Chess.com CEO：超级智能时代人类技能仍有价值Chess.com CEO: Human Skills Still Matter in the Age of Superintelligence
The Takeaway：即使机器在棋类上早已超越人类，棋类反而比30年前更受欢迎，因为人类本质上想做人类的事——与真人竞争、创造、解决问题。
Chess.com CEO Erik Allebest 从2005年以5.5万美元拍下域名起步，拒绝几乎所有VC，纯靠自有资金和盈利滚雪球，今天拥有超过2.5亿注册用户、约1000万日活、年收入超2亿美元、650人全远程团队。他指出，Deep Blue击败卡斯帕罗夫后很多人预言棋类消亡，但早期引擎让棋变得无聊，神经网络引擎（如Leela Chess Zero）却让棋重新变得激进、非常规且有趣，AI反而帮助人类更快进步：游戏复盘教练、个性化谜题、口袋教练都在加速学习曲线。
他强调没有捷径，只有重复：每天做5道谜题乘以1000天就是5000次练习。关于作弊，Chess.com投入大量资源用统计与机器学习模型检测，保护从休闲到职业赛事的诚信。对AGI/ASI，Allebest偏乐观，认为技术本身无关好坏，关键是文化与分配：希望它帮助解决气候、教育与公平问题，但也担心财富进一步集中。他总结：“人类想做人类的事。”棋类证明了即使机器更强，人们仍渴望习得并欣赏人类技能。
The Takeaway: Even after machines surpassed humans at chess decades ago, the game is more popular than ever because humans fundamentally want to do human things—compete against other humans, create, and solve problems.
Chess.com CEO Erik Allebest bought the domain out of bankruptcy for $55,000 in 2005, ignored nearly every VC who called it uninvestable, and bootstrapped to 250M+ members, ~10M DAU, $200M+ revenue and a 650-person fully remote team. Early engines made chess boring by pushing perfect play; neural-net engines like Leela Chess Zero made it aggressive and exciting again. AI now accelerates human improvement through game review coaches, personalized puzzles and on-demand pocket coaches. There are no shortcuts—only repetitions: five puzzles a day times a thousand days equals real progress. On AGI he is cautiously optimistic: the technology is neutral; the real risk is cultural and distributional. As he puts it, “Humans want to do human stuff.”
Chess.com CEO Erik Allebest 从2005年以5.5万美元拍下域名起步，拒绝几乎所有VC，纯靠自有资金和盈利滚雪球，今天拥有超过2.5亿注册用户、约1000万日活、年收入超2亿美元、650人全远程团队。他指出，Deep Blue击败卡斯帕罗夫后很多人预言棋类消亡，但早期引擎让棋变得无聊，神经网络引擎（如Leela Chess Zero）却让棋重新变得激进、非常规且有趣，AI反而帮助人类更快进步：游戏复盘教练、个性化谜题、口袋教练都在加速学习曲线。
他强调没有捷径，只有重复：每天做5道谜题乘以1000天就是5000次练习。关于作弊，Chess.com投入大量资源用统计与机器学习模型检测，保护从休闲到职业赛事的诚信。对AGI/ASI，Allebest偏乐观，认为技术本身无关好坏，关键是文化与分配：希望它帮助解决气候、教育与公平问题，但也担心财富进一步集中。他总结：“人类想做人类的事。”棋类证明了即使机器更强，人们仍渴望习得并欣赏人类技能。
The Takeaway: Even after machines surpassed humans at chess decades ago, the game is more popular than ever because humans fundamentally want to do human things—compete against other humans, create, and solve problems.
Chess.com CEO Erik Allebest bought the domain out of bankruptcy for $55,000 in 2005, ignored nearly every VC who called it uninvestable, and bootstrapped to 250M+ members, ~10M DAU, $200M+ revenue and a 650-person fully remote team. Early engines made chess boring by pushing perfect play; neural-net engines like Leela Chess Zero made it aggressive and exciting again. AI now accelerates human improvement through game review coaches, personalized puzzles and on-demand pocket coaches. There are no shortcuts—only repetitions: five puzzles a day times a thousand days equals real progress. On AGI he is cautiously optimistic: the technology is neutral; the real risk is cultural and distributional. As he puts it, “Humans want to do human stuff.”
The Takeaway: Even after machines surpassed humans at chess decades ago, the game is more popular than ever because humans fundamentally want to do human things—compete against other humans, create, and solve problems.
Chess.com CEO Erik Allebest bought the domain out of bankruptcy for $55,000 in 2005, ignored nearly every VC who called it uninvestable, and bootstrapped to 250M+ members, ~10M DAU, $200M+ revenue and a 650-person fully remote team. Early engines made chess boring by pushing perfect play; neural-net engines like Leela Chess Zero made it aggressive and exciting again. AI now accelerates human improvement through game review coaches, personalized puzzles and on-demand pocket coaches. There are no shortcuts—only repetitions: five puzzles a day times a thousand days equals real progress. On AGI he is cautiously optimistic: the technology is neutral; the real risk is cultural and distributional. As he puts it, “Humans want to do human stuff.”
查看原文 →
Chess.com CEO Erik Allebest bought the domain out of bankruptcy for $55,000 in 2005, ignored nearly every VC who called it uninvestable, and bootstrapped to 250M+ members, ~10M DAU, $200M+ revenue and a 650-person fully remote team. Early engines made chess boring by pushing perfect play; neural-net engines like Leela Chess Zero made it aggressive and exciting again. AI now accelerates human improvement through game review coaches, personalized puzzles and on-demand pocket coaches. There are no shortcuts—only repetitions: five puzzles a day times a thousand days equals real progress. On AGI he is cautiously optimistic: the technology is neutral; the real risk is cultural and distributional. As he puts it, “Humans want to do human stuff.”