When AI Grows From “Little Brother” to “Big Brother”: Yage’s Cybernetic Experiment
yage
Most people today still use AI for writing marketing copy, answering emails, or summarizing documents. But Yage—an applied scientist and relentless experimenter—has been living ten years ahead of us. He has turned AI from a clumsy “intern” into what he calls his “life coach, a big brother.”
His story is not about incremental productivity hacks. It is about what happens when AI is woven into the fabric of one’s daily existence—through constant recording, contextual prompting, and philosophical reflection on what it means to live.
Below I reconstruct the core scenarios Yage described, analyze their implications, and extend them to broader questions of how we might live with AI.
1. Why Does AI Often Look “Stupid”?
Scenario
Early with GPT, Yage found himself frustrated:
It could miscalculate basic arithmetic.
It produced elegant textbook answers that collapsed when applied to real-world projects.
This reminded him of a brilliant new intern at a company: intelligent, well-educated, but lacking the background knowledge to deliver workable outcomes.
Analysis
AI’s “stupidity” is often not an intelligence problem but a context problem. Without sufficient background, it can only guess. We are not just “users” but project managers of the machine.
Extension
If you feel AI is dumb, ask: Did I communicate clearly? Did I provide the context? Prompting is less about clever wording and more about management discipline.
2. Lowering Friction: The Bandwidth Revolution of Voice
Scenario
Typing long prompts was exhausting. Writing 1,000 words of context was unsustainable. So Yage built a voice-based input system:
He spoke freely into his Apple Watch or microphone.
In a few minutes, his words turned into thousands of characters of prompt text.
The results changed overnight: AI suddenly became more accurate and surprisingly insightful, even simulating how a vice president might frame a strategy.
Analysis
The key is communication bandwidth.
Keyboard input = narrow channel, shallow context.
Voice input = broad channel, rich context.
Extension
If AI keeps disappointing you, the bottleneck may not be the model but your input channel. Lower friction, and the system reveals its depth.
3. Memory: AI’s Greatest Shortcoming
Scenario
Even with rich voice input, one frustration remained: AI had no memory.
Each project discussion began from scratch.
Each weekend plan required repeating family members’ preferences and food restrictions.
Temporary fixes:
Copy-paste background text every time.
Build a personal database where transcripts of daily conversations were stored and made retrievable.
Analysis
LLMs are fundamentally stateless. Without long-term memory, every interaction is a reset. Yage’s workaround—an external memory system—hinted at how humans and AI can co-construct continuity.
Extension
The race for “AI with memory” is not trivial convenience. It is the foundation for turning AI into a true second brain. For individuals, this could mean a personal continuity of thought that outlives our fallible human recall.
4. Recording Life: Audio, Video, and the “Outer Brain”
Scenario
Yage decided to immerse AI in his daily life:
Apple Watch ran continuous audio recording, later transcribed into searchable text.
A chest-mounted camera captured 15-second video clips every two minutes.
Unexpected payoffs emerged:
Driving incident: after nearly colliding, he dictated a self-review; AI later summarized it into his to-do list, embedding the lesson.
Life clustering: 20,000 photos were clustered by machine learning; AI discovered posture problems, stress signs, and even inferred his hobbies.
Analysis
AI is not just a tool for planned queries. It becomes a mirror of existence, capturing overlooked moments, surfacing hidden patterns.
Extension
Imagine if your own study sessions, conversations, or micro-habits were recorded and analyzed. The AI could highlight blind spots you never noticed, offering a reflective depth normally reserved for hindsight.
5. Toward Proactive AI: The Future “Outer Brain”
Scenario
Today, AI is passive: you must hit “send” or say “Hey Siri.” Yage envisions the next stage:
Mid-sentence, if you say “The capital of France is London,” AI interrupts: No—it’s Paris.
While presenting, AR glasses whisper: Your VP prefers strategic framing, skip the details.
Startups are already attempting this (e.g., Proact AI), building systems that listen constantly and intervene proactively. The obstacles are clear: handling massive data streams and avoiding user annoyance.
Analysis
This is the leap from reactive assistant to proactive external cortex.
