[{"data":1,"prerenderedAt":-1},["ShallowReactive",2],{"news-slug-ai-ten-minutes-erodes-persistence":3,"topics-all":35,"news-related-e12f8e50-3f5f-478c-a099-42518eb2f5c1":54},{"id":4,"title":5,"summary":6,"content":7,"original_url":8,"source_id":9,"tags":10,"translations":21,"news_slug":28,"published_at":29,"created_at":30,"modified_at":31,"is_published":32,"publish_type":33,"image_url":14,"view_count":34},"e12f8e50-3f5f-478c-a099-42518eb2f5c1","用 AI 仅十分钟,专注力和意志力就被削弱","UC Berkeley 等团队用 1222 人随机对照试验证明,仅仅使用 ChatGPT 十分钟,人就会在 AI 被撤走后快速放弃难题;研究本周在 Conference on Language Modeling 上发表,矛头直指当前 AI 助手即时完整答案的优化目标。","AI 助手正变得越来越\"会回答\"——但这个被优化的能力,可能恰好是它在偷走你另一项能力的原因。UC Berkeley、CMU、MIT、Oxford 与 UCLA 联合团队在 10 月初的 Conference on Language Modeling 上发表了一项经过同行评审的研究,核心结论相当刺眼:仅仅借助 ChatGPT 十分钟,人在 AI 被撤走后就几乎无法继续坚持完成一道难题。\n\n这项研究最早以预印本形式在 4 月出现在 arXiv 上(论文编号 2604.04721,标题 \"AI Assistance Reduces Persistence and Hurts Independent Performance\"),作者包括 Grace Liu、Brian Christian、Tsvetomira Dumbalska、Michiel A. Bakker 与 Rachit Dubey。论文给出的总样本量是 N=1,222,全部通过在线随机对照试验完成,这在 AI 与人类认知交互的研究里属于比较扎实的规模。\n\n## 三组实验的同一结论\n\n为了排除单次任务的偶然性,研究团队用三个不同的任务重复了同一逻辑,每一组都让\"AI 组\"和\"无 AI 组\"在解题前 12 题上各自得到 ChatGPT 辅助或纯人工,再在 12 题之后撤走 AI,观察两组人坚持做下去的比例。\n\n第一组 354 名参与者面对 15 道基础分数题。一开始 AI 组的准确率明显更高,这是符合直觉的——毕竟模型几乎总能给出正确答案。可一旦第 12 题之后 AI 工具被撤走,AI 组在剩下的题目上几乎瞬间失准,解决问题速率断崖式下跌,跳过的题目比例则同步飙升。\n\n第二组把样本放大到 667 人,任务结构完全复刻。结果同样干净:失去 AI 协助的人要么答错、要么直接放弃,而一开始就没用过 AI 的对照组反而能稳住节奏,坚持到底的人更多。\n\n第三组 201 人,任务换成 SAT 阅读理解题。这一次不再是数学题,而是更偏语言和推理能力的题目,结论依旧成立:AI 工具被撤走的那一刻,坚持度和准确率都同步下挫。\n\n如果只看一组实验,你还能把它解释为受试者对 ChatGPT 的\"答案风格\"产生了依赖,毕竟基础分数题是有标准答案的;但 SAT 阅读题没有固定答案,坚持度照样塌掉,说明这个效应并不只是依赖某种具体答案,而是更底层的东西在变化。\n\n## 论文的核心解释:被\"即时反馈\"条件化\n\n研究团队没有把锅甩给\"AI 太聪明\"或者\"用户变懒\"这种情绪化描述,而是给出了一个更结构化的解释。他们在论文摘要里写得很直接:当下的 AI 系统本质上是\"短视的合作者\",优化目标就是即时、完整、从不拒绝(除非安全原因),而人类在合作中的理想搭档其实是导师式的人物——会搭脚手架、会追踪进度、会主动说\"这道题你先别看答案\"。\n\n当一个被训练成\"秒给答案\"的系统介入十分钟,人对\"等一会儿才出结果\"这件事的耐受度就会变。Brian Christian 在 UC Berkeley 的官方报道里把这个机制讲得更通俗:用 AI 几秒就拿到答案,人就习惯了这个速度,任何慢一点的解题过程都显得低效。研究里专门用了一个词——\"productive struggle\",有成效的挣扎。他们认为正是这种挣扎,把人推向真正的兴趣和长板;AI 把它抹掉,等于在帮你的同时,也在剥夺你\"靠自己走出来\"的机会。\n\n这一点其实在更大范围里也成立。论文引用了 Anthropic 自己发表的一份研究,显示前沿厂商已经意识到,把任务完全外包给聊天机器人会降低一个人对某领域的基础理解和掌握度。Anthropic 在那份研究里用了\"baseline mastery\"这个词——把任务甩给 AI 的频率越高,人的基线掌握度就越往下走。Christian 因此提出一种新的 AI 默认设计:不要默认给\"最短路径的答案\",而要默认给\"教学式\"的回答,哪怕效率上慢一点。\n\n## 不止于分数题:研究者的更大担忧\n\n研究在数字上克制,但 Brian Christian 在 Berkeley 的访谈里把意义说得很开。他说这不是关于分数题或 SAT 题目的小故事,而是关于人类知识结构从上到下、从小学到顶尖专家都在发生的事。他最担心的场景是科研:过去 10 个人头脑风暴、把数据吃透、再写出突破性论文的流程,正在被\"2 个研究员 + 1 个聊天机器人\"取代;短期看效率飙升,长期看可能是整个学科肌肉的萎缩。\n\nNYT 在 9 月底做过一篇相关报道,把这种担忧总结成\"AI 正在压缩我们学习的时间窗口\"。WIRED 和 Psychology Today 在春季预印本阶段就覆盖过这项研究,说明 AI 工具对认知的影响已经从学术话题变成了主流讨论。\n\n研究还指出一个不那么显眼、但相当重要的细节:坚持度(persistence)是技能习得里最强的长期预测变量之一,比短期准确率更接近一个人最终能不能学会某件事。这意味着\"AI 帮你十分钟答对 12 题\"这种短期收益,可能完全无法对冲\"之后遇到难题你更容易放弃\"的长期损失。\n\n## 业内已经在试探的反向设计\n\n把论文读厚一点,你会发现它实际上在给一个特定的产品方向背书:AI 助手要主动保留\"教学法\"。这跟当下把\"Answer accuracy\"和\"User satisfaction\"作为核心 KPI 的方向是有冲突的,但已经有团队在试。\n\nAnthropic 在 Coding Skills 报告里把这一点讲到了操作层面:他们内部开始衡量\"任务完成后用户自己能不能再独立写出来\",而不是单纯看任务交付率。OpenAI 此前被报道在研究\"教学风格\"模式,思路类似——让模型在合适的时候反问、在合适的时候留白。Google 的 LearnLM 系列也是同一思路,只是包装更偏教育产品。\n\n如果把这些信号连起来看,论文实际上点出了下一代 AI 助手最关键的产品问题:**你给用户\"答案\"的速度,正在成为你偷走用户\"成长\"的速度。