In complex engineering projects, project failures often stem from incorrect execution sequences. In the early stages, engineering teams often rely too heavily on Gantt charts for detailed schedules, only to encounter frequent reworks and delays during execution. The root cause is the lack of a rigorous task dependency structure design.
The division of labor between requirements analysis, Design Structure Matrix (DSM), and Gantt charts is crucial.
Core Conclusion: Requirements define the scope, DSM organizes the logic, and Gantt charts handle the time scheduling.
We use an EV Speed project as an example here. (The “EV Speed” initiative functions as a conceptual schema to architect the structural narrative and facilitate coherent technical exposition throughout the documentation.)
The Logic of Order: Integrating DSM and Gantt Charts
I. Prerequisite: Clarify Project Requirements
Before making any plan, engineers must clarify the ultimate goal of the project. Without accurate requirements input, subsequent schedule planning is meaningless.
1. The Core Role of Requirements Engineering
The requirements analysis phase needs to answer a fundamental question: What specific functions does the system need to deliver?
Standard operating procedures:
Identify core metrics: Lock in key data for the EV Speed project (e.g., cruising range).
Decompose specific tasks: Break down macro goals into executable micro steps.
Set acceptance criteria: Define specific engineering specifications for each task completion.
2. Connection Between Requirements and Subsequent Tools
Clear engineering requirements provide a complete task list:
The requirements phase identifies all tasks to be done.
The structural analysis phase clarifies the dependencies between tasks.
The Gantt chart phase assigns specific dates to individual tasks.
Without clear project requirements, DSM analysis will miss key steps.
II. Design Structure Matrix (DSM): Sorting Out the “Structure” First
DSM solves the underlying logical architecture problem: How do tasks depend on each other? Is the sequence reasonable?
1. Operational Mechanism
Engineers can understand this matrix as a “task dependency map.”
Basic approach:
List all decomposed tasks (e.g., battery management, system software).
Enter task names on both the horizontal and vertical axes of the table.
Mark the information transfer relationships between tasks at the intersections.
This table clearly demonstrates the operational logic within a complex system.
2. Assessing Structural Rationality
The value of the matrix lies in checking the health of the project’s logical structure.
Key assessment methods:
Diagonal Concentration: Tasks progress smoothly in order, and the engineering flow is extremely efficient.
Long-distance Dependencies: There are mutual waits across stages, requiring a readjustment of the work order.
High-coupling Blocks: Multiple tasks are intertwined and need to be handled by an independent technical group.
Circular Deadlocks: There are closed loops of repeated modifications between tasks, requiring early planning for multiple iterations.
3. Core Issues Solved
DSM does not focus on specific dates; it only optimizes the internal logic:
Reducing chaotic information transfer across development stages.
Significantly reducing the probability of technical rework in the later stages of the system.
Defining clear and reasonable boundaries for engineering team collaboration.
Identifying core R&D links that must be repeatedly tested and optimized.
III. Gantt Chart: Arranging the “Time” Next
Only after the logical relationships are completely sorted out is it the Gantt chart’s turn.
The Gantt chart primarily solves: When will each task start? How long will it take?
1. Core Functions
Set absolute start and end dates for engineering tasks.
Visually display the time span of individual tasks.
Mark the execution sequence on critical workflows.
Monitor daily execution progress and delivery delay risks.
Example of EV Speed project scheduling:
Battery development → Software adaptation → Vehicle integration testing.
Establish key milestones such as design freeze, engineering prototype, and mass production release.
2. Focus Points
Gantt charts focus on timeline progress management:
When the R&D team is ready to start.
How long equipment resources need to be occupied.
The impact of a delay in one link on the final product delivery date.
Essential positioning: A tool for time scheduling and project progress monitoring.
IV. Essential Differences: Requirements vs. Structure vs. Time
Briefly summarize the division of logic among the three engineering management tools:
Requirements Engineering: Determines what tasks the development team needs to do.
Design Structure Matrix: Determines which task to do first and which to do next.
Gantt Chart: Determines on which day the engineering plan will be completely finished.
Engineering logic chain: Requirements determine the development object. DSM determines the workflow architecture. Gantt chart determines the resource scheduling timing.
V. Why the Sequence is Irreversible
Skipping requirements and structural design to go straight to Gantt chart scheduling will lead to engineering disasters:
Missing hidden prerequisites, leading to complete reworks later.
Time assessments are completely inaccurate, making the schedule plan useless.
Modifying the critical path of project R&D every day.
If the EV Speed project does not clarify the dependency logic:
Software and hardware R&D teams wait for each other to deliver data.
Autonomous driving algorithms fail to receive underlying sensor information.
These hidden dangers will inevitably explode during the final vehicle integration test.
VI. Practical Suggestions (Applicable to complex projects like EV Speed)
1. Strictly Follow the Correct Sequence
When facing a complex system, first answer “which tasks to do.” Then clarify “who depends on whom.” Finally, fill in “specific delivery dates.”
2. Batch Intertwined Modules
When encountering intertwined task areas in the matrix table:
Encapsulate highly related tasks as an independent sub-project task force.
Allow high-frequency, rapid engineering trial and error within the group.
Require the group to provide stable and standardized module test results.
Example operation: Merge battery management and motor control into a Power Train R&D group.
3. Dedicated to Early-stage Design
DSM application scenarios:
Early system conceptual design phase.
Boundary division between R&D teams and technical modules.
Early planning of engineering technology roadmaps.
Do not use it as a tool for daily check-in progress scheduling.
4. Dedicated to Later-stage Execution
Gantt chart application scenarios:
Allocation of R&D team resources over time periods.
Establishment of key engineering milestone dates.
Tracking of daily technical development progress.
Do not use Gantt charts to design the underlying architecture of the system.
5. Recommended Standardized Workflow
Lock in all project R&D goals through requirements analysis.
Use the DSM table to sort out the underlying R&D logic network.
Divide complex and intertwined R&D tasks into independent task forces.
Reserve time for repeated iterations and modifications in advance.
Draw the project progress Gantt chart based on the organized logical framework.
Strictly monitor the technical team’s execution progress according to the calendar.
Conclusion
In a complex engineering system, the correct execution sequence is the cornerstone of project success.
First clarify the final business goal, then lay the system logical track, and finally start the train according to the calendar.
By following scientific engineering logic, the progress schedule will have true execution value.
This blog synthesizes concepts from physics, economics, cognitive science, and complex systems theory to construct a practical reasoning toolkit. The objective focuses on improving decision quality under uncertainty and navigating real-world complexity through structured thinking.
Engineers and scientists regularly operate with abstractions, models, and approximations. However, performance differences rarely originate from intelligence alone. Superior outcomes typically result from more effective thinking tools, which enable clearer reasoning, better trade-offs, and more adaptive strategies.
The framework progresses Thinking across six layers:
Worldviews establish foundational assumptions about reality, including beliefs about causality, determinism, and system behavior.
Identity & Agency define constraints, incentives, and behavioral capacity, shaping how decisions are executed within real environments.
Decision & Strategy formalize action under uncertainty using frameworks such as Bayesian inference and Value of Information, enabling rational updates and resource allocation.
Systems & Structure capture interactions across components using principles from complex systems theory, including the Free Energy Principle and network dynamics.
Uncertainty & Risk quantify variability through probability distributions, statistical inference, and theoretical limits such as the No Free Lunch Theorem.
Optimization & Growth address long-term dynamics through compounding effects, learning curves, and cognitive capital investment.
Thinking tools transform abstract understanding into actionable decisions. Actions generate feedback, which updates models through iterative learning cycles. This feedback loop drives continuous improvement and model refinement. Effective thinking tools enable practitioners to navigate trade-offs among accuracy, scalability, and efficiency, resulting in robust performance in complex, uncertain environments.
How to Use This Dictionary
The following section organizes core thinking tools into a structured reference system. Each model is a reusable lens for understanding complexity, improving decisions, and acting under uncertainty.
Narrative: This world is not a pile of facts, but a field of narratives; you must both step out of others’ stories and learn to set your own. 这个世界不是事实堆，而是叙事场；你既要跳出别人的故事，也要学会为自己设定故事。
Heavy Tail: This world is not evenly distributed, but full of extreme emergence; don’t get stuck in an additive world, seek multiplicative strengths and compound interest. 这个世界不是平均分配，而是极端涌现；别困在加法世界，要去寻找能做乘法的长板与复利
Agency: This world is already highly volatile, but old ideas still lag behind; don’t take steady-state survival logic as truth—be an agent, not a tool of old narratives. 这个世界早已高波动，旧观念却仍在滞后；别把稳态生存逻辑当真理，要做能动者，而不是旧叙事的工具。
Constraints: This world is not a wish-fulfillment machine, but a web of hard constraints; all actions must first assess resources, look for windows, respect rules, and acknowledge others. 这个世界不是愿望实现机，而是硬约束之网；一切行动都要先盘资源、看窗口、尊重规律、承认他人。
Possibility: This world will not give you absolute certainty; instead, meaning is generated by uncertainty. Don’t just seek stability—learn to manage, embrace, and even leverage uncertainty. 这个世界不会给你绝对确定，反而靠不确定性生成意义；不要只求稳定，要学会管理、拥抱，甚至利用不确定性。
Core: You do not have only one “self” but are driven by different layers of self; true growth is not about changing emotions or persona, but rewriting your core. 你并不只有一个“自我”，而是被不同层次的自我共同驱动；所以真正的成长，不是改情绪和人设，而是改写内核
Identity & Agency
Who you are and how you act: self-model, motivation, and internal stability for action.
