In the most recent Dev log, you discuss when to choose between investment in future growth (Expansion, Research, Construction) vs Immediate returns (Profit, Resources, Bought Goods). The choice should be made based on the entities willingness for risk, and the current competition present.

In this context, competition means other agents that intend on impeding the achievement of the AI's goals. A fleet may have bogeys, a corporation might hove competitors, an individual may have coworkers. Any time you might not achieve your goals because someone else could get there first (you could be shot down, made bankrupt, or replaced with someone more competent) then you need to work on immediate results.

But when you are free from competition you can invest. The fleet can purchase a constructor rather than a missile boat; the corporation can construct a starbase rather than hire contract miners; the individual can take night courses rather than put in overtime.

So based on an agent's risk-taking stat and their assessment of opposition, you get a straightforward way to allocate between investment and immediate returns.

Unfortunately, 'assessment of opposition'? That's hard. If my goal is 'expansion' and your goal is 'mine all the things!', are we in opposition? I want to own the asteroid field, and you want to mine it... we could be in opposition if I already own a bunch of mining robots, or I could hire you to mine my asteroids and everything could be peachy. I don't know how composite goals like this get resolved.