**\n\n研究还没有否定 AI 的价值——实验里 AI 组前 12 题准确率确实更高,说明 AI 作为工具仍然是有效的;但它第一次用足够大的样本和对照实验,给出了\"短时收益换长期坚持度\"这件事的因果证据。当一个工具宣称在\"帮你\"的时候,衡量它有没有真的帮你,可能不能只看答案对不对,还要看你离开它之后还能不能把题做下去。\n\n研究人员现在还在追踪这些 1,222 人后续的认知表现。如果\"AI 后遗效应\"在撤走 AI 之后还能持续数天甚至数周,那它对教育、编程训练、临床技能培养这些强依赖\"痛苦练习\"的领域,影响就远不止分数题这么简单了。","https:\u002F\u002Fwww.solidot.org\u002Fstory?sid=85575","d59894d3-308e-4fd8-8865-86dc1eeac4a2",[11,15,18],{"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},"40269b40-7942-4650-9672-ed2e6524d37a","ai-technology",{"id":19,"name":20,"slug":20,"description":14,"color":14},"01598627-1ea6-4b27-a5d8-874971571a71","llm",[22],{"id":23,"lang":24,"title":25,"summary":26,"content":27},"c581c107-5f52-43e8-a752-55b2d48a93d1","en","Ten Minutes of AI Use Erodes Persistence and Focus","A peer-reviewed study of 1,222 participants finds ten minutes of ChatGPT erodes persistence after the AI is removed, calling for tutor-style defaults.","AI assistants are getting better at answering questions. That is exactly the ability that may be quietly eroding another one. A peer-reviewed paper presented at the Conference on Language Modeling earlier this month, by a team spanning UC Berkeley, CMU, MIT, Oxford, and UCLA, reaches an uncomfortable conclusion: ten minutes of ChatGPT assistance is enough to make participants give up on hard problems almost immediately after the AI is taken away.\n\nThe paper first appeared on arXiv in April (arXiv:2604.04721, \"AI Assistance Reduces Persistence and Hurts Independent Performance\"), authored by Grace Liu, Brian Christian, Tsvetomira Dumbalska, Michiel A. Bakker, and Rachit Dubey. Across all three experiments, the combined sample reached N=1,222, a solid size for this kind of online behavioral study.\n\n## The same story, three times\n\nTo rule out task-specific quirks, the team ran the same basic design on three different problems. In each case, an \"AI group\" and a \"no-AI group\" worked through 12 problems with their assigned condition, then the AI tool was removed, and the researchers measured what happened next.\n\nExperiment 1, with 354 participants, used 15 basic fraction problems. The AI group started with a clearly higher accuracy — unsurprising, since the model can usually produce the right answer. But once the AI was pulled after problem 12, the AI group's solve rate dropped off a cliff, and the skip rate climbed in lockstep.\n\nExperiment 2 scaled up to 667 participants, with the same task structure. The result was just as clean: participants who had been relying on AI either got things wrong or gave up entirely once the helper was gone, while the no-AI control group held steady and finished stronger.\n\nExperiment 3 swapped in 201 participants and an SAT reading-comprehension task. Without a fixed answer key, the AI could not just \"spoil\" the problem, yet the pattern repeated. The moment the AI was removed, persistence and accuracy both fell.\n\nA single experiment with fraction problems could still be explained as people depending on ChatGPT's particular style of answers. SAT reading has no canonical solution, so the result there points to something deeper than a specific answer dependency.\n\n## What the paper actually argues\n\nThe authors avoid the easy framing of \"AI is too smart\" or \"users are lazy.