Identity身份认同
Source: ① Robert Kegan’s Constructive-Developmental Theory of adult mental development; ② Daniel Dennett’s framework of the intentional, design, and physical stances.
Definition: Identity is an individual’s sense of belonging to a social role or group, shaping behavior, values, self-perception, and social interaction. Advanced identity means using identity as a tool, not being bound by it, enabling self-direction and mental growth.
Application: Mastering identity means: proactively set and switch identities, avoid being driven by a single one; empower actions in social, work, and learning contexts; reconstruct others’ identities to reduce conflict and friction.
Insight: Advanced players treat identity like clothing: wear what’s appropriate for the occasion, and hang it up at home.
高阶玩家的身份认同像衣服，到什么场合穿什么款式，回家了就脱下来挂在门口。
Security 安全感
Source: ① Mary Ainsworth’s Attachment Theory; ② John Bowlby’s “secure base” and “safe haven” model; ③ Amy Edmondson’s research on psychological safety.
Definition: Security is an individual’s stable sense of trust in their environment and relationships, the most basic human need. It enables exploration and creativity, and saves cognitive bandwidth. Advanced security means being able to support oneself and also be a source of support for others.
Application: Recognize your attachment style, identify anxious or avoidant patterns, reduce old emotional burdens; build a self-harbor (self-compassion) and a controllable environment; provide a secure base and safe haven in relationships, enhance others’ exploration and psychological safety, and gain strong social capital.
Insight: Security is the prerequisite for exploration, and exploration determines the breadth of one’s life journey.
安全感是一个人敢探索的前提，而探索决定了这个人能活出多大的生命历程。
Self-Determination Theory自我决定理论
Source: First proposed as a framework by Edward Deci and Richard Ryan in the 1980s.
Definition: Self-determination theory states that human motivation ranges from external control to full autonomy. High-quality motivation comes from autonomy, competence, and relatedness, not just external rewards. Highly agentic people can actively choose, decide, and shape their lives.
Application: Managers, teachers, or parents can foster agency by providing choice, challenge and feedback, and relationship support; individuals can enhance self-drive by internalizing task meaning, making micro-decisions, and gamifying tasks; high agency helps create flow, achieve long-term learning, innovation, and excellence.
Insight: The strongest force is not self-discipline, but willingness.
人最强的不是自律，而是自愿。
Power-Seeking Theorems 能耐需求定理
Source: ① Alexander Turner’s 2021 research on AI agents; ② Theories in cognitive science and behavioural empowerment, such as Christoph Salge’s empowerment hypothesis; ③ The principle “the gentleman is not a mere instrument” in traditional Chinese thought, and Kant’s moral philosophy that humans are ends in themselves, not means.
Definition: The power-seeking theorem states: In uncertain environments, agents (including humans) are most likely to achieve long-term goals and growth by increasing their future options and capabilities. In short, don’t just pursue a single reward—actively enhance your abilities and degrees of freedom.
Application: Understanding this theorem means: consider more options in decisions, avoid path dependence on old narratives; add roles to your identity, don’t be trapped by a fixed one; learn not just specific tools but transferable understanding; maintain independence in social relations, avoid total dependence on a single power source.
Insight: The real difference between a gentleman and a petty person is being proactive vs. reactive.
君子和小人的真正区别是主动和被动。
Cognitive Decoupling认知解耦
Source: ① Keith Stanovich’s research on slow thinking and cognitive decoupling; ② Lisa Feldman Barrett’s theory of constructed emotion; ③ Modern emotion regulation research, such as James J. Gross’s cognitive reappraisal theory.
Definition: Cognitive decoupling is the ability to separate “the narrative in your mind” from “the facts before your eyes,” thus regulating negative emotions. It works by recognizing the difference between perception and reality, taking others’ perspectives, and reassigning meaning to events, breaking the chain of automatic emotional reactions and giving individuals active control over emotions and actions.
Application: Prevent emotional rumination, reduce stress and health damage; maintain rational judgment in complex interpersonal and social situations; reappraise negative events as constructive opportunities to guide positive action; enhance metacognition so individuals can observe their own thinking rather than be controlled by emotions.
Insight: You can use cognitive decoupling to let some emotions arise and dissipate without getting on the train and following them.
你可以用认知解耦让某些情绪就这样升起又消散，而自己不必上车跟着它们走。
Decision & Strategy
How choices are made: structured decision-making under constraints and limited information.
Game Selection赛道选择
Source: ① The biological concept of Niche Construction; ② Saras Sarasvathy’s Effectuation theory.
Definition: Game selection is a mental tool for choosing which “game” to participate in within society and career. Choosing different tracks means taking on different rules, risks, and growth paths. Advanced game selection skills can make effort yield exponential amplification, rather than mere repetitive labor or pursuit of stability.
Application: Identify the track that suits you, build your niche based on your resources and abilities, create unique value; use effectuation thinking to iteratively combine resources at hand and discover or create new opportunities.
Insight: The greatest fear is fantasizing about great achievements while stuck in the system, or longing for the system’s stability while living in a multiplicative world.
最怕的是身在体制内却幻想大闹天宫的成就，或者身处乘法世界却眷恋体制内的安稳。
Explore / Exploit探索与利用
Source: ① The multi-armed bandit problem in computer science and decision theory; ② The Gittins index proposed by mathematicians.
Definition: Explore vs. Exploit is the strategy of balancing “trying new things (exploration)” and “deepening known advantages (exploitation)”: exploration discovers opportunities, exploitation creates value, and the combination leads to sustained growth and vitality.
Application: When choosing a career or learning direction, first explore different opportunities, then deepen strengths; in art or project management, try diverse approaches, then focus on effective strategies; maintain novelty and continuous learning to extend vitality and growth.
Insight: First explore, then exploit; after achieving results through exploitation, explore again, then exploit again.
先探索再利用，利用出成绩之后再探索，再利用。
WOOP Wish-Outcome-Obstacle-Plan
Source: A set of cognitive strategies popularized by Gabriele Oettingen.
Definition: WOOP is a thinking process that turns wishes into executable actions. Through mental contrasting and implementation intentions, it awakens the brain from a drifting state and enables proactive decision-making. The four steps are:
Wish: clarify the goal you want to achieve;
Outcome: imagine the specific feelings and benefits after achieving the goal;
Obstacle: identify the most likely internal obstacles in the process;
Plan: design executable plans for each obstacle.
Application: Helps individuals reduce on-the-spot decisions in life, study, and work, improve task accuracy; bind situational triggers to action plans to automate key behaviors and reduce procrastination; support calm and effective action under high pressure or low control.
Insight: WOOP turns the drifter’s unsolvable problem into the next solvable step; it’s a technique for regaining some control in an uncontrollable life.
WOOP 把漂流者的无解变成可解的下一步，它是一种在不可控生活中夺回一点可控的技术。
Bayesianism贝叶斯主义
Source: A probability formula proposed in the eighteenth century by Thomas Bayes.
Definition: Bayesianism holds that probability is a measure of belief, not an attribute of objective things. It uses “prior” to represent your original judgment, updates the prior with “evidence” to obtain the “posterior” probability. Decision-making and reasoning are based on continuously updating beliefs, not just single outcomes or appearances.
Application: Daily decisions: judge health risks, investment opportunities, colleague reliability, etc.; risk management and strategy optimization: combine priors and evidence to avoid extreme overreactions; individual cognition: train rational thinking, keep room for “maybe I’m wrong,” avoid rigid or absent priors.
Insight: Priors are both our wealth and our cage.
先验既是我们的财富，也是我们的囚笼。
Value of Information (VOI)信息价值
Source: Originating in the mid-twentieth century, representing a paradigm shift in statistics and management science.
Definition: Value of information refers to the usefulness of information for actual decision-making: only when information can change your action does it have value. It measures the average additional gain from making the best choice with information versus without it.
Application: Prediction markets and arbitrage: capture high-value information for stable returns; business management: focus on internal bottlenecks, process optimization, and key decision data, not just macro news or hot topics; personal life: choose information that improves key actions, not just to satisfy curiosity or fear of missing out.
Insight: It’s fine to be a “knowledge person,” but if you want to get things done, you need VOI awareness.
做个“知道分子”也挺好，但如果你想做点实事儿，你得有VOI意识。
It compares the “drifter,” who gets sucked into the latest hot news and FOMO, to the “sage,” who knows what really matters. It points out that just scrolling through info isn’t the same as actually doing something; only the stuff that actually helps you make choices and take action is worth it, while everything else is just background noise.
Systems & Structure
How the world is organized: interaction systems, environment structure, and feedback loops.
Field Theory场域理论
Source: Proposed in the 1970s by French sociologist Pierre Bourdieu.
