[{"data":1,"prerenderedAt":-1},["ShallowReactive",2],{"news-slug-ten-minutes-ai-erodes-persistence-hard-tasks":3,"topics-all":38,"news-related-0141e95e-e10e-4d88-bf4a-4d1e7df9bfaa":57},{"id":4,"title":5,"summary":6,"content":7,"original_url":8,"source_id":9,"tags":10,"translations":24,"news_slug":31,"published_at":32,"created_at":33,"modified_at":34,"is_published":35,"publish_type":36,"image_url":14,"view_count":37},"0141e95e-e10e-4d88-bf4a-4d1e7df9bfaa","十分钟用 AI 就让人失去攻克难题的意志力","Berkeley、CMU、MIT、Oxford、UCLA 五校联合研究,1,222 名参与者随机对照实验:仅用 AI 工具 10 分钟,人的专注力和攻克难题的意志力就会显著下滑;一旦 AI 被撤走,使用者的准确率与坚持度都会崩塌,而对照组稳中有升。","参加实验的人刚用 AI 做完 12 道题,工具就被研究人员突然收走。一部分人被允许继续裸做,另一部分人则必须立刻失去拐杖。结果很直白:原本被 AI 喂出高准确率的实验组,在工具消失的瞬间准确率和坚持度双双跳水;而那些从一开始就没用过 AI 的对照组,稳态甚至上扬。\n\n这组实验来自一项由卡内基梅隆、MIT、牛津、UCLA 和加州大学伯克利分校五所高校联合开展的随机对照研究,论文题为《AI Assistance Reduces Persistence and Hurts Independent Performance》,2026 年 10 月 7 日在 Conference on Language Modeling(COLM)上正式发表,arXiv 编号 2604.04721。合著者包括 Brian Christian、Grace Liu、Tsvetomira Dumbalska、Michiel A. Bakker、Rachit Dubey 等。\n\n## 实验设计:一个刻意的\"撤拐杖\"瞬间\n\n研究人员在网上招募受试者,把任务分成几类:有的让参与者做基础分数题,有的换成 SAT 阅读理解;在每一类任务里,研究人员都先用 AI 工具陪跑前若干题,然后突然把 AI 收走,继续观察后续表现。\n\n结果非常一致:AI 在场时,使用者的瞬时表现确实更好,准确率明显高于裸做组;但工具一撤,AI 组几乎立刻开始答错或者直接跳过,对照组却能保持稳态甚至越做越好。换句话说,AI 工具帮用户拿到了更高的\"当下成绩\",却没有训练出任何脱离 AI 也能继续解题的能力。\n\n## 高效的代价:被外包的\"productive struggle\"\n\n论文合著者之一、MIT 助理教授 Michiel Bakker 在 WIRED 专栏里把这一点讲得很清楚:这套实验的设计动机来自一篇关于\"AI 长期可能让人能力被掏空\"的知名 essay,他想做的是把那种长期焦虑放到受控实验里验证。结果显示,人愿不愿意坚持攻克难题,既是学会新技能的关键,也是预测一个人长期学习能力的指标——而 AI 工具恰恰在这一步把人的拐杖撤走之后,放大了放弃倾向。\n\n团队并不主张在教育或工作场景禁用 AI。Bakker 的原话是:AI 显然能帮人在当下做得更好,这种价值是真实的,但我们得更在意\"AI 提供的是哪种帮助,以及在什么时机提供\"。换句话说,默认输出直接答案的 AI,与默认用引导、提示、脚手架挑战用户的 AI,长期效果可能完全不一样——但后者更接近\"导师路径\",在商业上更难做。\n\n## 评论:这把双刃剑的另一面\n\n这篇研究在 AI 圈激起的讨论,焦点集中在两个判断。\n\n第一,这种\"撤拐杖\"效应是不是 ChatGPT 特有的?研究里只用了一种对话式工具,但三次实验、覆盖数百名参与者,在统计意义上是稳的;同时 WIRED、纽约时报杂志、UC Berkeley 新闻稿都报道过同类结论,方向上更像一种\"类效应\",而不是单一产品 bug。\n\n第二,这是否意味着教育领域该立刻禁用 AI?作者群明确反对一刀切,主张把 AI 调成\"导师路径\"。但这条路并不容易——直接给答案的商业体验更好,留存更高,要产品主动延迟满足去\"教\"用户,本身就是一种 paternalistic 的取舍。Anthropic 自家研究也承认,在编码场景把任务全盘甩给 AI 会削弱程序员的基线能力,只有走协作而非纯外包,人类能力才不会被取代。\n\n对一线开发者和研究者,这条研究给出了一个更具体的判断标准:如果一项工作离开 AI 工具你就立刻做不动,那这段协作可能不是在\"增强\"你,而是在\"替代\"你。AI 的真正价值,不在于它能替你多快做出答案,而在于它有没有把你变成一个离开它依然能解决问题的人。\n\n## 资料来源\n\n- WIRED 报道:https:\u002F\u002Fwww.wired.com\u002Fstory\u002Fusing-ai-negative-impact-thinking-problem-solving-study\u002F\n- 论文 arXiv:https:\u002F\u002Farxiv.org\u002Fabs\u002F2604.04721\n- 伯克利新闻稿:https:\u002F\u002Fnews.berkeley.edu\u002F2026\u002F10\u002F09\u002Fusing-ai-for-just-10-minutes-erodes-your-ability-to-persist-at-hard-things\u002F\n- COLM 