# 20 Things Smart People Are Learning With AI Right Now While Everyone Else Waits.
**作者**: NeilXbt
**日期**: 2026-04-18T08:18:35.000Z
**来源**: [https://x.com/neil_xbt/status/2045416688808554571](https://x.com/neil_xbt/status/2045416688808554571)
---
There is a group of people who are not waiting.
They are not waiting for AI to mature. They are not waiting for better tools. They are not waiting for the right course to arrive or the right moment to feel ready.
They are using what exists right now, today, to learn things that used to take years, and they are getting a compounding head start that will be very difficult to close later.
The people still waiting have a list of reasons. They are watching to see how things develop. They are not sure which skill to focus on. They are busy with what is in front of them. They intend to start next month.
Next month, the gap will be larger.
This is not a list of the most exotic AI applications or the most technically impressive things being built in research labs. This is a list of practical things that real people are learning with AI right now that produce real returns, skills that pay, perspectives that change decisions, capabilities that used to require expensive courses or years of experience to develop.
Here are 20 of them.
1. How to Write in a Way That Actually Persuades People
Most people write to express what they think. Persuasive writing is different. It requires you to understand what the reader already believes, what their objections are before they voice them, and how to sequence information so that each sentence makes the next one easier to accept.
Smart people are using AI to study this in real time. They paste their writing in, ask Claude to identify where the logic drops, where a reader would push back, where the argument assumes too much. They ask for the three weakest sentences. They ask for a rewrite of the opening that would keep a skeptic reading.
This is an editorial feedback loop that used to cost money or require a skilled friend who was also a writer. It is now available in every draft.
2. How to Understand Financial Statements
The ability to read a balance sheet, understand what the income statement is actually telling you, and interpret a cash flow statement without getting lost is one of the most practically useful skills in business. Most people never develop it because the resources for learning it either assume you already know or bury you in jargon.
AI can explain a P&L line by line, in plain language, as many times as you need, at exactly your level of understanding. You can paste in a real company's annual report and ask Claude to walk you through what the numbers mean, which ones matter most, and what would concern a sophisticated investor.
This skill used to take a business degree or a very patient CFO mentor to develop. It now takes a few focused sessions with AI and a willingness to ask questions that would feel embarrassing in a room full of people.
3. How to Learn a New Language Faster Than Any App Teaches It
Duolingo teaches you to say "the cat drinks milk" in German. What it does not teach you is how to sound like a person rather than a textbook, how to understand fast native speech, or how to have a real conversation about something that matters to you.
Smart people are using AI as a conversation partner in their target language, asking it to correct their grammar, explain why a phrase sounds unnatural, teach them the idioms that native speakers actually use, and simulate real-world conversations around their specific interests and profession.
The feedback loop is immediate. The patience is infinite. And the ability to practice at two in the morning when you have twenty minutes and no language partner available makes the consistency that language learning requires actually achievable.
4. How to Think Through Complex Decisions Without Bias
Every major decision you make is shaped by biases you are not aware of. Confirmation bias. Availability bias. Sunk cost fallacy. The tendency to overweight recent events. The tendency to underweight base rates in favor of specific details.
AI gives you a thinking partner that can name the bias operating in your reasoning, steelman the position you are dismissing, and force you to engage with your assumptions rather than act on them unchecked. You describe the decision, explain your current thinking, and ask Claude to tell you what you are probably getting wrong.
This is not about outsourcing judgment. It is about having a mirror that reflects back the structure of your thinking so you can see it clearly rather than just feeling your way through it.
5. How to Build Something Online Without a Technical Background
A working landing page. A functional tool. A prototype that demonstrates an idea. A simple automation that saves ten hours per week. These things used to require a developer, which meant either money or a co-founder or months of learning to code.
Smart people are using AI to build real things without waiting. They describe what they want in plain language. They work through the problems that come up. They ship something that works, get feedback from real users, and iterate.
The gap between "I have an idea" and "I have a working version" has never been smaller. The people closing that gap are not waiting for perfect technical skills. They are learning the skill of working with AI to build, which is its own capability and one that compounds quickly.
6. How to Research Any Topic at a Professional Level
Shallow research produces shallow conclusions. Most people read the first three articles that appear in a search, form an opinion, and move on. Professional-quality research means understanding the primary sources, knowing what the actual debate is among experts, identifying where the evidence is strong versus where it is contested, and being able to articulate the steelman version of positions you disagree with.