Definition: Field theory views society as composed of multiple relatively independent “fields,” each a network of relationships or a competitive arena. Position, resources, and rules within a field determine success or failure. Success depends not only on ability or effort, but on understanding and following the field’s orthodox beliefs and habits.
Application: Identify the structure and key positions of your field, understand the field’s default beliefs and rules; reflect on your own habits to judge compatibility with the field; accumulate capital valued by the field; strategically choose to adapt to or change the field to achieve growth and influence.
Insight: Effort is not hard currency; compliance is.
努力不是硬通货，合规才是。
Resonance / Resonanz共鸣
Source: Proposed by Hartmut Rosa (1965–) in the mid-to-late 2010s.
Definition: Resonance is the mutual response and resonance between independent subjects, accompanied by a sense of mission. It means you are not just outputting, but receiving feedback in interaction. Resonance is divided into: “horizontal resonance” (between people), “diagonal resonance” (between people and things or work), and “vertical resonance” (between people and something greater).
Application: Transform comparative narratives in life into resonance narratives, such as sharing experiences or achieving goals together; find resonance points in career, art, or daily life to combine mission and joy.
Insight: Resonance requires you to allow the world to have its own voice: let materials talk back, let children be disobedient, let the market slap you, let a relationship take you somewhere you didn’t plan.
Source: ① Economics, Game Theory, and positive-sum thinking; ② Jon Levy’s 2025 book Team Intelligence.
Definition: Supply-side mindset means treating yourself as a module that provides verifiable value, proactively reducing collaboration friction, and embedding yourself in long-term repeated games and network effect structures. It emphasizes:
① Value creation: you must truly solve problems, not just talk;
② Friction elimination: make it easier and smoother for others to work with you;
③ Network reach: what level of collaborative circles you can access.
Application: Workplace: demonstrate problem-solving ability, provide practical solutions; Family and close relationships: optimize division of labor and processes, improve overall happiness; Global and policy: supply technology, standards, and systems, rather than zero-sum competition for resources.
Insight: In this highly interconnected, information-replicable, and complementary modern society, “being needed” is safer than “owning”.
在这个高度互联、信息可复制、充满互补性的现代社会，「被需要」是比「拥有」更安全的状态。
Free Energy Principle自由能原理
Source: First proposed around 2005 by Karl Friston.
Definition: The free energy principle holds that living beings maintain structure and boundaries by minimizing “free energy” (roughly equivalent to surprise), i.e., by actively predicting the environment and adjusting themselves or the environment to survive and remain stable.
Application: Individuals reduce surprise through perceptual and active inference, achieving efficient learning, habit improvement, and psychological stability; organizations and companies lower system free energy by collecting information, adjusting strategies, and actions to maintain adaptability; education and behavior design can provide controllable surprise and environmental stability to improve attention and learning efficiency.
Insight: To live is to align yourself with the environment in both directions.
活着就是让自己跟环境双向「对齐」。
Uncertainty & Risk
How reality behaves under noise: randomness, risk distribution, and long-term survival under uncertainty.
No Free Lunch Theorem无免费午餐定理
Source: Originating from algorithm optimization theory in computer science, proposed in 1997 by David Wolpert and William Macready.
Definition: The No Free Lunch Theorem states: there is no universally optimal decision or algorithm. Any method that performs well in a specific domain will necessarily perform poorly in others; optimization comes at a cost, and the effectiveness of decisions depends on individual biases and prior assumptions.
Application: Understand that there is no universally optimal decision in life and work; emphasize that decisions must first set values and biases, clarify goals and domains; conduct research, reasoning, and action under limited information and uncertainty; guide individuals to actively choose in adventures and adjust or exit decisions when necessary.
Insight: All things are impermanent, and decisions are always biased.
诸行无常，决策必有偏置。
Probability Distribution概率分布
Source: ① Foundations in statistics and probability theory; ② Research on outcome bias in psychology; ③ Decision theory and behavioral economics, such as Thinking in Bets by Annie Duke; ④ Practical case studies in decision systems, such as systematic thinking by Scott Adams.
Definition: Probability distribution is a mathematical function describing the possible outcomes of a decision or event and their probabilities. It focuses not only on single outcomes but on the full range of future possibilities. The core of decision-making is not to pick the single “best” outcome, but to manage the overall distribution and understand the uncertainty of risks and opportunities.
Application: Decision-makers analyze probability distributions rather than single outcomes to evaluate strategies; manage tail risks and volatility in career choices, investment, healthcare, insurance, etc.; build and optimize systematic behaviors or skill accumulation, focusing on long-term probability distribution optimization; help cultivate stable, rational decision temperament, and avoid overreacting to random outcomes.
Insight: After the arrow leaves the string, whether it hits the target or not, you must maintain a certain calm indifference—because that’s a sample of wind and noise.
箭离弦之后，是否正中靶心，你要保持某种冷静的漠然——因为那是风向和噪声的抽样。
Kelly Criterion凯利公式
Source: The 1956 paper A New Interpretation of Information Rate by John L. Kelly Jr. of Bell Laboratories.
Definition: The Kelly Criterion is a formula for calculating the optimal bet size in uncertain, repeatable, multiplicative environments. It uses win probability, odds, and cognitive advantage to decide how much to invest each time, maximizing compound growth over the long term while avoiding bankruptcy.
Application: The Kelly Criterion turns cognitive advantage into action size: increase investment when you have an edge, don’t bet when you don’t. It can be used for investment decisions, career choices, time allocation, and trust management; the core is to size positions based on the advantage, pursuing long-term compound growth rather than single wins or losses.
Insight: From the Kelly Criterion’s perspective, the fundamental freedom in life is always having the ability to make the next bet.
在凯利公式看来，人生的根本自由是你始终有下一次下注的能力。
Background: Bell Labs in 1956. Kelly transformed the communication noise problem into a gambling scenario and derived the optimal bet formula f* = (bp – q) / b. Emphasizes that the Kelly Criterion is the combination of information theory and capital compounding, and that growth is limited by cognitive bandwidth.
Non-Ergodicity非遍历性
Source: ① Ergodicity theory in statistical physics and probability theory; ② Research on economic systems by physicist Ole Peters; ③ Analysis of time averages versus ensemble averages by Murray Gell-Mann; ④ Discussions on risk theory by Nassim Nicholas Taleb.
Definition: Non-ergodicity means that the average result of a system as a whole does not represent the true fate of individuals over time. In multiplicative growth environments, the ensemble average may be positive, but the time average experienced by individuals may keep declining or even lead to ruin.
Application: In investment, entrepreneurship, and other multiplicative worlds, individuals must control variance and avoid going to zero: reduce frequent trading, use barbell strategies or Kelly sizing, diversify risk through index investing or risk sharing, thus lowering the probability of ruin from non-ergodicity.
Insight: The multiplicative world is full of non-ergodicity risk, which is bad for individuals but good for the house.
乘法世界中充满了非遍历性风险，它对个体很不利但是对庄家很有利。
Using Bitcoin as an example, it highlights how wild price swings from 2015 to 2026 hit different kinds of investors. It points out that non-ergodicity is key here: folks who don’t sweat the short-term ups and downs, have a long-term view, or plenty of cash to back them up are the ones who really do well in investing. Just remember, one big crash can totally wipe out everyday people!
Optimization & Growth
How performance compounds: compounding advantage, learning efficiency, and cognitive performance scaling.
Compound Interest复利
Source: ① Thomas Piketty’s discussion of r > g in Capital in the Twenty-First Century; ② Pierre Bourdieu’s theory of multiple forms of capital.
Definition: Compound interest means interest is calculated not only on the principal but also on accumulated interest, forming exponential growth over time. It applies not only to money but also to knowledge, skills, health, relationships, and other forms of capital. The key is to start early and persist long-term to form an uncatchable advantage.
Application: Compound interest helps accumulate wealth in financial investment and accumulate various forms of capital in personal growth: invest in human capital (knowledge, skills) in youth; deepen expertise, reputation, and social capital in middle age; focus on health, experience transfer, and influence in later years; ROI determines accumulation priority.
Insight: Interest rate is on you, compound interest grows in the system.
利率就在你身上，复利长在系统里。
Active High Cognitive Load主动高认知负荷
Source: Developed from the Cognitive Load Theory (CLT) proposed in 1988 by John Sweller.
Definition: Active high cognitive load is a thinking mode that deliberately mobilizes a large amount of attention resources to handle complex, high-uncertainty tasks. It increases germane load, enabling the brain to build, integrate, and optimize internal models, thus entering flow and improving output and learning efficiency.
Application: In study or work, occupy working memory with difficult tasks to achieve deep thinking and constructive learning; in daily life, upgrade ordinary activities to analysis, reasoning, or reverse engineering tasks to create cognitive challenges; in organizations and project management, improve team attention allocation and complex problem-solving ability.
Insight: Active high cognitive load means not waiting for the world to give you problems, but turning the world itself into a problem.
主动高认知负荷就是不等待世界给你难题，而是把世界本身变成一道难题。
This dictionary is not about providing final answers. Instead, it helps readers ask better questions and think more clearly before making decisions.
Reference
万维钢 (2026)，《现代思维工具 100 讲》与《现代思维工具辞典》。
Wan Weigang (2026).100 Lectures on Modern Thinking Tools & A Dictionary of Modern Mental Models.