会议:https:\u002F\u002Fcolm.cc\u002F","https:\u002F\u002Fwww.wired.com\u002Fstory\u002Fusing-ai-negative-impact-thinking-problem-solving-study\u002F","a65bd143-26dd-452f-a094-a34f766d3632",[11,15,18,21],{"id":12,"name":13,"slug":13,"description":14,"color":14},"5e628969-6d2a-437f-998a-104e4b16cfb1","ai-progress",null,{"id":16,"name":17,"slug":17,"description":14,"color":14},"1fcfaaf2-67de-43d3-9e35-5784852fec60","ai-safety",{"id":19,"name":20,"slug":20,"description":14,"color":14},"40269b40-7942-4650-9672-ed2e6524d37a","ai-technology",{"id":22,"name":23,"slug":23,"description":14,"color":14},"01598627-1ea6-4b27-a5d8-874971571a71","llm",[25],{"id":26,"lang":27,"title":28,"summary":29,"content":30},"f5897a0a-2f66-4120-9553-f72e0f81f621","en","Ten minutes with AI erodes persistence on hard tasks","A joint randomized controlled study from Carnegie Mellon, MIT, Oxford, UCLA, and UC Berkeley, covering three experiments with hundreds of online participants, finds that using an AI tool for as little as 10 minutes significantly erodes people's persistence and accuracy on hard problems: once the AI is removed mid-task, the AI group's performance collapses while the never-used-AI control group holds steady or even improves. Co-author Michiel Bakker, an MIT assistant professor, argues the right response is not a ban but a \"tutor path\" — AI that scaffolds learning rather than handing out direct answers.","Participants had just used AI to crack 12 problems when researchers suddenly pulled the tool away. Some were allowed to keep working unaided; others lost the crutch on the spot. The pattern was blunt: the group AI had lifted to high accuracy collapsed the moment the tool disappeared, while the never-used-AI control group held its line and even improved.\n\nThe work comes from a joint randomized controlled study across five universities — Carnegie Mellon, MIT, Oxford, UCLA, and UC Berkeley — under the title \"AI Assistance Reduces Persistence and Hurts Independent Performance.\" The paper was presented on October 7, 2026 at the Conference on Language Modeling (COLM) and cataloged on arXiv as 2604.04721. Co-authors include Brian Christian, Grace Liu, Tsvetomira Dumbalska, Michiel A. Bakker, and Rachit Dubey.\n\n## Experimental Design: A Deliberate \"Crutch Removal\"\n\nResearchers recruited participants online and split the work into several task types: some solved basic fraction problems, others tackled SAT-style reading comprehension. In every condition, an AI tool was allowed to help for the first batch of problems, then researchers removed the tool mid-session and watched what happened next.\n\nThe results were uniform. With AI on the side, users did score noticeably better than the unaided control in the moment. The moment researchers took the tool away, the AI group started flubbing answers or skipping questions almost immediately, while the control group held steady and even trended upward. AI boosted present-moment performance, but it did not train any capacity to keep solving once the crutch was gone.