AI compresses the time it takes to develop that depth without compressing the depth itself. You can go from "I have heard of this topic" to "I understand the main arguments, their strongest evidence, and the key objections" in a single focused session, which changes the quality of every decision, conversation, and piece of work that touches that topic.
7. How to Negotiate
Most people negotiate by stating a position, hearing a counteroffer, and compromising somewhere in the middle. Skilled negotiators know this is the most expensive way to negotiate.
Real negotiation skill involves understanding what the other party actually values, identifying the interests underneath stated positions, knowing which concessions cost you little but matter to them, and recognizing when to hold and when to move. These are learnable, specific skills and AI can teach them through case studies, role-plays, and real-time feedback on your reasoning.
You can describe a real negotiation you are preparing for, explain both parties' positions, and ask Claude to identify the leverage you are not using, the assumptions you are making that might be wrong, and the three most likely ways it goes badly. This kind of preparation used to require a coach.
8. How to Speak in Public Without Fear
Public speaking anxiety is almost never about speaking. It is about the combination of uncertainty, perceived judgment, and the gap between how you sound in your head and how you fear you sound out loud.
Smart people are using AI to prepare for specific speeches and presentations in a way that reduces uncertainty, which is the root of the anxiety. They practice their opening ten sentences until they know them cold. They work through every question they might receive. They ask Claude to identify the moment in their outline where a skeptical audience would check out.
Removing uncertainty through preparation is the most reliable path to composure under pressure. AI makes that preparation faster and more comprehensive than any other tool available.
9. How to Build Systems That Run Without Constant Attention
A system is any set of processes where doing the setup work once produces results repeatedly without proportional ongoing effort. Email sequences. Content pipelines. Client onboarding flows. Recurring reports. Standard operating procedures that survive the absence of the person who built them.
The ability to think in systems, to see where repeated effort exists and replace it with a process, is one of the highest-value skills in any professional context. AI makes it faster to design these systems, document them clearly, and build the automation layer that makes them run.
People who develop this skill are not just more productive. They are building leverage that compounds, because every system they create returns time that can be invested in the next one.
10. How to Give and Receive Feedback Without Damaging Relationships
Most people avoid giving honest feedback because they are afraid of how it will land. Most people receive feedback defensively because it arrives without enough context to understand the intent. Both problems make organizations, teams, and relationships worse over time.
AI can help you learn to frame difficult feedback in a way that is specific, actionable, and delivered with genuine care for the other person's growth. It can also help you understand how to receive feedback without the emotional shutdown that makes people stop giving it.
These are communication skills that used to get learned through trial and error, usually with real relationship damage along the way. They are learnable much faster through deliberate practice with AI, which has no ego in the outcome.
11. How to Understand the Technology Reshaping Your Industry
Every industry is being changed by AI and most people in those industries understand this at a headline level without understanding it at a practical level. They know "AI is affecting healthcare" or "AI is changing legal work." They do not know which specific tasks are being automated first, which roles are becoming more valuable, and what the person who thrives in their industry in ten years knows that they do not know yet.
AI can give you that specificity. You can ask Claude to explain exactly how AI is being deployed in your field, which jobs it is replacing versus augmenting, what skills the people thriving in your industry in five years will have, and what you should be learning now to be in that group rather than watching it from outside.
12. How to Write Code Well Enough to Use It
Not enough to be a professional developer. Enough to automate repetitive tasks in your own workflow, build simple tools that solve your own problems, and understand what engineers are telling you when they explain technical constraints.
This level of coding fluency is more achievable now than it has ever been, because AI explains every line you do not understand and helps you debug in real time. You do not need to learn syntax before you can build something that works. You learn syntax by building things that work and asking why they work.
The people developing this skill are not trying to become developers. They are trying to stop being limited by what they cannot build themselves.
13. How to Manage Your Own Psychology
The research on what actually drives performance, wellbeing, and sustained motivation is substantial and almost entirely ignored in the way most people approach their own lives. They manage their time but not their energy. They set goals but not the conditions that make achieving them likely. They address symptoms without understanding the systems producing them.