Recently, Y Combinator hosted Bob McGrew, the former Chief Research Officer at OpenAI and a veteran technologist from PayPal and Palantir. What surprised many was the line of questioning. Instead of asking him how to build the next GPT, founders kept pressing him on a very different topic: Palantir’s FDE model.
Bob admitted that over the past year, nearly every startup he’s advised has been obsessed with learning how this model works in practice.
What Exactly Is FDE?
FDE (Forward Deployed Engineer) is a model where engineers embed directly with customers to bridge the gap between what the product aspires to be and what the customer actually needs.
The idea traces back to Palantir’s early days working with U.S. intelligence agencies. The challenges were messy, complex, and had no off-the-shelf solutions. The only way forward was to “build on the ground” with the client. At the time, many dismissed it as unscalable, labor-intensive, and far from the clean SaaS ideal. Fast forward to today, and the very same approach is being embraced by AI startups building agents and enterprise solutions.
How It Works
Palantir structured its FDE teams around two roles:
Echo: the industry-savvy operator who lives inside the customer’s workflow, identifies core pain points, and challenges the status quo.
Delta: the technical builder who can spin up prototypes quickly, solving problems in real time.
Meanwhile, the core product team back at HQ takes these frontline hacks and turns them into platform features. Think of it as paving a permanent road where the FDEs first laid down gravel.
Why It Matters
The strength of the FDE model is that it forges unusually deep relationships with customers. It surfaces real market demand—things no survey or user interview could ever uncover. Done right, it creates a defensible moat.
But it’s also risky. Without discipline, FDE can collapse into traditional consulting or body-shop outsourcing. The litmus test of a healthy model is whether the core platform keeps evolving, making each new deployment faster, cheaper, and more scalable.
Different from Consulting
The distinction is critical:
Consulting delivers one-off solutions.
FDE is about feeding learnings back into the product, so the platform gets stronger with every customer.
This feedback loop—and the ability of product managers to abstract from bespoke requests—is what turns customer-specific fixes into reusable product capabilities.
Why AI Startups Love It
For AI Agent companies, the market is far too fragmented and unpredictable for a “one-size-fits-all” solution. No universal product exists. Embedding deeply with customers isn’t optional—it’s the only way to figure out what works, discover product-market fit, and build enduring platforms.
A Shift in Business Models
Unlike traditional SaaS, which scales on pure subscriptions, FDE contracts are more outcome-driven and flexible. The key lever is product leverage: doing the same amount of frontline work but translating it into larger contracts and less marginal customization over time.
The Bigger Picture
The rise of FDE highlights a paradox of modern tech: at scale, the best companies keep doing the things that “don’t scale.” The gulf between breakthrough AI capabilities and messy, real-world adoption is exactly where the biggest opportunities lie today.
It’s not an easy path—more trench warfare than blitzscaling—but for founders, it may be the only one that works.
This blog examines the rapidly evolving landscape of AI-powered search, comparing Google’s recent transformation with its Search Generative Experience (SGE) and Gemini integration against Perplexity AI‘s native AI-first approach. Both companies now leverage large language models, but with fundamentally different architectures and philosophies.
The New Reality: Google has undergone a dramatic transformation from traditional keyword-based search to an AI-driven conversational answer engine. With the integration of Gemini, LaMDA, PaLM, and the rollout of AI Overviews (formerly SGE), Google now synthesizes information from multiple sources into concise, contextual answers—directly competing with Perplexity’s approach.
Key Findings:
Convergent Evolution: Both platforms now use LLMs for answer generation, but Google maintains its traditional search infrastructure while Perplexity was built AI-first from the ground up
Architecture Philosophy: Google integrates AI capabilities into its existing search ecosystem (hybrid approach), while Perplexity centers everything around RAG and multi-model orchestration (AI-native approach)
AI Technology Stack: Google leverages Gemini (multimodal), LaMDA (conversational), and PaLM models, while Perplexity orchestrates external models (GPT, Claude, Gemini, Llama, DeepSeek)
User Experience: Google provides AI Overviews alongside traditional search results, while Perplexity delivers answer-first experiences with citations
Market Dynamics: The competition has intensified with Google’s AI transformation, making the choice between platforms more about implementation philosophy than fundamental capabilities
This represents a paradigm shift where the question is no longer “traditional vs. AI search” but rather “how to best implement AI-powered search” with different approaches to integration, user experience, and business models.
Keywords: AI Search, RAG, Large Language Models, Search Architecture, Perplexity AI, Google Search, Conversational AI, SGE, Gemini.
Google’s AI Transformation: From PageRank to Gemini-Powered Search
Google has undergone one of the most significant transformations in its history, evolving from a traditional link-based search engine to an AI-powered answer engine. This transformation represents a strategic response to the rise of AI-first search platforms and changing user expectations.
The Search Generative Experience (SGE) Revolution
Google’s Search Generative Experience (SGE), now known as AI Overviews, fundamentally changes how search results are presented:
AI-Synthesized Answers: Instead of just providing links, Google’s AI generates comprehensive insights, explanations, and summaries from multiple sources
Contextual Understanding: Responses consider user context including location, search history, and preferences for personalized results
Multi-Step Query Handling: The system can handle complex, conversational queries that require reasoning and synthesis
Real-Time Information Grounding: AI overviews are grounded in current, real-time information while maintaining accuracy
Google’s LLM Arsenal
Google has strategically integrated multiple advanced AI models into its search infrastructure:
Gemini: The Multimodal Powerhouse
Capabilities: Understands and generates text, images, videos, and audio
Search Integration: Enables complex query handling including visual search, reasoning tasks, and detailed information synthesis
Multimodal Processing: Handles queries that combine text, images, and other media types
LaMDA: Conversational AI Foundation
Purpose: Powers natural, dialogue-like interactions in search
Features: Enables follow-up questions and conversational context maintenance
Role: Provides advanced language processing capabilities
Applications: Powers complex reasoning, translation (100+ languages), and contextual understanding
Scale: Handles extended documents and multimodal inputs
Technical Architecture Integration
Google’s approach differs from AI-first platforms by layering AI capabilities onto existing infrastructure:
Key Differentiators of Google’s AI Search
Hybrid Architecture: Maintains traditional search capabilities while adding AI-powered features
Scale Integration: Leverages existing massive infrastructure and data
DeepMind Synergy: Strategic integration of DeepMind research into commercial search applications
Continuous Learning: ML ranking algorithms and AI models learn from user interactions in real-time
Global Reach: AI features deployed across 100+ languages with localized understanding
Perplexity AI Architecture: The RAG-Powered Search Revolution
Perplexity AI represents a fundamental reimagining of search technology, built on three core innovations:
Retrieval-Augmented Generation (RAG): Combines real-time web crawling with large language model capabilities
Multi-Model Orchestration: Leverages multiple AI models (GPT, Claude, Gemini, Llama, DeepSeek) for optimal responses
Integrated Citation System: Provides transparent source attribution with every answer
The platform offers multiple access points to serve different user needs: Web Interface, Mobile App, Comet Browser, and Enterprise API.
Core Architecture Components
Simplified Architecture View
For executive presentations and high-level discussions, this three-layer view highlights the essential components:
How Perplexity Works: From Query to Answer
Understanding Perplexity’s workflow reveals why it delivers fundamentally different results than traditional search engines. Unlike Google’s approach of matching keywords to indexed pages, Perplexity follows a sophisticated multi-step process:
The Eight-Step Journey
Query Reception: User submits a natural language question through any interface
Real-Time Retrieval: Custom crawlers search the web for current, relevant information
Source Indexing: Retrieved content is processed and indexed in real-time
Context Assembly: RAG system compiles relevant information into coherent context
Model Selection: AI orchestrator chooses the optimal model(s) for the specific query type
Answer Generation: Selected model(s) generate comprehensive responses using retrieved context
Citation Integration: System automatically adds proper source attribution
Response Delivery: Final answer with citations is presented to the user
Technical Workflow Diagram
The sequence below shows how a user query flows through Perplexity’s system.
This process typically completes in under 3 seconds, delivering both speed and accuracy.
The New Search Paradigm: AI-First vs AI-Enhanced Approaches
The competition between Google and Perplexity has evolved beyond traditional vs. AI search to represent two distinct philosophies for implementing AI-powered search experiences.
The Future of AI-Powered Search: A New Competitive Landscape
The integration of AI into search has fundamentally changed the competitive landscape. Rather than a battle between traditional and AI search, we now see different approaches to implementing AI-powered experiences competing for user mindshare and market position.