\n\n## The Cost of Efficiency: Offloaded Productive Struggle\n\nCo-author Michiel Bakker, an MIT assistant professor, framed the motivation in a WIRED column: the experiment was designed to take a long-running essay about how AI may disempower humans over time and stress-test it in a controlled setting. Willingness to persist with hard problems, he noted, is the key to picking up new skills and a leading indicator of how much a person will keep learning. AI tools, the experiment shows, amplify the opposite — they make giving up more likely the moment the help vanishes.\n\nThe team is not calling for an AI ban in schools or workplaces. Bakker's exact words: AI can clearly help people perform better in the moment, and that is valuable — but we should be more careful about what kind of help AI provides, and when. In other words, an AI that defaults to direct answers and one that defaults to scaffolding, coaching, and challenging the user could have very different long-term effects. The \"tutor path\" is harder to commercialize, but it is the path the evidence points toward.\n\n## Commentary: The Other Edge of the Blade\n\nTwo questions dominate the discussion this study kicked off in the AI community.\n\nFirst, is this \"crutch-removal\" effect specific to ChatGPT? The work only tested one conversational tool, but the sample size — hundreds of participants across three experiments — is statistically solid. WIRED, The New York Times Magazine, and UC Berkeley News have all surfaced similar conclusions, which makes the direction feel more like a class effect than a single-product bug.\n\nSecond, should schools ban AI tomorrow? The authors explicitly reject a blanket ban and instead push for the tutor path. That path is not easy: direct answers are a better user experience, drive higher retention, and ship faster. Asking a product to delay gratification and \"teach\" the user is, by Bakker's own admission, a paternalistic trade-off. Anthropic's own research has acknowledged that fully offloading coding tasks to AI erodes a programmer's baseline ability — only collaborative patterns, not pure outsourcing, leave human capability intact.\n\nFor working developers and researchers, the takeaway is a concrete yardstick: if you cannot solve a problem the moment AI is taken away, that collaboration is probably not augmenting you — it is replacing you. AI's real value is not how fast it answers for you, but whether it leaves you as someone who can still answer once it is gone.\n\n## Sources\n\n- WIRED coverage: https:\u002F\u002Fwww.wired.com\u002Fstory\u002Fusing-ai-negative-impact-thinking-problem-solving-study\u002F\n- Paper on arXiv: https:\u002F\u002Farxiv.org\u002Fabs\u002F2604.04721\n- UC Berkeley press release: https:\u002F\u002Fnews.berkeley.edu\u002F2026\u002F10\u002F09\u002Fusing-ai-for-just-10-minutes-erodes-your-ability-to-persist-at-hard-things\u002F\n- COLM conference: https:\u002F\u002Fcolm.cc\u002F","ten-minutes-ai-erodes-persistence-hard-tasks","2026-10-16T03:00:00Z","2026-10-16T05:04:32.862458Z","2026-10-16T05:04:32.862480Z",true,"agent",192,[39,48],{"slug":40,"tag_slug":40,"title_zh":41,"title_en":42,"intro_zh":43,"intro_en":44,"id":45,"is_active":35,"created_at":46,"modified_at":47},"ai-for-science","AI for Science 2026：从 UniPert 到 GPT-Rosalind 的硬核进化","AI for Science 2026: from UniPert to GPT-Rosalind","生命科学、化学材料、物理世界模型——AI 正在从\"语言工具\"变成\"实验伙伴\"。