行为调查员:以数据破解自我欺骗
大多数人使用 ChatGPT 或 Claude 仅限于提高工作效率,或者处理不想亲自动手的繁杂任务。然而,AI 能否真正改变你的生活轨迹,帮助你构建一个值得为之早起的人生?答案是肯定的,前提是将其作为深度对话的思考伙伴。我们将剖析阻碍我们摆脱生存模式的五大心智盲区(Cognitive Blind Spots: 个体无法自知或容易忽视的认知偏差与心理防线),针对每个盲区探讨其成因、AI 的辅助路径以及具体实操方法,并提供可直接套用的提示词。如果能认真践行这套方法 30 天,你将在人生的核心维度上迎来彻底的蜕变。
在人生的道路上,最危险的往往是那些视线之外的隐形障碍。正如美国高速公路上重达 8 万磅且极难转向的 18 轮重型卡车,每年有高达 84 万起车祸与变道和盲区相关。我们的人生也像这些卡车一样,承载着职业、家庭和未来的重担,而在变道时极易发生危机。第一个盲区是自我欺骗。我们为什么明知某些行为在伤害自己,却依然重复它?在航空领域,事故调查不会仅凭飞行员的记忆,而是依靠黑匣子(Black Box: 飞机上记录飞行数据与语音的抗毁设备)的数据,因为人在压力下的记忆往往是扭曲的。自我重塑也应从收集生命的数据开始。你的生活就是这个黑匣子,你可以让 AI 扮演事故调查员,从客观第三方的视角审视你提供的证据,从而揭示那些你未曾察觉的行为模式。
针对这一盲区,你可以使用行为调查员提示词与 AI 进行单问题交互式采访。在发现这些自我欺骗的盲点后,具体的实操方法是建立你的“负面反馈清单”。研究表明,来自熟知你的人的真实、具体的负面反馈能显著减少自我欺骗。
Original English
This is how you use AI to change your life. Most of us use ChatGPT or Claude for productivity hacks or to do work that we don't want to do. But can AI actually make a difference? Can it help you build a life worth waking up for? I think any thinking partner can if you use it the right way. So we're going to look at the five blind spots that keep you stuck in survival mode. For each blind spot, we're going to do three things. First, unpack why it happens. Second, show how AI can help. And third, what you can actually do about each blind spot. And for each one, I'll share a specific approach and a specific prompt that turns AI into your partner as you reinvent yourself. I'll also show you where AI is limited and why your judgment still has to stay in charge. Do this seriously for just 30 days and you may become unrecognizable in the only way that matters. The most dangerous thing on the road is not always the thing you can see in front of you. On US highways, you see these 18-wheelers, these massive trucks that weigh around 80,000 lb. They're very hard to turn. Some personal injury law firms estimate that 840,000 crashes a year in the US alone involve lane changes and blind spots. And you know, our lives feel like those giant trucks sometimes. The freight you carry is heavy, building your career, your family, a company, a future. The road is full of lane changes. The first blind spot is deception. Why do I keep doing the thing that I know is hurting me? Every plane carries a device called the black box. When something goes wrong, investigators don't start with the pilot's memory or his or her story because under pressure memory gets distorted. They start with the data recorded in the black box. Self-improvement ought to start in the same place. Your life is the black box. It has all the data. It has all the information, but you can make AI to act like the investigator. It can help you examine the evidence you can provide from a third party's perspective. The pattern that you may not have seen before become more obvious. Here's our first prompt. Act as an investigator for my behavior. This is the pattern I keep repeating. I don't give candid feedback to my team members, even when I know it would help them. I focus on positives and avoid the uncomfortable conversations that could help them improve dramatically. Now, this is a real pattern for me, but you could insert any pattern of your choice. For you, the pattern might be that you can't work out consistently, which is my problem, too. Or that work keeps getting your best attention while your partner or family gets the leftovers. Whatever it is, share here. Next part of the prompt. I want to understand three things. One, where my blind spots may be. Two, why I keep avoiding them. Three, how do I catch them before they're triggered? Interview me one question at a time. Unpack my answer and based on it, ask me the next better question. Use only my answer as evidence. When you're ready, give me your best hypotheses, the evidence behind it, and the evidence that's still missing. Be respectful, candid, and direct. Do not motivate me or flatter me. Do not suck up to me. That's the whole move. That's the prompt. One very important caveat here, AI doesn't have any divine insights. It can distort facts, it can overstate patterns, and it sounds more certain than it should. So, treat the answer as a hypothesis, not a diagnosis. AI is not a trained psychologist. So, why use it at all? Because your own mind is not neutral, either. The problem is that we rarely recognize our own self-deception. Once we've found a story that protects us, we stick to it. We stop asking a harder question. Now, what do you actually do in real life when you see it spots such a deception? A study from Frontiers in Psychology found