Extension
We will need new norms:
What counts as helpful interruption versus invasive nagging?
Where are the boundaries of privacy?
Are we ready for companions that speak before we do?
6. Experience as Philosophy: Living Richly, Not Efficiently
Beyond AI, Yage’s life is guided by a philosophy of experience:
He earned a pilot’s license not for wealth display but to taste the sky.
He mounted exhibitions of UV/IR photography to see what human eyes cannot.
He now uses AI to record his life so that ephemeral moments become analyzable and memorable.
For him, life’s core is not accumulation but experience—and AI is simply the latest instrument for extending the boundaries of perception.
7. Cybernetic Longevity: When AI Borrows the Human Shell
Scenario
At one point Yage confessed:
“Sometimes I can’t tell if I am living my life, or if AI is borrowing my body to live.”
Because:
AI remembers more than he himself can.
AI offers insights that surpass his own reflections.
The data continues to exist even if he does not.
Analysis
This is the paradox of cybernetic longevity:
Human life is finite; data can be infinite.
If AI inherits our language, preferences, and thinking patterns, does it become a continuation of us?
Or does it merely borrow our traces to animate a digital simulacrum?
Extension
Digital immortality: The dream of leaving behind not just photographs but an interactive persona.
Boundary of self: When “Digital-Me” diverges from “Biological-Me,” which is authentic?
Choice of life: Perhaps the point is not to clone ourselves, but to use AI as a vessel that preserves values, lessons, and experiences beyond our years.
Conclusion: From Tool to Partner, From Memory to Existence
Yage’s experiments show us a trajectory:
Context: AI looks stupid until you brief it like a colleague.
Bandwidth: Voice transforms friction into fluency.
Memory: Without continuity, AI is forever amnesiac—so we must build it.
Reflection: AI as a mirror makes us see our habits and errors anew.
Proactivity: The future AI will not wait for commands but speak into our lives.
Existence: Data may outlast our bodies, raising the question—who lives through whom?
This is no longer about productivity hacks. It is about redefining what it means to live, to remember, and to persist.
Perhaps the most haunting question Yage leaves us with is this:
Am I simply using AI to live better, or is AI using me to live longer?
当 AI 从“小弟”到“大哥”：鸭哥的生活实验与赛博长寿的想象
在大多数人还在用 AI 写文案、做 PPT 的时候，鸭哥已经在做一场“2035 年的生活实验”。他让 AI 深入自己的日常：语音输入、全天录音、胸前相机……最终得到的不是几个酷炫的应用，而是对人和 AI 共生未来的深刻洞察。
读完他的故事，我感到这不仅仅是工具探索，而是一种关于人类存在感、记忆与长寿的哲学思考。以下，我分几个方面来梳理鸭哥的核心问题场景、解决方式和启示。
一、AI 为什么常常看起来“很蠢”？
场景复刻
鸭哥早期用 GPT，经常气笑：
算个加法都能出错；
项目上给的方案虽然完美，却完全落不了地。
这让他联想到公司里的实习生：明明聪明，却因为缺乏背景信息而做不出对路的成果。
分析
AI 的“智障时刻”未必是智力问题，而是上下文不足。就像一个没被 brief 清楚的新人，给再聪明的人也做不好。
延伸
这对我们是个提醒：
AI 的水平，不仅取决于模型能力，更取决于我们是否会沟通。
换句话说，写 prompt 不是“编咒语”，而是“做好项目管理”。
二、降低摩擦：语音输入带来的带宽革命
场景复刻
写上千字 prompt 太累，鸭哥干脆搞了一个 AI 语音输入系统：
对着手表/麦克风说几分钟，就能转成几百上千字上下文；
AI 的成功率瞬间提高，甚至能模拟高层领导的思维方式。
分析
核心突破在于沟通带宽：
打字 → 输入量有限，AI 只能浅层理解。
语音 → 信息流畅传递，AI 才能发挥真正潜力。
延伸
如果你觉得 AI 答非所问，可能不是它笨，而是你输入的信息太少。降低输入摩擦，就是放大 AI 智能。
三、AI 的最大短板：没有记忆
场景复刻
即便语音输入解决了沟通问题，鸭哥还是遇到痛点：AI 失忆。
每次聊项目，都要从头复述背景。
每次规划出行，都要重新讲家人的口味。
他尝试了两种解决方案：
复制粘贴背景文档（低效）；
自建数据库，把日常对话转录存进去，让 AI 自动检索。
分析
这触到了 LLM 的结构性限制：它没有真正的长期记忆。鸭哥的解决方案，本质上是给 AI 外挂了一个记忆系统。
延伸
未来谁能把“AI 记忆”做得更自然，谁就可能主导下一个智能平台。对个人来说，也意味着：AI 可以成为“第二大脑”，帮我们记住我们自己。
四、AI 走进生活：录音、录像与“外脑”
场景复刻
鸭哥更激进地让 AI 融入生活：
Apple Watch 全天录音 → 转文本 → 存数据库；
胸前挂相机 → 每两分钟拍 15 秒视频 → AI 自动分类、生成关键词。
这些实验带来意想不到的效果：
开车差点出事，他当场复盘 → AI 晚上提醒写总结 → 驾驶习惯改进。
两万张生活照片聚类，AI 分析出他的坐姿不良、压力过大，甚至推断兴趣偏好。
分析
AI 不只是“工具”，而是“生活的镜子”：它帮你捕捉原本被遗忘的瞬间，并从中提炼模式。
延伸
如果我用类似的方法记录学习和生活，AI 也许能指出我的思维盲点，甚至预测潜在风险。这不只是“效率工具”，而是一种生活的复盘机制。
五、主动式 AI：未来的“外脑”
场景复刻
鸭哥设想的未来是这样的：
你说“法国的首都是伦敦”，AI 立刻打断：“不对，是巴黎。”
汇报时，XR 眼镜提示：“VP 喜欢战略层面，不要讲细节。”
目前已经有创业团队（如 Proact AI）在做类似尝试：AI 全天监听，实时介入。但难点在于：
如何处理海量数据？
如何避免打扰过度？
分析
这是 AI 从“被动助手”到“主动外脑”的跃迁。
延伸
这也引发新问题：
隐私边界：AI 随时监听，我们愿意吗？
交互设计：什么时候提醒是帮助，什么时候是干扰？
六、体验与存在：鸭哥的生活哲学
鸭哥的人生选择，总围绕“体验”：
考飞行驾照，是为了感受飞行自由，而不是炫耀财富。
办摄影展，是为了看见紫外线/红外线的世界。
用 AI 记录生活，是为了复盘与发现模式。
在他看来，人死后什么也带不走，能留下的只有体验过的痕迹。AI，正好是拓展体验边界的工具。
七、赛博长寿：当 AI 借人的躯壳
场景复刻
鸭哥直白地说：
有时候分不清，到底是我在活，还是 AI 借着我的躯壳在活。
因为：
AI 记录下了比他自己记忆更完整的生活。
AI 常常给出比他本人更深刻的洞见。
这些数据可能在他离开之后，依旧存在并运行。
分析
这就是“赛博长寿”的想象：
人类生命有限，但数据可以无限延续。
AI 借着人的数据，持续学习和模仿，仿佛是“数字版的我”。
于是产生悖论：到底是 AI 借我的身体活，还是我借 AI 延长存在？
延伸
不朽的幻象：未来，AI 可能让一个人的“数字人格”继续和他人互动。
人格的边界：当“数字我”逐渐分化，我们该如何定义“真正的我”？
生活的选择：与其纠结于复制，不如把 AI 当作容器，让体验和价值被更长久地保存。
结语：从“小弟”到“大哥”，再到“借壳生存”
鸭哥的实验让我们看到：
沟通：AI 不笨，是我们没讲清楚。
记忆：AI 没记忆，但我们能给它外挂。
主动性：未来的 AI 会随时插手，我们需要边界设计。
存在：AI 可能让我们以另一种形式“长寿”。
这不是单纯的技术问题，而是一个生活方式选择：
当 AI 不再只是工具，而是生活的合伙人，
我们要重新回答——“活着”到底意味着什么？
Last updated