Implementation Strategy Battle: Integration vs. Innovation
Google’s Integration Strategy:
Advantage: Massive user base and infrastructure to deploy AI features at scale
Challenge: Balancing AI innovation with existing business model dependencies
Approach: Gradual rollout of AI features while maintaining traditional search options
Perplexity’s Innovation Strategy:
Advantage: Clean slate design optimized for AI-first experiences
Challenge: Building user base and competing with established platforms
Approach: Focus on superior AI experience to drive user acquisition
The Multi-Modal Future
Both platforms are moving toward comprehensive multi-modal experiences:
Visual Search Integration: Google Lens vs. Perplexity’s image understanding capabilities
Voice-First Interactions: Google Assistant integration vs. conversational AI interfaces
Video and Audio Processing: Gemini’s multimodal capabilities vs. orchestrated model approaches
Document Intelligence: Enterprise document search and analysis capabilities
Business Model Evolution Under AI
Advertising Model Transformation:
Google must adapt its ad-centric model to AI Overviews without disrupting user experience
Challenge of monetizing direct answers vs. traditional click-through advertising
Need for new ad formats that work with conversational AI
Subscription and API Models:
Perplexity’s success with subscription tiers validates alternative monetization
Growing enterprise demand for AI-powered search APIs and integrations
Despite different starting points, both platforms are converging on similar technical capabilities:
Real-Time Information: Both now emphasize current, up-to-date information retrieval
Source Attribution: Transparency and citation becoming standard expectations
Conversational Context: Multi-turn conversation support across platforms
Model Diversity: Google developing multiple specialized models, Perplexity orchestrating external models
The Browser and Distribution Channel Wars
Perplexity’s Chrome Acquisition Strategy:
$34.5B all-cash bid for Chrome represents unprecedented ambition in AI search competition
Strategic Value: Control over browser defaults, user data, and search distribution
Market Impact: Success would fundamentally alter competitive dynamics and user acquisition costs
Regulatory Reality: Bid likely serves as strategic positioning and leverage rather than realistic acquisition
Alternative Distribution Strategies:
AI-native browsers (Comet) as specialized entry points
API integrations into enterprise and developer workflows
Mobile-first experiences capturing younger user demographics
Strategic Implications and Future Outlook
The competition between Google’s AI-enhanced approach and Perplexity’s AI-native strategy represents a fascinating case study in how established platforms and startups approach technological transformation differently.
Key Strategic Insights
The AI Integration Challenge: Google’s transformation demonstrates that even dominant platforms must fundamentally reimagine their core products to stay competitive in the AI era
Architecture Philosophy Matters: The choice between hybrid integration (Google) vs. AI-first design (Perplexity) creates different strengths, limitations, and user experiences
Business Model Pressure: AI-powered search challenges traditional advertising models, forcing experimentation with subscriptions, APIs, and premium features
User Behavior Evolution: Both platforms are driving the shift from “search and browse” to “ask and receive” interactions, fundamentally changing how users access information
The New Competitive Dynamics
Advantages of Google’s AI-Enhanced Approach:
Massive scale and infrastructure for global AI deployment
Existing user base to gradually transition to AI features
Deep integration with knowledge graphs and proprietary data
Ability to maintain traditional search alongside AI innovations
Advantages of Perplexity’s AI-Native Approach:
Optimized user experience designed specifically for conversational AI
Agility to implement cutting-edge AI techniques without legacy constraints
Model-agnostic architecture leveraging best-in-class external AI models
Clear value proposition for users seeking direct, cited answers
Looking Ahead: Industry Predictions
Near-Term (1-2 years):
Continued convergence of features between platforms
Google’s global rollout of AI Overviews across all markets and languages
Perplexity’s expansion into enterprise and specialized vertical markets
Emergence of more AI-native search platforms following Perplexity’s model
Medium-Term (3-5 years):
AI-powered search becomes the standard expectation across all platforms
Specialized AI search tools for professional domains (legal, medical, scientific research)
Integration of real-time multimodal capabilities (live video analysis, augmented reality search)
New regulatory frameworks for AI-powered information systems
Long-Term (5+ years):
Fully conversational AI assistants replace traditional search interfaces
Personal AI agents that understand individual context and preferences
Integration with IoT and ambient computing for seamless information access
Potential emergence of decentralized, blockchain-based search alternatives
Recommendations for Stakeholders
For Technology Leaders:
Hybrid Strategy: Consider Google’s approach of enhancing existing systems with AI rather than complete rebuilds
Model Orchestration: Investigate Perplexity’s approach of orchestrating multiple AI models for optimal results
Real-Time Capabilities: Invest in real-time information retrieval and processing systems
Citation Systems: Implement transparent source attribution to build user trust
For Business Strategists:
Revenue Model Innovation: Experiment with subscription, API, and premium feature models beyond traditional advertising
User Experience Focus: Prioritize conversational, answer-first experiences in product development
Distribution Strategy: Evaluate the importance of browser control and default search positions
Competitive Positioning: Decide between AI-enhancement of existing products vs. AI-native alternatives
For Investors:
Platform Risk Assessment: Evaluate how established platforms are adapting to AI disruption
Technology Differentiation: Assess the sustainability of competitive advantages in rapidly evolving AI landscape
Business Model Viability: Monitor the success of alternative monetization strategies beyond advertising
Regulatory Impact: Consider potential regulatory responses to AI-powered information systems and search market concentration
The future of search will be determined by execution quality, user adoption, and the ability to balance innovation with practical business considerations. Both Google and Perplexity have established viable but different paths forward, setting the stage for continued innovation and competition in the AI-powered search landscape.
Monitor the browser control battle and distribution channel acquisitions
Technology Differentiation: Assess the sustainability of competitive advantages in rapidly evolving AI landscape
Business Model Viability: Monitor the success of alternative monetization strategies beyond advertising
Regulatory Impact: Consider potential regulatory responses to AI-powered information systems and search market concentration
Conclusion
The evolution of search from Google’s traditional PageRank-driven approach to today’s AI-powered landscape represents one of the most significant technological shifts in internet history. Google’s recent transformation with its Search Generative Experience and Gemini integration demonstrates that even the most successful platforms must reinvent themselves to remain competitive in the AI era.
The competition between Google’s AI-enhanced strategy and Perplexity’s AI-native approach offers valuable insights into different paths for implementing AI at scale. Google’s hybrid approach leverages massive existing infrastructure while gradually transforming user experiences, while Perplexity’s clean-slate design optimizes entirely for conversational AI interactions.
As both platforms continue to evolve, the ultimate winners will be users who gain access to more intelligent, efficient, and helpful ways to access information. The future of search will likely feature elements of both approaches: the scale and comprehensiveness of Google’s enhanced platform combined with the conversational fluency and transparency of AI-native solutions.
The battle for search supremacy in the AI era has only just begun, and the innovations emerging from this competition will shape how humanity accesses and interacts with information for decades to come.
This analysis reflects the state of AI-powered search as of August 2025. The rapidly evolving nature of AI technology and competitive dynamics may significantly impact future developments. Both Google and Perplexity continue to innovate at unprecedented pace, making ongoing monitoring essential for stakeholders in this space.This analysis represents the current state of AI-powered search as of August 2025. The rapidly evolving nature of AI technology and competitive landscape may impact future developments.
Part One: The Singularity The story begins on the Moon, beside China’s Tianhe Base, with the sudden appearance of a black obelisk. It was unquestionably not of human origin from the very first day of its discovery. Composed of an unidentifiable, perfectly smooth black material, it reflected no light and emitted no heat, as if it were a three-dimensional void against the cosmic background. The astronauts who discovered it named it “Witness.”
For twenty years, human scientists exhausted every technological means to study it, yet not a single atom could be removed from its surface. As research approached a deadlock, teetering on the edge of becoming a symbolic relic, a “point” was discovered.
On the side of the obelisk facing Earth, at its exact center, there was a point.
It was neither a mark nor a dent nor a protrusion. It seemed intrinsic to the material itself—a geometric perfection made manifest. The point was discovered by quantum metrologist Dr. Yun Tianming during a holographic surface scan aimed at mapping quantum fluctuations on the obelisk. Amid the torrent of data, he identified an absolute “nothingness,” a singularity with zero information entropy.
When the image was translated into visible-light models, the point appeared there: a perfect, dimensionless point.
The following decade became the most maddening ten years in physics.
The team first used an atomic force microscope (AFM) to examine the point at the nanoscale, hoping to resolve its edge structure. By conventional expectation, any solid surface should reveal electron cloud distributions and quantum fluctuations. Yet the force curves remained perfectly flat, devoid of noise or disturbance, as if the probe were suspended in a vacuum, unable to detect any structural signal.
Next, they turned to scanning tunneling microscopy (STM) to measure the local density of electronic states. Regardless of voltage adjustments, the tunneling current remained zero—no energy levels for electrons to occupy, as though the region did not belong to the three-dimensional material world.
To rule out instrument limitations, the team deployed laser interferometry to approach the precision of the Planck length. Still, the data remained perfectly symmetrical: the distances from the point to the four edges of the obelisk were exactly equal—not approximately, but to a precision beyond the limits of quantum measurement. Every terminal value in the dataset entered an infinite loop of zeros, seemingly mocking humanity’s grasp of physical law.
“This makes no sense,” Yun murmured to his colleague Dr. Cheng Xin after countless sleepless nights. “According to the Heisenberg uncertainty principle, we cannot determine a particle’s position with infinite precision. The very existence of this point undermines the foundations of physics itself. It is an ontological miracle, something that should not exist.”
Cheng Xin pointed to the rotating holographic obelisk model, streams of data cascading like a waterfall. “Perhaps we’re approaching this incorrectly, Tianming. We keep trying to measure it, treating it as part of our universe. But… what if it’s not?”