something very interesting. They found that real, concrete, negative feedback can reduce self-deception. I was surprised to find that, too. And I don't mean feedback from social media trolls. Ignore all of that. I mean feedback from someone who has worked with you, who knows you well. That person has already told you the truth at least once in the past. Here's one thing that I do. I keep a list of every hard piece of feedback I've gotten over the years. And I read it sometimes. I go back to it. I don't kick myself for it, but again, it's a good mirror. Also, if you want a more in-depth look at topics like this one, please join our newsletter. Every Tuesday, one insight, one tool, and one practice. I love writing it. Link is below, and it's totally free.
对话演练方:打破人际关系中的叙事投射
在建立起识别自我欺骗的防线后,我们需要面对第二个盲区:人际关系,即我们对他人意图的主观臆断。心理学家 Eli Finkel 在西北大学对 120 对关系逐渐恶化的已婚情侣进行了一项研究。他们发现,仅需花 20 分钟站在客观中立且关怀双方的第三方视角写下最严重的冲突,就能成功阻止亲密关系的下滑。这是因为任何沟通都包含四个层次:你本想表达的、你实际说出口的、对方听到的,以及对方所理解的。而在日常冲突中,我们往往带着一套预设剧本(如“她无视我”或“员工很懒”),并将所有新细节都作为强化旧剧本的呈堂证供,使沟通变成了单方面的指控。
此时,AI 的角色将从行为调查员转变为对话演练方,帮助你构建对方立场的最强辩护词。你可以使用对手戏提示词,甚至开启语音模式,与 AI 模拟真实的冲突对抗。在现实生活中,克服这一盲区的关键在于用事实观察替代动机指控(Observation over Accusation: 专注于描述具体行为和客观影响,而非审判对方的意图)。不要说“你根本不在乎我”,而是说:“当连续两天没有收到你的回复时,我开始怀疑这件事情对你是否重要。我的看法客观吗?”
Original English
Now, let's go to the second blind spot, relationship. The stories we tell ourselves about others. There is a psychologist named Eli Finkel who ran a study at Northwestern University with 120 married couples. And all of those couples had their relationships slowly getting worse. Now, some of these couples were asked to write about their worst fights from a view of a neutral outsider who wanted the best for both of them. And you just had to do it for 20-odd minutes. And the surprising fact was the decline for those couples stopped. The rest who didn't go through that writing exercise kept sliding further. And I think that happened because writing brought them some clarity that every communication has four parts. What you meant to say, what came out of your mouth, what the other person heard, and what they interpreted. And all four of these can be very different. Because in every conversation, we bring a story. She ignored me. He disrespected me. My partner doesn't love me. My employees are lazy. On and on. And too often, the goal of that conversation is to reaffirm that story that's playing in our head. Every new detail is recruited to reinforce that old story. And that's why relationships are such powerful blind spots. Because the distance between a conversation and a prosecution is so tiny. This is where AI can help. In the previous blind spot, we used AI to act as an investigator. For this one, we use AI as the counterparty. Use it to build the stronger version of the other person's case. Here's the prompt. Act as the person I am in conflict with, or the person and trying to communicate with. Do not pretend that you know their inner life. Here's the situation in my own words. Be as honest as you can. Here's what I plan to say. Tell AI what you're trying to say, your side of the equation. Do not take my side. Argue their case as they might, assuming they're smart and fair and have reasons that make sense to them. Then, tell me the following. One, what they most likely feel that I'm not seeing. Two, the weakest part of my argument. And three, what they may need to hear from me. Do not flatter me. Do not turn this into a therapy session. Help me rehearse the actual conversation. That's the prompt. And by the way, for this one, turn the sound mode on and have an actual spoken conversation with AI. Now, what do we do about this in real life? Well, the study that I talked about gave us a clue. That small habit of taking a neutral, third-party perspective helped stop the decline in those couples. So, that's definitely a good move. One other thing I learned, a little late in my life, unfortunately, was to replace accusation with observation. You know, we sometimes don't know the difference between attacking the pattern versus attacking the person. The useful move is much simpler. Name the behavior, name the impact, and then ask the question. Be curious. So, instead of saying, "You don't care about me." try this. You know, when I don't hear back from you for 2 days, I start wondering if this is important to you. Is that fair? Am I being fair? What do you think? That's the human work that we have to do. AI can help you unpack the story, but you still have to face that person yourself in that room.