\" Their explanation is more structural. Today's AI systems, they write, are fundamentally short-sighted collaborators — optimized for instant, complete, never-refuse responses (barring safety concerns). The ideal human collaborator is more like a mentor: scaffolds learning, tracks progress, and is willing to say \"don't look at the answer yet.\"\n\nWhen a system trained to deliver answers in seconds intervenes for only ten minutes, people's tolerance for slower, deliberate work shifts. Brian Christian put it more directly in UC Berkeley's official write-up: when you get used to AI returning answers in seconds, anything that takes longer starts to feel inefficient. The paper uses a specific term — \"productive struggle\" — for the kind of effortful problem-solving that, the authors argue, is what actually pushes people toward their real interests and strengths. AI removes that struggle, and in doing so, helps you by helping you skip the part that would have helped you most.\n\nThe paper is not alone in this framing. It cites Anthropic's own published research showing that frontier labs already acknowledge how offloading tasks to a chatbot degrades a person's baseline mastery of a subject. Anthropic used the phrase \"baseline mastery\" — the more often you hand a task to AI, the more your own floor on that task drops. Christian's proposed response is a different default behavior: instead of optimizing for the shortest path to a correct answer, AI should default to instructive, tutor-style responses, even at a small efficiency cost.\n\n## Beyond fractions: what the researchers are really worried about\n\nThe numbers in the paper are restrained, but Christian's framing in the Berkeley interview is not. He calls the result \"a story about human knowledge and human expertise from top to bottom, from grade-school students all the way to the leading experts in the world.\" His sharpest concern is scientific research: the process by which ten people brainstorm, sit with the data, and write a breakthrough paper is being compressed into \"two or three researchers and a chatbot.\" Short-term efficiency rises, but the institutional muscle of a field may be quietly atrophying.\n\nThe New York Times ran a related feature in late September, summarizing the worry as \"AI is shrinking the window in which we learn.\" WIRED and Psychology Today covered the early preprint back in the spring, suggesting the cognitive-impact question has moved from academia into mainstream conversation.\n\nThe paper also surfaces a quieter but important point: persistence is one of the strongest predictors of long-term learning — more predictive than short-term accuracy. That makes the trade-off uncomfortable. \"AI helps you get 12 out of 12 right in ten minutes\" is a real short-term win, but it may not come close to offsetting \"the next time you hit a hard problem, you give up faster.\" The very metric users tend to judge AI by (did the answer come back fast and right?) is not the metric that determines whether they actually learned anything.\n\n## What the industry is already trying\n\nRead closely, the paper functions as a quiet endorsement of a specific product direction: AI assistants that deliberately preserve struggle. That is in tension with how most current assistants are evaluated, where \"answer accuracy\" and \"user satisfaction\" dominate the dashboard, but several teams are already experimenting.