Without going over the report yet I have some basic questions.
I did not feel the gold was "too much", but was similar to my previous experience. I did not track the net gold accumulated or spent per turn during this test (but will try to do this next game). I accumulated gold on almost all turns (31 gold on final turn to learn Sedentary Lifestyle) while averaging about 94% "earned" beakers from slider gold allocation to Science to the end of Sedentary Lifestyle, running at 100% about 70% of the turns, and running at 80% or less for about 40 turns in a row (while learning 10 techs). TD accounted for only about 6% of my technology learning (roughly equivalent to dropping the gold usage slider from 100% Science to 95%). Just a guess, but I think the 5 AIs on my continent enjoyed a roughly similar TD boost. This gold rate and positive cash flow is normal compared to my previous games, except I don't normally have streaks that long at reduced science rates. I do tend to quickly delete obsolete units and build gold generating buildings or resources when my cash flow starts getting close to negative. I was not inconvenienced in terms of Gold - for me, keeping an overall positive cash flow is one of the game challenges.
I learned Tribalism at turn 393 (about 11880 BC, after learning 68 other techs - 57 on the Fast Path plus 11 optional Techs). I averaged about 93% gold going to Science, running at 100% about 68% of the turns, and running at 80% or less for about 40 turns in a row. I just realized from your questions that there should have been a early and noticeable ramp-up in my TD science boost (increasing as I met more AIs) but now I think I missed recording it. The first TD bonus I logged was was at turn 154 for 7 beakers (almost a 50% bump to my "earned" beaker rate).
By Sedentary Lifestyle, I had 3 cities. I reached Tribalism at turn 393, built my second city at turn 409, and my third at turn 428. For each new city, I pre-build access roads and basic kick-start units (usually 2 enforcers, 2 healers, 2 story tellers, 2 merchants, 2 food merchants, a couple of defensive military units, and however many unique subdued animals I have). As a game boost, I manage my technology progress so that I can build the Captured Fire National Wonder and get a Golden Age right after founding the third city.
According to WorldBuilder, the 7 AI civilizations had between 1 and 4 cities each (average 2.3) when I reached Sedentary Lifestyle. I had met 5 AIs on our very large "Old World" continent (I think this means we were all eligible for TD bonuses). The largest AI (4 cities) was on its own small and mountainous continent, and the other AI (2 cities) on a medium-sized continent.
As for Complex Traits, I found they were interesting and challenging and generally balanced by the other Civics and growth rate changes. I did not notice any significant change to overall
or to game technology progress as I was playing , although thinking about it now I wonder if the early game was slower until getting my first Trait at turn 136, and the late game faster for the last 48 turns after taking the Scientific I trait at turn 448 ( giving me 3 positive Traits and only 1 negative Trait, versus without Complex Traits where my usual leader Elizabeth has 1 highly positive trait, 1 so-so positive trait, and 1 not-so-bad negative trait throughout the whole game).
I plan to start a new game in the next day or two, updating to the latest SVN level and keeping the same options except for switching to either Epic game speed (my preference) or to Immortal Difficulty. I saved the starting map from this game using WorldBuilder, and tested loading as a Scenario to allow measuring technology progress at a different game speed or difficulty level on the same map. I think this avoids significant measurement differences due to running on a different map with different starting locations, random map resource placement, and AI opponent styles. I think this will give a more accurate comparison specific to game speed selection or difficulty level selection (although there will still be differences due to SVN changes (can be avoided if necessary but does not make measurements up-to-date), and to game play differences to take advantage of different random events. Before I start this, I would appreciate feedback on the feasibility and value of this approach, and any related advice, suggestions or comments.