转型制图师:重构流动的身份认同
理顺人际关系中的投射后,我们必须面对与自身身份认同的博弈。在瞬息万变的时代,我们需要不断进行自我重塑。心理学中存在一种历史终结错觉(End of History Illusion: 人们能清晰意识到过去的巨大变化,却倾向于认为自己未来的性格和价值观将保持稳定)。对 1.9 万名 18 至 68 岁人群的研究表明,人们普遍认为当下的自己已经是“最终版本”,这导致了心智维度的身份盲区。如果将自我意识与某种特定角色(如职业运动员、CEO)融合得过于紧密,在转型时就会面临巨大的痛苦。正如演讲者自己曾从灵修僧侣、职业乐手、数学家、华尔街分析师到 AI 公司 CEO、投资者乃至如今的创作者,只有把这些角色视为随时可以脱下的衣物而非皮肤,才能在身份转换时保持从容。
在这种迷茫的导航期,AI 将扮演你的转型制图师。你可以使用身份转型提示词让 AI 评估你的资产转移能力,并明确哪些曾经的优势在新的阶段会演变成包袱。明确差距后,最核心的现实行动是运行最小可行性测试(Minimum Viable Test: 以最低成本和风险验证新方向可行性的微型实验),例如在副业展现出清晰的商业闭环前,不要轻易辞去全职工作。
Original English
Now, the third blind spot is the relationship you have with your own identity. We're entering an era where you may have to reinvent yourself again and again, and that's one of the trickiest transformations. That's where we go next. There is a strange illusion in psychology called the end of history illusion. Researchers studied more than 19,000 people between the ages of 18 and 68, and they found that people can look backward and see how much they have changed, but when they start looking forward, they assume they're not going to change that much. We know how much we changed in the last 5 years, but we feel certain we'll be the same for the next 5 years. As if we have arrived at some final version of who we are. That's never the case. That is the identity blind spot, and it gets worse. Sports researchers have found that athletes who have their sense of self fused too tightly with the sport they play often have the hardest time when they retire. And you know, I have had a lot of personal experience with this idea of ever-changing identity. I have pivoted so many times in my life. When I was young, I wanted to be a monk. That's all I wanted to be. I trained for it in an ashram. Well, that didn't work out. Then I tried to be a full-time musician, then a mathematician, an equity analyst on Wall Street, a CEO of an AI company, a board member, then an investor, and now this YouTube channel. So, technically I am a creator, but now when I look back and think about it, none of those roles were fused with my identity. They didn't form my identity. They were just clothes that I was wearing, not my skin. I wore them when it was time, I took them off. So, that's one good way to look at life. The confusion feels existential, but it's just navigational. This is where AI can help. For this blind spot, AI is your mapmaker. It can give you a map from the old territory to the new landscape. Here's the prompt. Act as my transformation partner. You are my mapmaker. Map out my next move. Here's where I am right now. So, here you can put your current role, your strengths, your habits, your fears, what people reward you for, and so on and so forth. Give it more details. Here's where I'm trying to go. Here you'll describe your new chapter. Be honest, you know, your new job, your new company, new career, new relationship, new phase of life, health reset, whatever lies ahead. Now, interview me one question at a time to find the gap. Then, give me an action plan and a map that bridges that gap. Respond to these five questions. Number one, which of my strengths still will carry over? Number two, which strengths will become liabilities in this new chapter? Number three, what I need to leave behind even though it's worked before? Number four, what I've never built but is now needed? And number five, what I should test for at least 30 days before I make a bigger commitment. Do not give me abstract purpose advice. Do not flatter me. Do not give me a 10-year life plan. Give me a practical transformation map. And do not suck up to me. You have to keep saying that. All right, so that was our prompt. Again, one caveat, AI won't know the emotional cost of whatever you're leaving behind to become someone new. It will make this reinvention of your identity sound crisp and clean. But, life is usually pretty messy. I've gone through such transformations many times, and I know there is grief, there is fear, there is anxiety. But, there's also growth. Now, what do you actually do once you see the gap? You should run the smallest real test that you can. For instance, don't quit your day job until your new adventure shows some promising signs of working.