Part Two: The Anchor of Dimensions Cheng Xin’s words struck Yun like lightning. He began feverishly developing new mathematical models. No longer did he consider the obelisk a three-dimensional object; instead, he hypothesized it was a higher-dimensional entity “sliced” into our three-dimensional space.
“Imagine this,” he explained at an international physics conference, his holographic presence tinged with fervor, “an infinitely thin needle piercing through an infinitely large sheet of paper. For two-dimensional beings on the paper, they would perceive only a perfect point. No matter how precise their measurement tools, the point would always appear at the ‘center’ they can perceive. They cannot comprehend the needle, because the third dimension is beyond them.”
His theory caused a stir. Most dismissed it as philosophical speculation. Yet it perfectly explained the point’s “perfect centering.”
“This point,” Yun continued, “is not a feature on the obelisk’s surface. It is the obelisk itself! Or rather, it is the projection of a higher-dimensional object’s ‘axis’ into our universe. We are not measuring a point on a two-dimensional plane; we are gazing upon a reality-piercing, higher-dimensional spine.”
The theory became known as the “Anchor of Dimensions” Hypothesis. The point anchors a four-dimensional—or even higher-dimensional—object into our three-dimensional space. The civilization that left it had used the simplest, most elegant method to demonstrate a physics beyond our imagination.
They were saying: You exist, but not in the space you can perceive.
Part Three: The Response How could this hypothesis be tested? It could not be verified by measurement. Yun proposed a bold experiment: do not measure the point’s “position,” but perturb the reality around it.
A massive ring-shaped device was constructed around the Witness. It emitted no particles or energy, but generated an extraordinarily precise, twisted spacetime field—a “whisper of gravitational waves.” If the point truly was a higher-dimensional projection, disturbances in our dimension might elicit a response from the anchor.
The day of the experiment drew the eyes of the world to the Moon. Yun and Cheng Xin stood at the control center, their hearts racing.
“Spacetime field generator, 1% power.”
Nothing happened.
“10%… 30%… 70%…”
The obelisk remained silent. The readouts were unchanging, despairingly stable.
“100%.”
A moment of silence.
Suddenly, the room felt gripped by an invisible hand; the air seemed to collapse. Heartbeats across Earth faltered, as if drawn toward a nonexistent direction.
Walls stretched, floors sank, control panels warped, faces elongated into unseeable dimensions. It was a sensation beyond language—like a drowning person inhaling air for the first time, or a blind man suddenly scorched by sunlight.
Then, they saw.
The point was no longer a point but a luminous spine piercing reality, extending into dimensions that could not be named. It was not dazzling, yet clearer than any star.
A torrent of conceptual information flooded their minds—not words, not sound, but pure ideas:
—Very good.
—You have finally abandoned the ruler and begun measuring the universe with thought.
—This door opens for you. We await on the other side.
The messages faded, and perception collapsed back into three-dimensional space. The room was unchanged; instruments stable. Only their breathing and trembling eyes revealed the magnitude of what had occurred. It was as if they had been swept by a cosmic tsunami and returned to shore.
Part Four: The Beginning The secret of the Witness was revealed. It was not a monument, a warning, or a work of art.
It was a test: the most concise, ruthless, and elegant test.
The civilization that left it used a perfect geometric singularity to filter cosmic civilizations. Only when a species transcended three-dimensional thinking and began to understand higher dimensions could it “graduate” and earn the invitation to higher-dimensional existence.
Humanity took thirty years to solve the puzzle. Thirty years to earn a single answer.
Yun Tianming and Cheng Xin stood by the viewport, gazing at the serene black obelisk. That point, the enigma that had tormented generations of scientists, was no longer a point—it was a nail, pinning human civilization onto the test paper of the cosmos.
Perhaps countless intelligent species in innumerable galaxies had faced similar points. Some solved them, some failed, some still wandered the labyrinth. Humanity was merely one example, granted the privilege to step through the doorway to higher dimensions. It was both an invitation and a judgment.
And it began with understanding that perfect, infinitesimal point. Now, the gaze of all humanity was drawn to the invisible line extending to higher dimensions, calling them onward.
Suddenly, Yun shivered: Humanity has been chosen. But does being chosen mean fortunate?
The Moon remained silent; the black obelisk unchanged. A signpost, or perhaps a chain.
Afterword When I was a child, I read a story—I can no longer verify which magazine or author it was—but it left a profound impression on me. In that story, humanity discovered a black obelisk from an extraterrestrial civilization. Though seemingly ordinary, every measurement—height, width, every geometric feature of its surface—conformed to a perfect, infinitely precise golden ratio.
Scientists exhausted themselves trying to decode any physical message: cosmic coordinates, mathematical formulas, or warnings. Ultimately, they realized the obelisk was the information itself. It was not language, but a tool for measuring and filtering. This civilization had elegantly, nonviolently, demonstrated a force beyond physical scale.
This story fascinated me and inspired my creation of The Singularity. I further concretized the idea of perfection, transforming it into a dimensionless point—a miracle challenging the foundations of physics. It is not a display of technology, but a test of human thought itself.
This work pays homage to that childhood story, reminding us that the deepest cosmic mysteries may not lie in distant stars, but in the simplest of concepts.
“On the Sincerity and Mastery in Large Models” is a two-part essay inspired by Sun Simiao’s classical Chinese text On the Absolute Sincerity of Great Physicians. Written in classical Chinese style, it warns against superficial understanding and blind faith in large language models (LLMs). It calls for practitioners to uphold a spirit of diligence (“精”) and sincerity (“诚”)—to understand the inner principles of algorithms and the biases within data. The model is but a tool; its moral compass lies in the human operator. Only by combining technical rigor with ethical restraint can AI serve humanity and avoid causing harm. This is both a philosophical treatise on AI and a critique of today’s hasty tech culture.
Mark Zuckerberg’s recent move to bring Alex Wang and his team into Meta represents a bold and strategic maneuver amid the rapid advancement of large models and AGI development. Putting aside the ethical considerations, Zuckerberg’s approach—laying off staff, then offering sky-high compensation packages with a 48-hour ultimatum to Top AI scientists and engineers from OpenAI , alongside Meta’s acquisition of a 49% stake in Scale AI—appears to serve multiple objectives:
1. Undermining Competitors
By poaching key talent from rival companies, Meta not only weakens their R&D teams and disrupts their momentum but also puts pressure on Google, OpenAI, and others to reassess their partnerships with Scale AI. Meta’s investment may further marginalize these competitors by injecting uncertainty into their collaboration with Scale AI.
2. Reinvigorating the Internal Team
Bringing in fresh blood like Alex Wang’s team and Open AI Top talents could reenergize Meta’s existing research units. A successful “talent reset” may help the company gain a competitive edge in the race toward AGI.
3. Enhancing Brand Visibility
Even if the move doesn’t yield immediate results, it has already amplified Meta’s media presence, boosting its reputation as a leader in AI innovation.
From both a talent acquisition and PR standpoint, this appears to be a masterstroke for Meta.
However, the strategy is not without significant risks:
1. Internal Integration and Morale Challenges
The massive compensation packages offered to those talents could trigger resentment among existing employees—especially in the wake of recent layoffs—due to perceived pay inequity. This may lower morale and even accelerate internal attrition. Cultural differences between the incoming and incumbent teams could further complicate internal integration and collaboration.
2. Return on Investment and Performance Pressure
Meta’s substantial investment in Alex Wang and Scale AI comes with high expectations for short-term deliverables. In a domain as uncertain as AGI, both the market and shareholders will be eager for breakthroughs. If Wang’s team fails to deliver measurable progress quickly, Meta could face mounting scrutiny and uncertainty over the ROI.
3. Impacts on Scale AI and the Broader Ecosystem
Alex Wang stepping away as CEO is undoubtedly a major loss for Scale AI, even if he retains a board seat. Leadership transitions and potential talent departures may follow. Moreover, Scale AI’s history of legal and compliance issues could reflect poorly on Meta’s brand—especially if public perception ties Meta to those concerns despite holding only non-voting shares. More broadly, Meta’s aggressive “poaching” approach may escalate the AI talent war, drive up industry-wide costs, and prompt renewed debate over ethics and hiring norms in the AI sector.
Conclusion Meta’s latest move is undeniably ambitious. While it positions the company aggressively in the AGI race, it also carries notable risks in terms of internal dynamics, ROI pressure, and broader ecosystem disruption. Only time will tell whether this bold gamble pays off.
应对人工智能的独特错误模式：布鲁斯·施奈尔（Bruce Schneier）和内森·E·桑德斯（Nathan E. Sanders）（网络安全视角）指出，人工智能系统，特别是大型语言模型（LLMs），其错误模式与人类错误显著不同——它们更难预测，不集中在知识空白处，且缺乏对自身错误的自我意识。他们提出双重研究方向：一是工程化人工智能以产生更易于人类理解的错误（例如，通过RLHF等精炼的对齐技术）；二是开发专门针对人工智能独特“怪异”之处的新型安全与纠错系统（例如，迭代且多样化的提示）。
Abstract: this article presents three distinct, cross-disciplinary strategies for ensuring the safe and beneficial development of Artificial Intelligence.
