(s) it survives against the lower ability score attacker....
The system works as is. I've played around with it a lot over the decades and I've always come back to the same conclusions - nothing is perfect, but the RAW are a pretty good balance of utility and simplicity.