Well, not exactly. The developer's complaint was that the older games didn't "play the game". His position on the change in design can be better-summarised as going from "AI that tries to stop the human from winning the game" (I-IV) to "AI that tries to win the game" (V-VI). The point remains that the I-IV approach is both more efficient from a coding/manhours perspective - since its much easier to do with a bad AI - and in practice. Why they've settled on a system that takes more developer effort to code in terms of AI sophistication but produces worse results because the AI has the wrong goal is baffling. To put this in context, I actually defended this approach in Civ V and I enjoyed the fact that that game gave me some of the best late-game peaceful challenges in the series, since the AI would sometimes win and could be sneaky about it at times (I have memories of sessions where my diplo victory was actively disrupted by AIs conquering my city states as I approached victory) in what seemed targeted behaviour), so the idea of making the AI try to win is not without merit - but that should not be its primary goal. It should not, as Civ VI does, just treat the human as another AI player and ignore it in favour of rushing their own attempt at victory, since the AI is not going to win a race against anyone who isn't either inexperienced or deliberately handicapping themselves. I wasn't aware of those bonuses. The problem with a percentile bonus like that in Civ VI may be that, without something like the slider to give them science and culture from tiles and with specialist resources locked behind districts, they develop cities too inefficiently to gain very much of a boost. Also, the AI is not good at amenity management and the -% modifier from that offsets some of this bonus. Those sorts of bonuses could matter in Civ games where production was specialised, and Civ IV's paint-by-numbers approach was the most extreme in the series in that regard (build all the stuff that produces bulbs or multiplies bulbs in the same place and, voila, you have a science city). I found that simplistic and unsatisfactory as a player, but it's very AI-friendly.. Civ VI doesn't strongly reward city specialisation - generally a city with hills and a river is better than one without for pretty much everything and no individual resource can be boosted beyond the third tier. You don't have the food-settler link or specialists to make 'Great People Cities' akin to Civ IV. Look at the table in the link - the AI gets percentage-based discounts on nearly all forms of production, and since costs increase over time so does the discount in terms of individual hammers, gold etc. This is more effective than a + modifier on production, since the latter doesn't do anything meaningful in a city that has low production of that resource, but a fixed cost is always going to be lower - even in a low production city you save the same 20 hammers off a 100 hammer item that you would in a high production one. Every game in the series has done that. The Civ IV link lists the AI free techs and units it received at the start above those the player did. The issue is that in Civ VI that's most of what the AI gets (I'd previously thought it was all of it, hence this comment) - it doesn't get much that helps it once the human threatens to overcome its early advantage, and it can't capitalise on that early advantage well. There's a row in the Civ V difficulty summary for "Production free units per city". I don't recall the exact mechanism described on the Civ V forum, but it may be that the new units were spawned when the AI settled cities.