All articles
Articles from mendoi-apps, published in English after a language check.
Search terms are limited to 120 characters.
12 articles
Most relevant first
- Shadowverse: Worlds BeyondI Gave a Brain-Dead AI Three-Turn Foresight and It Got Even Worse — What 46,900 Market-Simulation Games and a Separate AI Research Program Taught Us About “Seeing the Future but Scoring It Wrong”The original goal was painfully simple: find the strongest current Shadowverse: Worlds Beyond deck.
- Will 12,480 games make the AI smart? No. They can make the wrong answer more preciseWhen people build a card-game simulator, the first instinct is usually: run more games.
- AI Scored 99.95% on 2D Games Humans Pick Up Quickly: What ARC-AGI-3 Says About Discovering Unknown Rules—and How That Idea Could Be Used to Search for Trading EdgesAI has been strangely uneven.
- Shadowverse: Worlds BeyondAfter Letting AI Play Ten Thousand Games, the Human Only Had to Recognize the “Lion”: Diamond Aggro Nightmare, Repeat Opponents, and How Calculation DisappearedAI did the expensive search first.
- Shadowverse: Worlds BeyondLet’s Build the Strongest AI with AI — A Shadowverse WB Bot That Got Weaker with Three-Turn Foresight Now Reasons Backward from Win Lines, Loss Lines, and Proof-Scoped “Forced” OutcomesTen thousand repetitions do not turn a mistake into truth. They can create a machine that is extremely consistent at being wrong.
- Shadowverse: Worlds BeyondFrom “Storm Face” to a Research Lab — How to Build a Card-Game AI Simulator from Zero, Using Shadowverse as the ExampleA research simulator does not need to start with card art, animations, voice lines, dragging, and flashy evolution scenes.
- Shadowverse: Worlds BeyondI Became the CEO Before I Even Started the Game — Let AI Employees Play a Million Matches and Evict Card Knowledge to GitHubThe old workflow was simple: research it yourself, implement it yourself, test it yourself.
- AI cut the job by 40%. Why do we cram in 67% more work instead of going home?A sensible way to judge an AI subscription is to ask how many hours it gives back.
- I Let Mahjong AI Do Too Much and Became an “AI Operator” Who Rarely Finishes Last but Cannot Win — The Domino Hell of Chasing a Few Percent in a 25% GameIn four-player Mahjong, if all four players are equally strong, the starting point for first place is 25%. That is the first brutal fact.
- We expected “manually enter 826 cards.” Instead, the lab started appearing in half a dayA card-game simulator sounds terrifying if you imagine entering every card by hand: name, cost, stats, text, special behavior, repeat hundreds of times.
- Hunter × HunterAre Mental Leaps a Bug or a Superpower? — “Bungee-Gum Thinking”: Stretch Far, Verify, and Pull the Idea BackA conversation starts with children’s activity and suddenly jumps through handheld games, smartphones, dogs, childhood exploration, and a milkshake.
- Why do kids keep loving the same kinds of things for 100 years? Cards, ghost stories, chain letters, and gacha may share an old human pattern“Draw a ‘word you should never search’ from a gacha, then get sent to the search results a few seconds later.”