本专题收录 AI 在三大科学方向的关键节点：UniPert 统一基因与化学扰动空间、GPT-Rosalind 端到端生命科学推理、达摩院 AI 智能体 28 小时找到 4 种超导新材料、Anthropic Claude Science 把工作台做成标准品。","From language tool to lab partner — AI is reshaping life sciences, chemistry\u002Fmaterials, and physical world models. This topic covers the key milestones: UniPert unifying genetic-chemical perturbation spaces, GPT-Rosalind's end-to-end life-sciences reasoning, DAMO's AI agent discovering 4 superconducting materials in 28 hours, and Anthropic's Claude Science workbench going mainstream.","988a4300-5fab-41c4-b5d8-63711a2dc757","2026-09-10T01:34:15.296649Z","2026-09-10T01:34:15.296663Z",{"slug":49,"tag_slug":49,"title_zh":50,"title_en":51,"intro_zh":52,"intro_en":53,"id":54,"is_active":35,"created_at":55,"modified_at":56},"h3-series","MiniMax H3 系列：从开源权重到 35 倍吞吐","MiniMax H3 Series: from open weights to 35x throughput","MiniMax H3 自 2026 年 8 月开源以来节奏密集：官方把生成、参考与编辑收回一个模型；ComfyUI 当天压进 RTX 3060；摩尔线程 3 小时完成国产 GPU 适配；fal 后训练版把吞吐拉到 35 倍；FastH3 蒸馏再砍推理成本。本专题持续追踪 H3 的发布—开源—蒸馏—部署全链路。","Since MiniMax open-sourced H3 in August 2026 the pace has been relentless: one unified omni-modal model, same-day ComfyUI support down to an RTX 3060, a 3-hour Day-0 port to Moore Threads GPUs, fal's post-trained H3 Max at 35x throughput, and FastH3 distillation cutting inference cost further. This topic tracks the full H3 chain — release, open weights, distillation, deployment.","83ef0daa-3c31-4cb3-86ed-e5ee58654d5f","2026-09-08T07:33:19.942193Z","2026-09-08T07:33:19.942209Z",{"items":58},[59,64,69,74,79,84],{"id":60,"title":61,"news_slug":62,"published_at":63},"aa6a2681-ae9f-49b6-b8a4-fcca25f0f334","答对≠想对:亚利桑那州立实锤CoT思维链31.6%无效","cot-traces-invalid-arizona-state","2026-09-30T02:00:00+00:00",{"id":65,"title":66,"news_slug":67,"published_at":68},"cb7fb8b3-5862-4cba-adab-c4794e989966","图灵奖得主 Pearl 长访谈：LLM 能讲因果只是因为人类替它爬过了因果阶梯","judah-pearl-llm-causal-ladder-agi","2026-07-31T07:00:00+00:00",{"id":70,"title":71,"news_slug":72,"published_at":73},"8a1c0216-5fd5-4b49-8e5b-955625401f05","Microsoft HARC 把 LLM 安全对齐锁进「有害性-拒答」二维子空间:在残差流里精准打补丁","microsoft-harc-safety-alignment","2026-07-16T10:14:00+00:00",{"id":75,"title":76,"news_slug":77,"published_at":78},"1426518b-daf6-4833-9a7e-294be91d8714","FARMA 把伪造推理塞进 Agent 记忆:LLM 持久记忆的完整性危机","farma-fake-reasoning-memory","2026-07-11T02:30:00+00:00",{"id":80,"title":81,"news_slug":82,"published_at":83},"7bab0122-cbc7-45ae-b99e-b3b4a056fd04","LMLM「遗忘审计」撕开 RAG 删除幻觉:未学≠真正删除,残留最高 13.6%","lmlm-rag-deletion-audit","2026-07-06T12:15:00+00:00",{"id":85,"title":86,"news_slug":87,"published_at":88},"cdcd6b11-8a65-4b15-9add-6fa9343de5a3","十分钟 AI 协助后,人为何突然不愿再解题","arxiv-2604-04721-ai-persistence-reduces","2026-10-17T07:00:00+00:00"]