AI gives you access to the practical implications of that research in a way that is personalized to your actual situation. You describe what is happening, how you are responding to it, and what you want to be different, and you can explore the cognitive patterns, behavioral interventions, and environmental changes that are most relevant to your specific case. This is not therapy. It is evidence-based self-knowledge, applied.
14. How to Run Meetings That Produce Decisions
Most meetings produce no decisions. They produce conversation, alignment-building, status updates, and vague agreement to think about the topic more and discuss it again next week. The people who know how to run meetings that end with clear decisions, clear owners, and clear timelines are rare and valued everywhere they work.
This is a learnable skill and AI can accelerate the learning substantially. You can study meeting design, prepare agendas that force decisions rather than enable deferral, and get real-time feedback on why specific meetings in your experience went nowhere.
15. How to Build an Audience Around What You Know
Every person who has developed real expertise in something has knowledge that other people would pay attention to. Most of them never share it because they do not know how, do not have the confidence, or do not understand that the bar for valuable content is much lower than the bar they are setting for themselves.
Smart people are using AI to bridge that gap. Not to replace their voice but to help them find it. To turn their half-formed thoughts into coherent pieces. To understand which of their ideas will resonate with the people they want to reach. To build a consistent output without the creative overhead that stops most people before they start.
16. How to Understand Data Well Enough to Question It
Statistics can be used to support almost any conclusion. The people who know how to examine a study, identify the assumptions baked into a dataset, and recognize when a percentage is being used to mislead are functionally operating in a different reality than the people who accept numbers at face value.
AI can teach this skill through specific examples. You paste in a statistic, a study summary, or a chart, and ask Claude to identify what questions a statistically sophisticated reader would ask about it. After enough of these sessions, you start asking those questions automatically.
17. How to Solve Problems Systematically
Most people approach problems the way they always have, by doing the first thing that comes to mind, adjusting when it does not work, and continuing until something sticks. This produces solutions but rarely the best ones, because the first solution that comes to mind is almost never the one that would survive careful analysis.
Systematic problem-solving is a set of specific tools: defining the actual problem before jumping to solutions, separating symptoms from causes, generating options before evaluating them, and stress-testing proposed solutions against the conditions most likely to break them. AI can walk you through this framework on real problems you are facing, which is how you actually learn it rather than just knowing it exists.
18. How to Build Genuine Confidence
Not performed confidence. Not the kind that collapses under pressure or requires constant validation. The kind that comes from accumulated evidence of your own competence, from understanding your actual strengths and weaknesses clearly rather than through the distorted lens of other people's opinions or your own anxiety.
This is not something AI builds for you. But AI can accelerate the process by helping you reflect honestly on your experience, identify patterns in what you do well, and build the skills that produce the competence that confidence is actually built on.
19. How to Write Proposals That Win
A proposal is a sale in document form. Most proposals lose not because the work being proposed is not good but because the document focuses on the deliverables instead of the transformation. It describes what will be done rather than what will be different as a result.
Smart people are using AI to understand the structure of proposals that win, practice writing them against specific scenarios, and get feedback on where their current proposals are selling the process rather than the outcome. This skill pays directly and immediately in anyone who ever needs to convince another person or organization to say yes to something.
20. How to Learn Anything New in Half the Time
This one is the meta-skill underneath all the others.
Most people learn by reading linearly and hoping comprehension accumulates. The most effective learners engage differently. They start with the big picture structure of the subject before filling in details. They actively test their understanding rather than passively reviewing material. They teach back what they have learned to expose the gaps. They seek the specific type of example that makes an abstract concept concrete.
AI can be configured to teach in exactly this way, building learning experiences around your existing knowledge, testing comprehension in real time, generating the specific examples that connect new ideas to things you already understand, and making explicit the structure of a new subject that would otherwise take months of immersion to perceive.
The people who develop this meta-skill are not just learning faster. They are building a capability that accelerates every area of learning they turn it toward for the rest of their lives.
## The Reason to Start Now
There is a window in which learning these skills with AI produces disproportionate returns. That window exists because the people who learn now build advantages while most people are still deciding whether to start.
It will not stay open indefinitely. The tools will become more mainstream. The skills will become table stakes rather than differentiators. The first-mover advantage will compress.
None of that means starting later is fine. It means starting now is better than starting later, and starting later is better than not starting at all.
