能用它來談2026年的 AI 嗎：某 AI 大師譚 讓‧皮亞傑的一句話："Intelligence is not what you know, it's what you do when you don't know" 約翰·霍爾特在1964年出版的一本書中對其進行了重新表述並使其廣為人知：“智力的真正考驗不在於我們知道多少，而在於當我們不知道該做什麼時，我們的行
我想起了發展心理學先驅讓‧皮亞傑一句精彩卻又似曾相識的名言：
“L'intelligence, ce n'est pas ce que l'on sait, mais ce que l'on fait quand on ne sait pas.”
I was reminded of this wonderful, yet apocryphal, quote by developmental psychology pioneer Jean Piaget:
"L'intelligence, ce n'est pas ce que l'on sait, mais ce que l'on fait quand on ne sait pas."
"Intelligence is not what you know, it's what you do when you don't know"
Apparently this is not a real quote but a synthesis of Piaget's positions on the topic. Psychologist John Holt reformulated and popularized it in a 1964 book: "The true test of intelligence is not how much we know how to do, but how we behave when we don't know what to do."
These simple quotes clarify many questions surrounding AI today:
- Why the accumulation of declarative knowledge seems like Intelligence but is not intelligence.
- Why the accumulation of skills seems like intelligence but is not intelligence.
- Why the ability to solve new problems "zero shot" without prior training on said problem is an important metric of intelligence.
- Why intelligence is not a collection of learned skills but an ability to acquire new skills very quickly, with very little or no training.
Our mental model of reality gives us the ability to predict the consequences of our actions, which gives us the ability to plan, which gives us the ability to apprehend new situations without prior training
在這場以「智慧普及化將重塑每個人」為核心的演講中，黃仁勳回顧了從1950年代的 AI 先驅學者（如 Allen Newell 和 Herbert Simon）到1979年成立機器人研究所，強調 CMU 數十年來始終處於技術革命的最前線。 [1]
您可以透過 卡內基美隆大學官方 YouTube 頻道 觀看完整的畢業典禮致詞影片，深入了解他對 AI 時代未來人才與產業發展的見解。 [1, 2, 3]
AI OVERVIEW： 黃仁勳在卡內基美隆大學（CMU）2026屆畢業典禮演講中，強調CMU是人工智慧（AI）真正的發源地之一。他指出，該校研究人員在1950年代發明了史上第一個AI電腦程式「邏輯理論家」（Logic Theorist）。「邏輯理論家」是由CMU的艾倫·紐厄爾（Allen Newell）、司馬賀（Herbert A. Simon，後獲諾貝爾獎）與約翰·克里夫·肖（J. C. Shaw）於1955年與1956年間共同開發。這個程式被廣泛認為是歷史上第一個能自動進行符號推理與解決問題的AI程式。它成功證明了數學巨著《數學原理》（Principia Mathematica）中的多個定理，甚至找出了比原著更簡練的證明方法。這段歷史象徵了AI發展的開端，也呼應了黃仁勳在演講中鼓勵畢業生，在這個由AI重塑運算的新時代中「親手創造未來」的期許。
Herbert A. Simon先生紀念 (2026 0615；1 《紐約時報》，《明周文化 MP Weekly 明報周刊文化 ) 村上春樹的短篇小說〈四月某個晴朗的早晨遇見100%的女孩〉（收錄於同名短篇小說集《遇見100%的女孩》中）。：很簡單地說，人生的決策準則是「滿意」（他創英文新字），非最佳或極值…..。 我們每一人都有一套價值與哲學，Simon 選擇四十幾年每天步行來回學校，吃穿等到極簡單，多發表各領域的深入論文，課程創新
The Nobel-Winning Psychologist Who Believed He Found the Secret to Happiness
May 12, 2026
A peach balancing on top of another peach, against a dark background.
Credit...Daniel Ribar
Listen · 7:03 min
Share full article528
By David Epstein
Mr. Epstein is the author of “Inside the Box: How Constraints Make Us Better.”
If in making decisions you are often guided by a search for the best, you are going about decision making all wrong — and you’re also probably less happy for it.
In an age of information and choice abundance, we assume we can find the best of everything if we look long and hard enough. Psychologists call that tendency maximizing.
But searching for the best is the wrong goal. That is because searching is itself a cost, and most people forget to account for it. If you did, you would see that the optimal strategy isn’t optimizing at all.
There’s a better way to make decisions. To understand it, you should know about Herbert Simon, a pioneer of artificial intelligence and cognitive psychology, as well as a Nobel laureate in economics. Mr. Simon demonstrated that for most decisions, humans can’t really evaluate the options available — there are too many, our information about them is incomplete and our minds aren’t built to weigh them all — and so we rely on mental shortcuts. He coined the term “satisficing” — a portmanteau of satisfy and suffice — to describe how we consider a limited set of options, then choose one that is good enough and move on to live our lives.
Sign up for the Opinion Today newsletter Get expert analysis of the news and a guide to the big ideas shaping the world every weekday morning. Get it sent to your inbox.
When Mr. Simon faced a decision, he considered a few alternatives, sometimes asked for advice, chose and moved on. He didn’t agonize, and he didn’t second-guess. “The best is enemy of the good” was the mantra he lived by.
