I was having a nice conversation with Ringo on Discord and we decided it would for the best to post it here. Sorry for the long text but lets go:

As HK is now, each consecutive star in the same category gives increasing amounts of fame. This has 3 effects. First, it promotes specialization: you'd rather a few things (say science, money and influence) to gain its 9 stars than spread your development amongst each category to gain 9 stars split amongst the seven (with no more than 2 stars on a individual category, for example), since the first 9 starts will give more fame than the second 9 stars. This, on a genre that historically suffers from overspecialization and yield inflation, is another step towards the same trend - it rewards getting a few things very high over spreading your growth over all the fame categories. 

Second, it favours the leader: the more ahead you are, the more you can linger in the same era to gain those high value stars at the end of the star categories. When you're behind, you'll naturally be pressured to advance as soon as possible, since advancing an era will lead to more techs and a new LT/EQ/EU, generally increasing your power level. As you advance early and forgo those high value stars (which, after you advance, you may NEVER get back), the leader solidifies its Fame lead by using that extra time it has and nabbing a few extra very high value stars. Again, this exacerbates the problems the genre historically has, as runaway leaders are a problem all 4x games have to fight against due the incremental nature of its mechanics. 

 Third, it makes harder for the AI to compete (fame-wise) with a human player. Any AI will have a hard time fighting a human with a good grasp on the game mechanics, and Humankind does have a very intelligent way to buff the AI: have it advance sorta early, with few level 3 stars. By having the AI advance early, it can ensure the AI will have the power buffs it needs to be a challenge to the player. However, this come to a cost of Fame: any player that can survive the discrepancy and stay a little longer in each era will quickly gain a huge Fame lead on the AI.

My proposal is to invert this, making each consecutive star in a category give less and less fame. The first Expansionist star should give you more fame than the second Expansionist star and the second, more than the third. This makes that you're actually rewarded for not overspecializing, and while staying in an era to get more stars will still net you more fame, the players that can manage to spread out and score stars on all categories more evenly will gain more fame than the one that specializes on a few categories. This should lower the Fame opportunity cost of advancing earlier and help the fame race end more even, with a lower lead for the first players. 

Ringo, however, has a counterpoint that it reduces player expression, especially on early game. On his opinion,

"Humankind is a streamlined game that allow you to swap every now and then between different culture, because of this aspect of the game we don't (as of yet) have very different and unique culture that totally change your way of playing the game (the closest of that are huns and mongols). That mean that the game can feel a bit samey from time to time. The only thing that is preventing two games from feeling too similar are in fact the level 3 era stars and the amount of fame they give. Because 3-stars era stars are important, you try to get them, that force you to prepare a game-plan and be flexible. For example if you play Egyptians in Ancient era you'll have a totally different game than if you play Babylonians or Harappans, and that's because the first economy you build on will be different and you'll need to build your other economy as the game goes on and that's if the AI or someone doesn't get the culture you wanted to take, in which case you'll need to improvise. Current system give some amount of skill expression. If you nerf 2/3-stars era stars then you decrease the skill expression and end up with a bland ancient era where everyone play different cultures the same way and it will change the feel of the whole game going forward. "

 I think that there are pros and cons both to have increasing returns and diminishing returns to the era star values. It all depends on designer intent, which is something I cannot judge as I am not one of the designer nor I have access to the design documents of the game. Increasing returns promotes being a specialist, diminishing returns promote being a generalist.  IMHO, Humankind has two interconnected "ratings" for each empire. Power and Fame. 

Power is the ability to do more things and gain more bonuses - its your tech, your cities, your military, your income, your pops, so on and so forth. This side of the game is common to all 4x games and well known, not only by veteran players but also by the devs. Fame is the victory condition, and here lies the one of the main differences between HK and most other games in its genre. While the conventional gameplan of most 4x is to directly translate Power into victory, HK creates another layer that interacts with the Power layer, making the power to victory relationship not so direct and straightfoward. Right now, as most other 4x, the Power layer promotes specialization and has a "winner takes all" bias. The more things (tech, city, pops, military, etc) you have, the more things (yada yada) you gain. This is expected and feeds into the power fantasy of games like this, so there aren't much HK can do to help the underdog compete (other than prevent glaring exploits) or it'll deliver a much worse experience. 

However, due to the positive marginal Fame gain of high tier stars, the Fame mechanic is ALSO pushing in the same direction, favouring the leader and promoting specialization. Giving the high tier stars a negative marginal Fame gain would give the game a nudge in the opposite direction, making it, IMHO, lower the "runaway leader" effect and giving more opportunities of close Fame races and make the game better, IF the designer intent is to promote a more generalist approach to building your empire instead of overspecializing. 

Anyway, this change should be quite easy to code (it is simply to increase the value of tier 1 - bronze - stars and increase the value of tier 3 - gold - stars). So since we probably won't have another OpenDev before launch, what I'm proposing right now is:

Test a build with the VIPs and internal QAs where the Fame gain per Star favor the first stars instead of the last ones to see if it leads to a more satisfying game.

I think making this change would make the game better and less prone to runaway leaders, with both more chance for a player to lose 1st place and more chance for a player to overtake the AI when they're behind, but only playtesting can say that for sure. 

IT VERY GOOD