Addressing Idiosyncratic AI Error Patterns (Cybersecurity Perspective): Bruce Schneier and Nathan E. Sanders highlight that AI systems, particularly Large Language Models (LLMs), exhibit error patterns significantly different from human mistakes—being less predictable, not clustered around knowledge gaps, and lacking self-awareness of error. They propose a dual research thrust: engineering AIs to produce more human-intelligible errors (e.g., through refined alignment techniques like RLHF) and developing novel security and mistake-correction systems specifically designed for AI’s unique “weirdness” (e.g., iterative, varied prompting).
Updating Ethical Frameworks to Combat AI Deception (Robotics & Internet Culture Perspective): Dariusz Jemielniak argues that Isaac Asimov’s traditional Three Laws of Robotics are insufficient for modern AI due to the rise of AI-enabled deception, including deepfakes, sophisticated misinformation campaigns, and manipulative AI interactions. He proposes a “Fourth Law of Robotics”: A robot or AI must not deceive a human being by impersonating a human being.Implementing this law would necessitate mandatory AI disclosure, clear labeling of AI-generated content, technical identification standards, legal enforcement, and public AI literacy initiatives to maintain trust in human-AI collaboration.
Establishing Rigorous Protocols for AGI Detection and Interaction (Astrobiology/SETI Perspective): Edmon Begoli and Amir Sadovnik suggest that research into Artificial General Intelligence (AGI) can draw methodological lessons from the Search for Extraterrestrial Intelligence (SETI). They advocate for a structured scientific approach to AGI that includes:
Developing clear, multidisciplinary definitions of “general intelligence” and related concepts like consciousness.
Creating robust, novel metrics and evaluation benchmarks for detecting AGI, moving beyond limitations of tests like the Turing Test.
Formulating internationally recognized post-detection protocols for validation, transparency, safety, and ethical considerations, should AGI emerge.
Collectively, these perspectives emphasize the urgent need for innovative, multi-faceted approaches—spanning security engineering, ethical guideline revision, and rigorous scientific protocol development—to proactively manage the societal integration and potential future trajectory of advanced AI systems.
Here are the full detailed content：
3 Ways to Keep AI on Our Side
AS ARTIFICIAL INTELLIGENCE reshapes society, our traditional safety nets and ethical frameworks are being put to the test. How can we make sure that AI remains a force for good? Here we bring you three fresh visions for safer AI.
In the first essay, security expert Bruce Schneier and data scientist Nathan E. Sanders explore how AI’s “weird” error patterns create a need for innovative security measures that go beyond methods honed on human mistakes.
Dariusz Jemielniak, an authority on Internet culture and technology, argues that the classic robot ethics embodied in Isaac Asimov’s famous rules of robotics need an update to counterbalance AI deception and a world of deepfakes.
And in the final essay, the AI researchers Edmon Begoli and Amir Sadovnik suggest taking a page from the search for intelligent life in the stars; they propose rigorous standards for detecting the possible emergence of human-level AI intelligence.
As AI advances with breakneck speed, these cross-disciplinary strategies may help us keep our hands on the reins.
AI Mistakes Are Very Different from Human Mistakes
WE NEED NEW SECURITY SYSTEMS DESIGNED TO DEAL WITH THEIR WEIRDNESS
Bruce Schneier & Nathan E. Sanders
HUMANS MAKE MISTAKES all the time. All of us do, every day, in tasks both new and routine. Some of our mistakes are minor, and some are catastrophic. Mistakes can break trust with our friends, lose the confidence of our bosses, and sometimes be the difference between life and death.
Over the millennia, we have created security systems to deal with the sorts of mistakes humans commonly make. These days, casinos rotate their dealers regularly, because they make mistakes if they do the same task for too long. Hospital personnel write on patients’ limbs before surgery so that doctors operate on the correct body part, and they count surgical instruments to make sure none are left inside the body. From copyediting to double-entry bookkeeping to appellate courts, we humans have gotten really good at preventing and correcting human mistakes.
Humanity is now rapidly integrating a wholly different kind of mistakemaker into society: AI. Technologies like large language models (LLMs) can perform many cognitive tasks traditionally fulfilled by humans, but they make plenty of mistakes. You may have heard about chatbots telling people to eat rocks or add glue to pizza. What differentiates AI systems’ mistakes from human mistakes is their weirdness. That is, AI systems do not make mistakes in the same ways that humans do.
Much of the risk associated with our use of AI arises from that difference. We need to invent new security systems that adapt to these differences and prevent harm from AI mistakes.
IT’S FAIRLY EASY to guess when and where humans will make mistakes. Human errors tend to come at the edges of someone’s knowledge: Most of us would make mistakes solving calculus problems. We expect human mistakes to be clustered: A single calculus mistake is likely to be accompanied by others. We expect mistakes to wax and wane depending on factors such as fatigue and distraction. And mistakes are typically accompanied by ignorance: Someone who makes calculus mistakes is also likely to respond “I don’t know” to calculus-related questions.
To the extent that AI systems make these humanlike mistakes, we can bring all of our mistake-correcting systems to bear on their output. But the current crop of AI models—particularly LLMs—make mistakes differently.
AI errors come at seemingly random times, without any clustering around particular topics. The mistakes tend to be more evenly distributed through the knowledge space; an LLM might be equally likely to make a mistake on a calculus question as it is to propose that cabbages eat goats. And AI mistakes aren’t accompanied by ignorance. An LLM will be just as confident when saying something completely and obviously wrong as it will be when saying something true.
The inconsistency of LLMs makes it hard to trust their reasoning in complex, multistep problems. If you want to use an AI model to help with a business problem, it’s not enough to check that it understands what factors make a product profitable; you need to be sure it won’t forget what money is.
THIS SITUATION INDICATES two possible areas of research: engineering LLMs to make mistakes that are more humanlike, and building new mistake-correcting systems that deal with the specific sorts of mistakes that LLMs tend to make.
We already have some tools to lead LLMs to act more like humans. Many of these arise from the field of “alignment” research, which aims to make models act in accordance with the goals of their human developers. One example is the technique that was arguably responsible for the breakthrough success of ChatGPT: reinforcement learning with human feedback. In this method, an AI model is rewarded for producing responses that get a thumbs-up from human evaluators. Similar approaches could be used to induce AI systems to make humanlike mistakes, particularly by penalizing them more for mistakes that are less intelligible.
When it comes to catching AI mistakes, some of the systems that we use to prevent human mistakes will help. To an extent, forcing LLMs to double-check their own work can help prevent errors. But LLMs can also confabulate seemingly plausible yet truly ridiculous explanations for their flights from reason.
Other mistake-mitigation systems for AI are unlike anything we use for humans. Because machines can’t get fatigued or frustrated, it can help to ask an LLM the same question repeatedly in slightly different ways and then synthesize its responses. Humans won’t put up with that kind of annoying repetition, but machines will.
RESEARCHERS ARE still struggling to understand where LLM mistakes diverge from human ones. Some of the weirdness of AI is actually more humanlike than it first appears.
Small changes to a query to an LLM can result in wildly different responses, a problem known as prompt sensitivity. But, as any survey researcher can tell you, humans behave this way, too. The phrasing of a question in an opinion poll can have drastic impacts on the answers.
LLMs also seem to have a bias toward repeating the words that were most common in their training data—for example, guessing familiar place names like “America” even when asked about more exotic locations. Perhaps this is an example of the human “availability heuristic” manifesting in LLMs; like humans, the machines spit out the first thing that comes to mind rather than reasoning through the question. Also like humans, perhaps, some LLMs seem to get distracted in the middle of long documents; they remember more facts from the beginning and end.
In some cases, what’s bizarre about LLMs is that they act more like humans than we think they should. Some researchers have tested the hypothesis that LLMs perform better when offered a cash reward or threatened with death. It also turns out that some of the best ways to “jailbreak” LLMs (getting them to disobey their creators’ explicit instructions) look a lot like the kinds of social-engineering tricks that humans use on each otherfor example, pretending to be someone else or saying that the request is just a joke. But other effective jailbreaking techniques are things no human would ever fall for. One group found that if they used ASCII art (constructions of symbols that look like words or pictures) to pose dangerous questions, like how to build a bomb, the LLM would answer them willingly.
Humans may occasionally make seemingly random, incomprehensible, and inconsistent mistakes, but such occurrences are rare and often indicative of more serious problems. We also tend not to put people exhibiting these behaviors in decision-making positions. Likewise, we should confine AI decision-making systems to applications that suit their actual abilities—while keeping the potential ramifications of their mistakes firmly in mind.
Asimov’s Laws of Robotics Need an Update for AI PROPOSING A FOURTH LAW OF ROBOTICS
Dariusz Jemielniak
IN 1942, the legendary science fiction author Isaac Asimov introduced his Three Laws of Robotics in his short story “Runaround.” The laws were later popularized in his seminal story collection I, Robot.
FIRST LAW: A robot may not injure a human being or, through inaction, allow a human being to come to harm.
SECOND LAW: A robot must obey the orders given it by human beings except where such orders would conflict with the First Law.
THIRD LAW: A robot must protect its own existence as long as such protection does not conflict with the First or Second Law.