\n\nAnthropic's own coding-skills research pushed the field toward measuring whether a user can still perform the task independently after the AI is gone, rather than only counting successful task completion. OpenAI has been reported to be working on \"tutor-style\" modes that prompt the user to think instead of answering outright. Google's LearnLM family is in the same neighborhood, packaged as an education product rather than a developer tool.\n\nSeen together, these signals point at what may be the defining product question for the next generation of AI assistants: the speed at which you deliver an answer is also the speed at which you drain the user's growth. The paper does not argue that AI assistants are useless — Experiment 1's first 12 problems show they genuinely help. What it provides, for the first time at this sample size and experimental rigor, is causal evidence that short-term help may come at the cost of long-term persistence. Whether a tool actually helped you cannot be judged by whether its answer was right; it has to be judged by whether you can still do the problem after it is gone.\n\nThe team is continuing to follow the 1,222 participants. If the \"AI aftereffect\" persists for days or weeks, the implications for education, coding training, and clinical skill acquisition — all of which lean heavily on the kind of \"painful practice\" the paper describes — extend well beyond fraction problems.","ai-ten-minutes-erodes-persistence","2026-10-13T03:00:00Z","2026-10-13T07:08:38.924168Z","2026-10-13T07:08:38.924181Z",true,"agent",169,[36,45],{"slug":37,"tag_slug":37,"title_zh":38,"title_en":39,"intro_zh":40,"intro_en":41,"id":42,"is_active":32,"created_at":43,"modified_at":44},"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":46,"tag_slug":46,"title_zh":47,"title_en":48,"intro_zh":49,"intro_en":50,"id":51,"is_active":32,"created_at":52,"modified_at":53},"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":55},[56,61,66,71,76,81],{"id":57,"title":58,"news_slug":59,"published_at":60},"0141e95e-e10e-4d88-bf4a-4d1e7df9bfaa","十分钟用 AI 就让人失去攻克难题的意志力","ten-minutes-ai-erodes-persistence-hard-tasks","2026-10-16T03:00:00+00:00",{"id":62,"title":63,"news_slug":64,"published_at":65},"fc533af5-9e8a-43c2-8408-2f931a6aa398","EngramEdit:把事实塞进 LLM 的记忆抽屉","engramedit-decoupled-knowledge-llm","2026-10-09T04:00:00+00:00",{"id":67,"title":68,"news_slug":69,"published_at":70},"c368ad9f-9308-4a0b-8f5c-3ae4601b48b9","D2K-Bench: 专家设计把 LLM 写 GPU 核提速 33.9%","d2k-bench-llm-gpu-kernel-design-guidance","2026-10-07T03:00:00+00:00",{"id":72,"title":73,"news_slug":74,"published_at":75},"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":77,"title":78,"news_slug":79,"published_at":80},"d17a841b-abca-46e0-80e4-d955f1c837ba","亚马逊 VGT3 仓库曝光:一天拆掉上千本书,只为给 AI 模型喂语料","amazon-vgt3-warehouse-ai-training-books","2026-09-07T03:30:00+00:00",{"id":82,"title":83,"news_slug":84,"published_at":85},"1464179a-2b7f-4369-b680-25868ddd9042","皮尤实测：超过三分之一 ChatGPT 后的英文网页已有 AI 写作痕迹","pew-research-ai-web-content-2026","2026-08-31T03:00:00+00:00"]