AI Should be Prohibited from Getting Military Techs in First 30-40 Turns

You've got a point, however if you play a game mode with a reasonable amount of pirates, or a smaller map, you might find yourself researching military sooner than you normally do.  

The AI should definitely have different research priorities for example on an insane map with low players and no pirates, versus a small map with a high number of players and abundant pirates.  I usually start to pay attention to military when I have first contact, and I KNOW its a galactic arsehole race which will quickly find me weak and ripe for conquest.  

I'm not sure how the AI weighs its tech research decisions, but for military it should be related to many factors including "which" races it encounters (or how diplomacy is developing per turn).

Example, if I declare war on a "peaceful" neighbor researching an economic line, I would expect the AI the next turn to change its research to a military focus to protect themselves from extermination.

There's 3 factors:

Personality and strategy both have weight values to determine what the race 'prefers' at any given time. Then the techs themselves have weights on which determine how valuable they are to races with those weightings. The AI then creates a random weighted array of all the techs available and rolls a virtual dice - so say thre's one tech with 10 weight and one with 50, it has a 5/6 chance of picking the weight 50 tech.

Now, the only thing in there which changes at all once the game is in action is the strategy, which AIs will change after a given number of turns. There's 5 triggers for how the AI picks a strategy: turn, victory condition, at war, tech age and personality. You can't check 'have I met someone', or 'do I have this many colonies' or 'have I bumped into some pirates'. The AI will always pick the first strategy in the list which fulfills all the valid triggers, and ignore any others which also fulfill all the triggers. This means that we can't actually tell it to anticipate a war - it will only react once it's IN a war.

The AI's early strategies in vanilla already have strong negative weight on military techs, so it generally avoids them til turn 25, and then doesn't have much bonus weight to them til turn 150ish. Marigoldran plays against Godlike AIs, so he's probably just seeing a combination of the ones that that AI gets for free at that level, and the very, very, very occasional random roll that lands military.

Thanks for the explanation.  However, a good intelligence, be it human or artificial, should have the capacity to predict.

Indeed it should. I did ask Paul if we could get some more triggers exposed in the xml for modders a few weeks back; he didn't really answer. There's a number of triggers in the hardcode which thus far aren't availalbe to us... but these don't seem to impact on which script it chooses to run atm.

I would like the ability to have the AI have more fixed (scripted) abilities that could make this happen.  

Just being able to have it check it's cash balance as a trigger, so we could throw it into 2-3 turn mini-bursts of intensive saving when it's <50 cash would be useful. One to check for hostiles within X hexes of it's border would be awesome, too - so we could tell it to quickly research military tech and crank out a few warships when it bumps into pirates. And one to check for touching cultural borders.

Just these three basic things would revolutionize what we can do with the AI. And any three of a couple of dozen more, too

I disagree that getting war tech early is useless. I agree that you should MOSTLY invest in economy and research at the start, but if you neglect war tech too much you will have to research it from scratch later on which means that if you end up in a war you will be at a disadvantage. It's worth it to research war techs even during times of peace simply so that you don't fall too much behind.

With the increased power from research techs, you will catch up, and most likely go further than if you researched weapon stuff early. The point here is you can't get punished for not researching those weapons tech early on, since planet invasions is a tech era later, so why bother? In addition, one properly outfitted modern ship can steamroll all those crappy early ships on it's own, making the investment in early weaponry totally moot (lambs to the slaughter). Hence the OP topic about no point in researching them at all. 

However, if there were some plunder mechanics or something, you might be right in your approach to strategy, but currently there isn't something like that.