决策模拟器:沙盘推演未知的命运分叉
在看清转型差距并设计好测试后,我们常常卡在具体决策的分水岭,这构成了第四个盲区:决策分叉口。正如电影《荒岛余生》(Cast Away)的结尾,主角站在十字路口面临抉择。分叉口之所以痛苦,是因为两条道路都伴随着机会成本与风险,且你无法同时体验。为了预测这些风险,我们可以引入决策专家 Gary Klein 创立的事前剖析法(Pre-mortem: 在决策实施前,假设项目已彻底失败并反向推导原因的决策工具)。通过想象最糟糕的失败场景,逼问自己忽略了什么、因为过度兴奋错过了什么,或者由于害怕而不敢承认什么。
在此场景下,AI 是你的决策模拟器。使用事前剖析模拟提示词,让 AI 分别对选项 A 和选项 B 运行独立的失败沙盘推演,客观对比两者的失败路径、早期预警信号以及不可逆的决策成本。分析完成后,我们可以套用杰夫·贝佐斯的门类型决策原则:
- 双向门决策(Two-Way Door Decisions: 试错成本低、可随时返回的决策):应当快速行动,迅速迭代。
- 单向门决策(One-Way Door Decisions: 成本高昂、一旦跨过便无法回头的决策):必须放慢速度,进行极度严谨的推演。
Original English
The fourth blind spot is forks, the decision you keep staring at because it's hard to make. There is a movie called Cast Away. And at the end of the movie, the main character, played by Tom Hanks, is standing in the middle of a crossroad, a map in hand, weighing his options. We've seen everything that he's gone through and what brought him here. And for a moment, the whole movie boils down to one question, which road will he take? We all know that feeling, standing at a fork, staring at a decision we don't know how to make. Do I go left or right? Do I quit or stay? Propose or wait? Forks are hard because both roads have a cost. Both roads have a risk, and you don't get to live both of them simultaneously. So, how do you forecast these risks? Gary Klein created a decision tool called pre-mortem. Before you make the decision, you should imagine the absolute worst scenario. Describe it in detail. Write it down. Then ask, "What went wrong? What did I ignore? What was I too excited to miss? What was I too afraid to admit?" This is all hypothetical, so create the most realistic scenario you can. And this is where AI can help you tremendously. In this blind spot, AI becomes your simulator. It runs that pre-mortem with you. And here's the prompt. "Act as my decision simulator. I am facing a fork ahead." Here you can describe the decision you have to take, the options you have, etc. Option A is describe option A. Option B is describe option B. Here's what I know. Now you can add all the constraints, money, time, people, organization, timeline, whatever. The rest is the same. Interview me one question at a time to understand the decision, and then run a separate pre-mortem on each option. For each one, imagine that it's 2 years later and that choice failed badly, and work backward from there, and tell me why. Treat them independently, so I can compare how each road fails differently. And then for each road, tell me these three things. Number one, the most likely way it will fail. Number two, the warning signs I should see first. And number three, the cost I cannot reverse. Do not tell me what to choose. Be very rigorous and systematic about your analysis. All right. So, that was the prompt. Again, you can add anything you want on top of it. You're not asking AI to choose the road for you. It cannot. No one can. But AI can pressure test your assumptions. It can surface risks that you haven't thought about yet, and help you understand the trade-offs. That's what that exercise is for. Now, what do you do in real life after AI has analyzed it? I like Jeff Bezos' rule. He talks about two types of decisions, one-way doors and two-way doors. A two-way door is when you can walk through a door and if it doesn't work out, you walk back. The cost is bearable. In those cases, you should move fast. And a one-way door is the one you can't reverse. Those are rare, but they do happen. When you're facing one of those, slow down.