While drawn from works of fiction, these laws have shaped discussions of robot ethics for decades. And as AI systems—which can be considered virtual robots—have become more sophisticated and pervasive, some technologists have found Asimov’s framework useful for considering the potential safeguards needed for AI that interacts with humans.
But the existing three laws are not enough. Today, we are entering an era of unprecedented human-AI collaboration that Asimov could hardly have envisioned. The rapid advancement of generative AI, particularly in language and image generation, has created challenges beyond Asimov’s original concerns about physical harm and obedience.
THE PROLIFERATION of AI-enabled deception is particularly concerning. According to the FBI’s most recent Internet Crime Report, cybercrime involving digital manipulation and social engineering results in annual losses counted in the billions. The European Union Agency for Cybersecurity’s ENISA Threat Landscape 2023 highlighted deepfakes—synthetic media that appear genuine—as an emerging threat to digital identity and trust.
Social-media misinformation is a huge problem today. I studied it during the pandemic extensively and can say that the proliferation of generative AI tools has made its detection increasingly difficult. AI-generated propaganda is often just as persuasive as or even more persuasive than traditional propaganda, and bad actors can very easily use AI to create convincing content. Deepfakes are on the rise everywhere. Botnets can use AI-generated text, speech, and video to create false perceptions of widespread support for any political issue. Bots are now capable of making phone calls while impersonating people, and AI scam calls imitating familiar voices are increasingly common. Any day now, we can expect a boom in video-call scams based on AI-rendered overlay avatars, allowing scammers to impersonate loved ones and target the most vulnerable populations.
Even more alarmingly, children and teenagers are forming emotional attachments to AI agents, and are sometimes unable to distinguish between interactions with real friends and bots online. Already, there have been suicides attributed to interactions with AI chatbots.
In his 2019 book Human Compatible (Viking), the eminent computer scientist Stuart Russell argues that AI systems’ ability to deceive humans represents a fundamental challenge to social trust. This concern is reflected in recent policy initiatives, most notably the European Union’s AI Act, which includes provisions requiring transparency in AI interactions and transparent disclosure of AI-generated content. In Asimov’s time, people couldn’t have imagined the countless ways in which artificial agents could use online communication tools and avatars to deceive humans.
Therefore, we must make an addition to Asimov’s laws.
FOURTH LAW: A robot or AI must not deceive a human being by impersonating a human being.
WE NEED CLEAR BOUNDARIES. While human-AI collaboration can be constructive, AI deception undermines trust and leads to wasted time, emotional distress, and misuse of resources. Artificial agents must identify themselves to ensure our interactions with them are transparent and productive. AI-generated content should be clearly marked unless it has been significantly edited and adapted by a human.
Implementation of this Fourth Law would require
mandatory AI disclosure in direct interactions,
clear labeling of AI-generated content,
technical standards for AI identification,
legal frameworks for enforcement, and
educational initiatives to improve AI literacy.
Of course, all this is easier said than done. Enormous research efforts are already underway to find reliable ways to watermark or detect AI-generated text, audio, images, and videos. But creating the transparency I’m calling for is far from a solved problem.
The future of human-AI collaboration depends on maintaining clear distinctions between human and artificial agents. As noted in the IEEE report Ethically Aligned Design, transparency in AI systems is fundamental to building public trust and ensuring the responsible development of artificial intelligence.
Asimov’s complex stories showed that even robots that tried to follow the rules often discovered there were unintended consequences to their actions. Still, having AI systems that are at least trying to follow Asimov’s ethical guidelines would be a very good start.
What Can AI Researchers Learn from Alien Hunters?
THE SETI INSTITUTE’S APPROACH HAS LESSONS FOR RESEARCH ON ARTIFICIAL GENERAL INTELLIGENCE
Edmon Begoli & Amir Sadovnik
THE EMERGENCE OF artificial general intelligence (systems that can perform any intellectual task a human can) could be the most important event in human history. Yet AGI remains an elusive and controversial concept. We lack a clear definition of what it is, we don’t know how to detect it, and we don’t know how to interact with it if it finally emerges.
What we do know is that today’s approaches to studying AGI are not nearly rigorous enough. Companies like OpenAI are actively striving to create AGI, but they include research on AGI’s social dimensions and safety issues only as their corporate leaders see fit. And academic institutions don’t have the resources for significant efforts.
We need a structured scientific approach to prepare for AGI. A useful model comes from an unexpected field: the search for extraterrestrial intelligence, or SETI. We believe that the SETI Institute’s work provides a rigorous framework for detecting and interpreting signs of intelligent life.
The idea behind SETI goes back to the beginning of the space age. In their 1959 Nature paper, the physicists Giuseppe Cocconi and Philip Morrison suggested ways to search for interstellar communication. Given the uncertainty of extraterrestrial civilizations’ existence and sophistication, they theorized about how we should best “listen” for messages from alien societies.
We argue for a similar approach to studying AGI, in all its uncertainties. The last few years have shown a vast leap in AI capabilities. The large language models (LLMs) that power chatbots like ChatGPT and enable them to converse convincingly with humans have renewed the discussion of AGI. One notable 2023 preprint even argued that ChatGPT shows “sparks” of AGI, and today’s most cutting-edge language models are capable of sophisticated reasoning and outperform humans in many evaluations.
While these claims are intriguing, there are reasons to be skeptical. In fact, a large group of scientists have argued that the current set of tools won’t bring us any closer to true AGI. But given the risks associated with AGI, if there is even a small likelihood of it occurring, we must make a serious effort to develop a standard definition of AGI, establish a SETI-like approach to detecting it, and devise ways to safely interact with it if it emerges.
THE CRUCIAL FIRST step is to define what exactly to look for. In SETI’s case, researchers decided to look for certain narrowband signals that would be distinct from other radio signals present in the cosmic background. These signals are considered intentional and only produced by intelligent life. None have been found so far.
In the case of AGI, matters are far more complicated. Today, there is no clear definition of artificial general intelligence. The term is hard to define because it contains other imprecise and controversial terms. Although intelligence has been defined by the Oxford English Dictionary as “the ability to acquire and apply knowledge and skills,” there is still much debate on which skills are involved and how they can be measured. The term general is also ambiguous. Does an AGI need to be able to do absolutely everything a human can do?
One of the first missions of a “SETI for AGI” project must be to clearly define the terms general and intelligence so the research community can speak about them concretely and consistently. These definitions need to be grounded in disciplines such as computer science, measurement science, neuroscience, psychology, mathematics, engineering, and philosophy.
There’s also the crucial question of whether a true AGI must include consciousness and self-awareness. These terms also have multiple definitions, and the relationships between them and intelligence must be clarified. Although it’s generally thought that consciousness isn’t necessary for intelligence, it’s often intertwined with discussions of AGI because creating a self-aware machine would have many philosophical, societal, and legal implications.
NEXT COMES the task of measurement. In the case of SETI, if a candidate narrowband signal is detected, an expert group will verify that it is indeed from an extraterrestrial source. They’ll use established criteria—for example, looking at the signal type and checking for repetition—and conduct assessments at multiple facilities for additional validation.
How to best measure computer intelligence has been a long-standing question in the field. In a famous 1950 paper, Alan Turing proposed the “imitation game,” more widely known as the Turing Test, which assesses whether human interlocutors can distinguish if they are chatting with a human or a machine. Although the Turing Test was useful in the past, the rise of LLMs has made clear that it isn’t a complete enough test to measure intelligence. As Turing himself noted, the relationship between imitating language and thinking is still an open question.
Future appraisals must be directed at different dimensions of intelligence. Although measures of human intelligence are controversial, IQ tests can provide an initial baseline to assess one dimension. In addition, cognitive tests on topics such as creative problem-solving, rapid learning and adaptation, reasoning, and goal-directed behavior would be required to assess general intelligence.
But it’s important to remember that these cognitive tests were designed for humans and might contain assumptions that might not apply to computers, even those with AGI abilities. For example, depending on how it’s trained, a machine may score very high on an IQ test but remain unable to solve much simpler tasks. In addition, an AI may have new abilities that aren’t measurable by our traditional tests. There’s a clear need to design novel evaluations that can alert us when meaningful progress is made toward AGI.
IF WE DEVELOP AGI, we must be prepared to answer questions such as: Is the new form of intelligence a new form of life? What kinds of rights does it have? What are the potential safety concerns, and what is our approach to containing the AGI entity?
Here, too, SETI provides inspiration. SETI’s postdetection protocols emphasize validation, transparency, and international cooperation, with the goal of maximizing the credibility of the process, minimizing sensationalism, and bringing structure to such a profound event. Likewise, we need internationally recognized AGI protocols to bring transparency to the entire process, apply safety-related best practices, and begin the discussion of ethical, social, and philosophical concerns.
We readily acknowledge that the SETI analogy can go only so far. If AGI emerges, it will be a human-made phenomenon. We will likely gradually engineer AGI and see it slowly emerge, so detection might be a process that takes place over a period of years, if not decades. In contrast, the existence of extraterrestrial life is something that we have no control over, and contact could happen very suddenly.
The consequences of a true AGI are entirely unpredictable. To best prepare, we need a methodical approach to defining, detecting, and interacting with AGI, which could be the most important development in human history.
C. Cui's Blog 