Pick two from this list. The two that produce the most return in your specific life, in the work you do, in the goals you are building toward.
Use AI to develop them deliberately, not casually, over the next 90 days.
The people who are learning these things right now are not exceptionally talented.
They are not technically sophisticated.
They are not waiting for the right moment.
They just decided to start while everyone else was still watching.
## 相关链接
- [NeilXbt](https://x.com/neil_xbt)
- [@neil_xbt](https://x.com/neil_xbt)
- [5.9K](https://x.com/neil_xbt/status/2045416688808554571/analytics)
- [Upgrade to Premium](https://x.com/i/premium_sign_up)
- [4:18 PM · Apr 18, 2026](https://x.com/neil_xbt/status/2045416688808554571)
- [5,964 Views](https://x.com/neil_xbt/status/2045416688808554571/analytics)
- [View quotes](https://x.com/neil_xbt/status/2045416688808554571/quotes)
---
*导出时间: 2026/4/19 11:09:56*
---
## 中文翻译
# 聪明人正在用 AI 学习的 20 件事，而其他人还在观望。
**作者**: NeilXbt
**日期**: 2026-04-18T08:18:35.000Z
**来源**: [https://x.com/neil_xbt/status/2045416688808554571](https://x.com/neil_xbt/status/2045416688808554571)
---
有这样一群人，他们没有在等待。
他们没有在等 AI 变成熟。没有在等更好的工具出现。没有在等一门合适的课程、或等到自己感觉准备好的那一刻。
他们正在利用当下已有的东西，去学习那些过去需要花上多年才能掌握的东西，并且正在积累一种后期很难被追赶上的复利式领先优势。
仍在观望的人有一堆理由。他们在观望事情会如何发展。他们不确定该专注哪项技能。他们忙于眼前的事务。他们打算下个月再开始。
下个月，差距只会更大。
这并不是一份关于最奇特 AI 应用、或研究实验室里技术最惊艳成果的清单。这是一份实用的清单，列出真实的人此刻正在用 AI 学习、并能产生真实回报的东西——能带来收入的技能、能改变决策的视角、过去需要昂贵课程或多年经验才能培养的能力。
以下是其中的 20 项。
1. 如何写出真正能说服人的文字
大多数人写作是为了表达自己的想法。有说服力的写作则不同。它要求你理解读者已经相信什么，在他们说出口之前预判他们的反对意见，以及如何安排信息的顺序，让每一句话都为下一句的接受铺路。
聪明人正在用 AI 实时研究这一点。他们把自己的文字粘贴进去，让 Claude 找出逻辑在哪里断裂、读者会在哪里反驳、论证在哪里预设得过多。他们要求指出最薄弱的三句话。他们要求重写开头，让一个持怀疑态度的人愿意继续读下去。
这在过去需要花钱请编辑，或需要一个既是朋友又是写作者的人来反馈。现在，每一次草稿都能获得这样的反馈循环。
2. 如何看懂财务报表
能够读懂资产负债表、理解利润表真正在告诉你什么、以及解读现金流量表而不迷失方向，是商业中最实用的技能之一。大多数人从未掌握它，因为相关的学习资源要么假设你已经懂了，要么让你淹没在术语中。
