While the second could be fun, it could also be balance-destroying in the right circumstance. Frankly your first point all on its own would do a lot: Stop the scaling, and simply increase the base research cost for each Era (Especially II and III).

I'm with Romeo on this.

The era-defining projects sound a lot like a tech-oriented version of Endless Legend's Deeds, and I believe the devs have said that they plan to implement something similar to that.

I definitely agree on flattening the costs, though. The scaling system is completely broken.

DesertFoxx wrote:

Currently, there has been much debate about the design of technology tree. Basically, choosing one research over others is too punishing as science cost becomes exponential. This is not really fun as the current design reduces flexibility/replay value and goes against the theme of the game.

To address this, I propose the following design change:

1. All projects within the same era cost the same amount of Science. Sounds crazy, but my second point below addresses potential risks of this approach. Back to the point, based on this design, as your empire gets bigger, it actually gets easier to research remaining projects within that era. After all, if your empire is growing within that era, projects within that era should get easier to research. Also, this design helps you as a player to be flexible as you explore the galaxy.

2. Each era gives you four very unique research projects. You can research only one per era. These are Era-Defining projects. Ask yourself, "What is my legacy in this era?”

Era-defining projects can be researched in the background while you do your normal research. They tend to be science-expensive and is researched at the rate of X% of your science/turn. If you don't research any normal project, 100% science/turn can be devoted to era-defining projects. Therefore, depending on how quickly you want to finish era-defining projects, you may choose to finish normal projects quickly, intermittently as you adapt, or late if you need the era-defining project as soon as possible.

For example, let's say there are following four era-defining projects in the first era (you can research only one!):

• "Reach for the Star" (Exploration) - Make every single planet type colonizable. +X Ecologist Impact
• "Dust is yellower here" (Economy) - System improvement to extract dust from fine soil, which feeds the industrial complex. +1 dust & +0.5 industry per pop in all colonized planets. +X Industrialist Impact
• "The harbinger program" (Military) - Explosive Probes can be sent out by Attacker ships. Requires manual detonation and causes small HP splash damage to enemy ships within its visual range. Probes can be restocked per x turns, and once sent out, direction cannot be controlled and lasts x turns. +X Military Impact.
• "The Castle" (Influence) – System improvement to turn X% of empire influence to additional system man-power (i.e. max cap increased). Empire influence generation per turn is not reduced. +X Pacifist Impact.

This type of tech-tree design would help add more flavor to addressing the unique need of your empire while increasing flexibility and replay value.

Good ideas.

Also I propose to make it competitive: put every era a number of this "big techs" equal to players in game. Each player can go for one tech, and first to reach enough science points gets it. If you lost you have to go for another one, and as compensation for your efforts, you receive 40% of science distributed within techs not researches of this era, 40% to your new "big tech" research, and 20% is lost. 

That sounds like a tech-version of "wonders" from Civ games.  That's an interesting idea!