复盘伙伴:在废墟上重建心智秩序
即便我们做出了最谨慎的决策,痛苦与挫败依然可能在不经意间降临,这就是第五个盲区:挫折与创伤。当裁员或背叛这类突发性失败发生时,身体的防御机制会让我们陷入“我为什么总是把事情搞砸”或“我不该相信任何人”的破坏性叙事中。德克萨斯大学奥斯汀分校的研究表明,连续 4 天、每天仅花 15 分钟写下痛苦经历的人,数月后的身心健康状况显著优于对照组。写作能够帮助我们将主观痛苦客观化。
此时,AI 是你的复盘伙伴与倾听者。你可以使用创伤复盘提示词,要求 AI 在不提供廉价安慰或即时解决方案的前提下,通过单问题追问帮你理清事实与主观臆断的边界,区分哪些是自己可控的,哪些是不可控的,并提炼出需要带向未来的教训。需要强调的是,大语言模型绝非专业心理咨询师(Licensed Psychologist: 接受过系统训练的心理治疗专家)的替代品。对于无法独自承受的严重创伤,必须寻求专业人士或亲友的帮助。古老的佛教智慧告诉我们:痛苦与受苦是两码事——痛苦是已然发生的事实,而受苦则是我们在事后不断对自己重复的悲情叙事。就像初学吉他时手指会破裂流血,但当老茧形成后,音乐便会随之诞生。
Original English
Now comes the fifth blind spot, trials. What do you do with the trauma and pain after you have failed? Our trials and traumas don't ask for our permission. They arrive unannounced. They don't care for our opinions. Long-term viewers know this story that I've shared before. At one of my jobs, my boss invited me to a very fancy restaurant for drinks and dinner. And I was thinking that I was getting promoted. And in the middle of the dinner, he handed me a brown envelope and said, "We've decided to let you go." When failure comes this suddenly, your entire body just shuts down. But the blind spot isn't the event. It's the story that stays behind. Why do I keep screwing up? I shouldn't trust people. I should have never tried that. On and on and on. A researcher at UT Austin found something strange. People who wrote about a painful experience for even 15 minutes four days running were way healthier months later. And that has been my experience, too. When we fall, writing about it helps a lot. And this is why AI can help. For this blind spot, AI becomes your debrief partner, a companion. Here's the prompt. Act as a debrief partner for something difficult that I have gone through. Here's what happened. Now you can describe the failure, loss, your setback, betrayal, rejection, grief in plain language. The story I've been telling myself about this is this. Share what you've been thinking about, about yourself, about other people, trust, risk, love, work, future, all of it. The rest, of course, is the same. Interview me one question at a time. Do not comfort me or try to find a solution. I just want to talk. And then help me understand the following four things. Number one, the facts of what happened. Number two, the meaning I keep attributing that isn't about the facts. Number three, what was mine to control and what wasn't. And number four, one thing to carry forward, one thing to leave behind. That's the prompt. And again, remember, AI is not a therapist. And for some traumas and some trials, you're going to need one. If this is a grief, pain, or a weight you can't lift alone, a language model is not the way to go. Talk to a close friend or an expert. Take care of yourself. Loss and grief change you. They rearrange you forever. You don't get to go back to who you were before. But you don't have to let your wound become your way of life. There is an old Buddhist teaching that pain and suffering are not the same things. Pain is what happened. Suffering is the story we keep telling ourselves after. You know, when I first learned guitar, the skin on my fingertips would split. It would bleed. It hurt. But then the skin hardened. First, there's pain. Then there's music. Thank you. And I love you.