AI 可以用通俗的语言，逐行解释损益表，按你需要的次数反复讲解，并且完全贴合你的理解水平。你可以粘贴一份真实公司的年报，让 Claude 带你理解这些数字的含义，哪些最重要，以及哪些地方会让一位资深投资者感到担忧。
这项技能过去需要商科学位或一位非常有耐心的 CFO 导师才能养成。现在，只需要与 AI 进行几次专注的对话，以及愿意提出那些在满屋子人面前会觉得尴尬的问题。
3. 如何比任何 App 都更快地学会一门新语言
多邻国教你用德语说“猫喝牛奶”。但它不会教你如何说话像一个人而不是一本教科书，如何听懂语速很快的母语者，或如何就你真正关心的话题进行一场真实的对话。
聪明人正在把 AI 用作目标语言的对话伙伴，让它纠正自己的语法、解释为什么某个短语听起来不自然、教他们母语者真正使用的习语，并围绕他们特定的兴趣和职业模拟真实世界的对话。
反馈是即时的。耐心是无限的。而且，能在凌晨两点、只有二十分钟、又没有语言伙伴可用时进行练习，这让语言学习所必需的坚持变得真正可以实现。
4. 如何不受偏见影响地思考复杂决策
你做出的每一个重大决策，都受到你未曾察觉的偏见的影响。确认偏误。可得性偏误。沉没成本谬误。过度看重近期事件的倾向。用具体细节压倒基础比率的倾向。
AI 给了你一个思考伙伴，它能指出你推理中正在起作用的偏见，为你要否定的立场提出最强辩护，并迫使你正视自己的假设，而不是不加检验地直接据此行动。你描述这个决策，解释你当前的想法，然后让 Claude 告诉你，你可能在哪些地方想错了。
这不是把判断力外包出去。而是拥有一面镜子，映照出你思考的结构，让你能清晰地看见它，而不是仅凭感觉摸索前行。
5. 如何在没有技术背景的情况下在网上做出一个东西
一个能用的落地页。一个实用的工具。一个能展示想法的原型。一个每周节省十小时时间的简单自动化。这些东西过去都需要一名开发者，这意味着要么花钱，要么找联合创始人，要么花几个月学编程。
聪明人正在用 AI 直接动手，而不是等待。他们用通俗的语言描述自己想要什么。他们逐一解决过程中出现的问题。他们发布一个能用的东西，从真实用户那里获得反馈，然后迭代。
“我有个想法”和“我有个能运行的版本”之间的鸿沟从未像今天这么小。那些正在跨越这道鸿沟的人，并没有在等待完美的技术功底。他们在学习与 AI 协作构建的技能——这本身就是一种能力，而且能快速复利增长。
6. 如何以专业水准研究任何主题
浅薄的研究产生浅薄的结论。大多数人只读搜索结果里排在前三的文章，形成一个观点，然后就结束了。专业水准的研究意味着理解一手资料、了解专家之间真正的争论点在哪里、识别证据在哪里坚实可靠而在哪里存在争议，并且能够清晰地阐述你所不同意的立场中最有力的版本。
AI 能压缩培养这种深度所需的时间，却不会压缩深度本身。你可以在一次专注的会话中，从“我听说过这个话题”进阶到“我理解主要论点、它们最有力的证据以及关键的反驳意见”，这会改变每一个涉及该话题的决策、对话和工作成果的质量。
7. 如何谈判
大多数人谈判的方式是：陈述一个立场，听取对方的还价，然后在中间某个地方妥协。熟练的谈判者知道，这是成本最高的谈判方式。
真正的谈判技巧包括：理解对方真正重视什么、识别表面立场之下的利益诉求、知道哪些让步对你成本很低却对对方很重要，以及懂得何时坚持、何时让步。这些是可以学习的具体技能，AI 可以通过案例研究、角色扮演和对你的推理的实时反馈来教授它们。
你可以描述一场你正在准备的真实谈判，说明双方的立场，然后让 Claude 指出你没有使用的筹码、你可能做出的错误假设，以及最有可能导致谈判失败的三种情形。这种准备工作以前需要一位教练。
8. 如何毫无恐惧地进行公众演讲
公开演讲焦虑几乎从来不是关于演讲本身。它源于不确定性、感受到的评判，以及你脑中自己的声音与你担心自己实际发出的声音之间的落差。
聪明的人正在用 AI 以降低不确定性的方式为特定的演讲和展示做准备，而不确定性正是焦虑的根源。他们练习开场的前十句话，直到烂熟于心。他们推演可能收到的每一个问题。他们让 Claude 找出大纲中持怀疑态度的听众可能会失去兴趣的那个时刻。
通过准备消除不确定性，是在压力之下保持镇定最可靠的途径。AI 让这种准备比任何其他可用工具都更快、更全面。
9. 如何构建无需持续照看的系统
