文章以2025年12月26日为背景，指出AI正把“优化”从工作渗透到娱乐。早在2016年，阿尔伯塔大学用模拟器与机器学习为冰壶“最后一投”找最优策略，结论是按模型决策比奥运级选手更常赢；如今高水平队伍与国家项目广泛使用软件分析。

在牌桌与赌桌上，这种趋势更量化：研究者在1990—2000年代用博弈论与“遗憾最小化”打磨扑克AI，并在2019年由卡内基梅隆的机器人在多人无限注德州中击败职业玩家。赛马与幻想体育也被算法重塑：美国赛马场押注金额中，计算机下注联盟占20%—40%，改变赔率结构；在线拉米在印度有“数千万”真实货币玩家，顶级机器人据称因赢太多而被平台封禁。

作者借哲学家Susan Schneider提出的“智识拉平”担忧：LLM可能把每个人都推向同样的节目、目的地与游戏。即使NFL球队雇AI专家（但比赛中禁止使用），或高尔夫出现AI球童，风险仍是把判断与互动替换为更冷静的最优解。九子棋在1990年代被电脑分析出“永不输”的策略后变得乏味；而赛马赌徒Bill Benter称，过去一两年是他近20年来最兴奋的时期，显示“自动化的乐趣”需求仍在增长。

Dated Dec 26, 2025, the piece argues AI is quietly optimizing leisure, not just work. In curling, University of Alberta researchers built a simulator and ML strategy for the decisive final shot; in 2016 they found teams would win more by following the model than by relying on Olympic-level judgment, and analytics are now common in elite programs.

Games and gambling show the clearest numerical shift. After decades of work in the 1990s–2000s (game theory and regret minimization), a Carnegie Mellon bot beat professional players in multiplayer no-limit Texas Hold’em in 2019, accelerating AI’s use as both trainer and substitute opponent. In US horse racing, computer syndicates account for roughly 20%–40% of dollars wagered, reshaping odds; similar bot-driven competition appears in fantasy sports and real-money rummy played online by tens of millions in India.

The cultural concern is a “flattening” of taste and interaction: LLMs could steer users toward the same shows, destinations, and games while presenting results as personalized. Even where AI is restricted (e.g., NFL teams can study with AI but cannot use it during games), optimization can drain drama—like nine men’s morris, where 1990s computer analysis revealed near-unbeatable play. Yet demand persists: veteran bettor Bill Benter says the last 1–2 years have been his most exciting in about 20 years.