A quick guide on how to not get destroyed right away:
Select a grassland patch somewhere on the edge without many players around, many bots and an access to a see/river. More advanced version would be to start in a center and do more diplomacy later.
Send 20% of troops right away when game starts, then move a slider in a bottom left (attach ratio) to 35% and expand every time you get 40% of your population (like 6k, 8k, 10k checkpoints). Avoid PvP.
When there are no more free land left, start conquering bots. Try to encircle them, because you will annex them without spending troops, otherwise just try to get ones located on mountains last, because there's a penalty on mountains. You'll get gold every time you finish out the nation, spend this gold on cities first, then ports and some good forts to defend yourself.
The shift from PvE to PvP must be deliberate and opportunistic. Only attack when you have an advantage like when they're busy with another war, too expanded. Get some allies because breaking an alliance gives a penalty. Don't rush to conquer as it will make you weak in a short term.
Pick a strategy for your mid-game:
City-Maxxing - this strategy involves investing heavily in Cities to achieve an enormous maximum population cap. A player with many cities can field a colossal army, aiming to overwhelm opponents through sheer numbers and a rapid troop regeneration rate. This is a land-centric, brute-force approach.
Trademaxxing - this strategy focuses on building numerous Ports to create a vast trade network. The goal is to generate immense quantities of gold, which is then used to fund a powerful navy and a large nuclear arsenal. This is a sea-centric, wealth-based approach that aims to win through economic and technological superiority.
In the late game build Missile Silos and SAMs in fortified, mountainous locations, break some alliances with a huge attach armies, or break defenses with nukes.
To win launch MIRV on a biggest threat and swarm the area
(Btw, I (OP) am not the writer of this piece, I'm just quoting it. But I am also a neuroscientist who's interested in brain preservation, for what it's worth.)
The focus of this model is to be able to do iterative editing and/or use other images as a source while the focus of that bet is to consistently one shot a specific image 9/10 times with the same prompt. Given the canyon between those two focuses I don't think so, but maybe if you had an inventive enough prompt?
You don't strictly need it, it just makes it a tiny bit more convenient since you can set it up to override DNS on any connected device, and Tailscale sets up a private VPN mesh between your devices I've come to get take for granted - a tangential feature that goes well with centrally managed DNS.
I think LLMs are usually give very simple ideas if you don't do "deep research" and just give one paper, I assume that's what you're doing? At least I saw some of my projects were not good, and I need to run more steps on deduplication/improvement
Oh, that's cool! I will have to read your article on scoring more carefully a bit later but I think that in short, I decided to take a simpler and more heuristic approach for now based on TRL (a well established metric) and a combination of some other factors. I also didn't want the user to consider the score too much as a kind of "divine word". In general I found that the heuristic used is OK for the given score plus/minus maybe 1 point.