系统是指任何这样一套流程：一次性的搭建工作能够反复产生成果，而不需要成比例的持续投入。邮件序列。内容流水线。客户入职流程。周期性报告。即使创建者不在也能照常运转的标准操作程序。
以系统的方式思考的能力——看到哪里存在重复性劳动并用流程取代它——是任何职业场景中价值最高的技能之一。AI 让设计这些系统、清晰地记录它们、以及构建让它们运转的自动化层都变得更快速。
培养这项技能的人不仅效率更高。他们还在积累会复利的杠杆，因为他们创建的每一个系统都会释放出时间，而这些时间又可以投入到下一个系统的构建中。
10. 如何在不伤害关系的情况下给予和接受反馈
大多数人回避给予诚实的反馈，因为害怕反馈会造成的影响。大多数人带着防备心接受反馈，因为反馈到达时缺乏足够的背景来理解其意图。这两种问题都会让组织、团队和关系随着时间的推移变得更糟。
AI 可以帮助你学会以具体、可操作、并且真正出于对对方成长的关怀的方式来表达棘手的反馈。它还可以帮助你理解如何在不产生情绪封闭的情况下接受反馈——正是这种情绪封闭让人们不再愿意给出反馈。
这些沟通技能过去只能通过反复试错来学习，通常还会在这个过程中造成真实的关系损伤。而通过与 AI 进行刻意练习，可以更快地学会它们，因为 AI 对结果没有自我自尊心。
11. 如何理解正在重塑你所在行业的技术
每个行业都在被 AI 改变，而这些行业中的大多数人对它的理解只停留在标题层面，而非实践层面。他们知道“AI 正在影响医疗”或“AI 正在改变法律工作”。但他们不知道哪些具体任务会最先被自动化、哪些角色正变得更有价值，也不知道十年后在他们行业中如鱼得水的人知道哪些他们尚不知道的东西。
AI 可以为你提供这种具体性。你可以让 Claude 详细解释 AI 在你的领域是如何被部署的、它正在取代哪些工作而不是增强哪些工作、五年后在你行业中成功的人将具备哪些技能，以及你现在应该学什么才能成为那个群体的一员，而不是站在外面旁观。
12. 如何把代码写到够用的程度
不是达到专业开发者的水平，而是足以自动化你自己工作流程中的重复任务、构建能解决你自己问题的简单工具、以及理解工程师向你解释技术限制时所说的内容。
这种程度的编码流利度现在比以往任何时候都更容易实现，因为 AI 会解释每一行你不理解的代码，并帮助你实时调试。你不需要先学会语法才能构建出能运行的东西。你是在构建能运行的东西并追问它们为何能运行的过程中学习语法的。
培养这项技能的人并不是想成为开发者。他们只是想不再被自己无法亲手构建的东西所限制。
13. 如何管理自己的心理状态
关于什么真正驱动表现、幸福感和持续动力的研究相当丰富，但绝大多数人在对待自己生活的方式上几乎完全忽视了这些成果。他们管理时间却不管理精力。他们设定目标却不创造使目标得以实现的条件。他们处理症状却不理解产生症状的系统。
AI 让你能够以贴合自身实际情况的方式，获取这些研究的实际应用启示。你描述正在发生的事情、你的应对方式以及你希望改变什么，就可以探索与你具体情况最相关的认知模式、行为干预和环境调整。这不是心理治疗。这是有循证依据的自我认知，付诸应用。
14. 如何主持能产出决策的会议
大多数会议产出不了任何决策。它们产出的是谈话、建立共识、状态汇报，以及“再多想想、下周再讨论”这种模糊的同意。懂得如何主持以明确决策、明确责任人和明确时间线收尾的会议的人，凤毛麟角，且无论在哪工作都备受重视。
这是一项可习得的技能，而 AI 能大幅加速学习过程。你可以研究会议设计，准备能促成决策而非纵容拖延的议程，并针对你经历过的某些会议为何毫无结果获得实时反馈。
15. 如何围绕你所知的事物建立受众
每一个在某个领域拥有真正专长的人，都有其他人会关注的知识。但大多数人从不分享，因为他们不知道怎么分享、缺乏信心，或者不明白有价值内容的门槛远低于他们为自己设定的标准。
聪明的人正在用 AI 来弥合这一差距。不是为了取代他们的声音，而是帮助他们找到自己的声音。把半成形的想法变成连贯的文章。理解他们的哪些想法能引起目标受众的共鸣。在没有创意负担压垮大多数人的情况下，保持稳定的输出。
16. 如何充分理解数据，从而能够质疑它
统计学几乎可以用来支持任何结论。懂得审视一项研究、识别数据集中隐含的假设、并识别百分比何时被用来误导人的人，与那些照单全收数字的人，实际上生活在不同的现实之中。
AI 可以通过具体例子教你这项技能。你粘贴一条统计数据、一份研究摘要或一张图表，让 Claude 找出一位具备统计学素养的读者会提出的疑问。经过足够多次这样的练习，你就会开始自动提出这些问题。
17. 如何系统性地解决问题
大多数人解决问题的方法一成不变：做脑子里冒出的第一件事，行不通就调整，一直试到某个办法奏效为止。这能产出解决方案，但很少是最优的，因为脑海中冒出的第一个解决方案几乎从不是经得起仔细推敲的那一个。
系统性问题解决是一套具体的工具：在急于寻找解决方案之前先定义真正的问题，把症状与原因区分开，在评估之前先产生多个备选方案，以及针对最可能使方案失效的条件对拟议方案进行压力测试。AI 可以引导你将这一框架应用到你正面临的实际问题上，这才是真正学会它的方式，而不仅仅是知道它存在。
18. 如何建立真正的自信
不是表演出来的自信。不是那种在压力下崩溃或需要不断寻求肯定的自信。而是那种来自你自身能力的累积证据的自信，来自清晰地理解自己真实的优势与劣势，而不是透过他人意见或自身焦虑的扭曲滤镜去看待。
这不是 AI 能替你建立的东西。但 AI 可以加速这一过程：帮助你诚实地反思自己的经历，识别你擅长之事中的模式，并培养那些产生真正能力——自信真正根基——的技能。
19. 如何撰写能赢得认可的提案
提案就是以文件形式呈现的销售。大多数提案失败，不是因为所提议的工作不够好，而是因为文件聚焦于交付物而非转变。它描述的是将做什么，而不是结果将有什么不同。
聪明的人正在用 AI 来理解获胜提案的结构，针对具体场景练习写作，并获得反馈，了解自己当前的提案在哪些地方是在推销过程而非结果。对于任何需要说服他人或组织对某件事说“是”的人来说，这项技能都会直接且立即带来回报。
20. 如何用一半时间学会任何新东西
这一条是所有其他技能背后的元技能。
大多数人通过线性阅读来学习，指望理解会自然累积。最高效的学习者则有不同的参与方式。他们在填充细节之前先把握学科的整体结构。他们主动测试自己的理解，而不是被动复习材料。他们通过向别人讲解所学来暴露知识的漏洞。他们寻找那种能让抽象概念变得具体的特定类型的例子。
AI 可以被配置成完全按照这种方式教学：围绕你已有的知识构建学习体验，实时测试理解程度，生成能将新想法与你已理解的事物联系起来的具体例子，并将一门新学科的结构显性化——否则你可能需要数月的沉浸才能察觉。
掌握了这一元技能的人不只是学得更快。他们是在建立一种能力，能够在余生中加速他们在任何领域的学习。
## 现在就开始的理由
借助 AI 学习这些技能存在一个能产生超额回报的时间窗口。这个窗口之所以存在，是因为现在就开始学习的人能在大多数人还在犹豫要不要开始时，就建立起优势。
这个窗口不会永远敞开。工具会变得更加主流。这些技能会成为基本功，而不再是差异化优势。先发优势会被压缩。
但这并不意味着晚点开始也无所谓。它的意思是，现在开始比晚点开始好，晚点开始也总比不开始强。
从这份清单里挑两项。挑在你的具体生活中、在你的工作中、在你正在努力实现的目标中，能产生最大回报的那两项。
在接下来的 90 天里，有意识地——而不是随意地——用 AI 来培养它们。
现在正在学习这些东西的人并非天赋异禀。
他们并非技术高超。
他们也没有在等待所谓的合适时机。
他们只是在其他人还在观望的时候，决定开始了。
## 相关链接
- [NeilXbt](https://x.com/neil_xbt)
- [@neil_xbt](https://x.com/neil_xbt)
- [5.9K](https://x.com/neil_xbt/status/2045416688808554571/analytics)
- [升级到 Premium](https://x.com/i/premium_sign_up)
- [2026年4月18日 下午4:18](https://x.com/neil_xbt/status/2045416688808554571)
- [5,964 次浏览](https://x.com/neil_xbt/status/2045416688808554571/analytics)
- [查看引用](https://x.com/neil_xbt/status/2045416688808554571/quotes)
---
*导出时间: 2026/4/19 11:09:56*