Mr. Simon was, as he put it, an “incorrigible satisficer.” His eldest daughter, Katherine, recalled that he wore one brand of socks to avoid selecting color or style each morning, and he owned exactly one black beret at a time, made at a particular haberdashery in Europe.
According to Katherine, he said that one needed only three sets of clothes: “one on one’s body, one in the wash and one in the closet ready to wear.” He always ate the same breakfast — oatmeal, half a grapefruit, black coffee — and lived in the same house for 46 years.
“My father simplified his life in terms of his daily habits,” Katherine wrote, “thus eliminating the need to make little decisions about everything.” By taking the small decisions off his plate, that simplification freed his attention for the people and work that actually mattered to him. The mathematician John Allen Paulos illustrated the same principle with a thought experiment in his 1988 book “Innumeracy”: How should you choose your final romantic partner? First, he argued, you should estimate the number of people you might plausibly date in your lifetime. Then date roughly the first third with no intention of committing. Use that time purely to calibrate what you liked, what you didn’t like and what you might be missing.
After that, commit to the very next person you like better than everyone you’ve already dated. Mr. Paulos was illustrating a well-known result in probability, which shows that this rule gives you the best chance of ending up with the best partner in the whole sequence. Keep pushing past that point, and you’re more likely to end up with a worse match or no one at all. The core insight — that the path to the best outcome runs directly through the willingness to stop searching long before you’ve exhausted the options — extends far beyond dating.
Psychologists who followed up on Mr. Simon’s work have shown that his personal philosophy was both efficient and wise. Shortly after Mr. Simon’s death in 2001, a team of researchers created a maximization scale to measure where a person falls on the spectrum between maximizer and satisficer. They found that it’s usually bad to be a maximizer.
Maximizers tend to be less satisfied with their decisions and their lives. They are typically less happy, more prone to regret and more likely to compare themselves endlessly with others. Satisficers don’t necessarily have low standards. Their standard is “good enough for me” rather than “the best out there,” and that makes it possible to feel satisfied with their choices, instead of haunted by the ones they didn’t make.
The psychologist Mihaly Csikszentmihalyi, who first used the term “flow” to describe states of complete absorption in an activity, put it well. By making up one’s mind to invest in a choice, regardless of more attractive options that may come along later, “a great deal of energy gets freed up for living, instead of being spent on wondering about how to live.”
Advertisement This is critical today because chronic maximizing has never been easier. In 2006 an economist calculated that the consumer options available to citizens of modern economies exceeded those of preindustrial societies roughly by a factor of 100 million. That is an almost incomprehensible multiplication of choice, and it extends well beyond consumer goods into questions of who to be, how to live, where to work and whom to love.
Social media has intensified the problem by functioning as an infinite comparison engine. When you can see a curated highlight reel of everyone else’s career, relationship, home and vacation, the very concept of “good enough” begins to feel like settling.
The pull to keep searching for something better has poisoned even the most mundane moments. Research shows that giving viewers many videos to flip between makes them more bored than if they focus on just one. One way to interpret the findings is that the mere notion that something better might be out there spoils the moment.
Studies in the United States and China show that since about 2010, young people have reported becoming increasingly bored. Dating apps have offered a version of Mr. Paulos’s thought experiment, with users forever wondering what might be beyond that next swipe — maximizing in its purest form.
And now artificial intelligence promises to help us optimize everything: our schedules, our diets, our wardrobes, our creative output. If Mr. Simon was right, the hidden danger of these tools is that they will expand the menu of options and comparisons even further. The Japanese novelist Haruki Murakami captured the maximizer’s tragedy in a short story. A lonely boy and girl meet on a street corner and intuitively recognize that they are the perfect match for each other. It’s a miracle. They hold hands and talk for hours. But then a sliver of doubt creeps in: “Was it really all right for one’s dreams to come true so easily?” They decide on a test. If they truly are perfect for each other, they can part and will inevitably meet again. Then they’ll know for sure. The boy walks off to the west, and the girl to the east. They really were perfect for each other. Years later, they pass in the street, but their memories have faded. They never meet again.
Mr. Simon would not have been surprised they never met again. Whether you’re searching for a dishwasher or a date, set a good-enough standard. Stop when it’s met. Save your cognitive resources for things that matter.
David Epstein is the author of, most recently, “Inside the Box: How Constraints Make Us Better” and “Range: Why Generalists Triumph in a Specialized World.”
The Times is committed to publishing a diversity of letters to the editor. We’d like to hear what you think about this or any of our articles. Here are some tips. And here’s our email: letters@nytimes.com.
Follow the New York Times Opinion section on Facebook, Instagram, TikTok, Bluesky, WhatsApp and Threads.
++++++
Herbert A. Simon先生紀念5 從MIT公開下載的書： Scientific Discovery: Computational Explorations of the Creative ProcessUnavailable By Patrick W. Langley, Herbert A. Simon, Gary Bradshaw, Jan M. Zytkow 到 2026年5月18 " //今天 Google IO 活動同步發表了兩篇 Nature 論文，值得學研界朋友關注"
Scientific Discovery: Computational Explorations of the Creative ProcessUnavailable
Scientific discovery is often regarded as romantic and creative—and hence unanalyzable—whereas the everyday process of verifying discoveries is sober and more suited to analysis. Yet this fascinating exploration of how scientific work proceeds argues that however sudden the moment of discovery may seem, the discovery process can be described and modeled.
Using the methods and concepts of contemporary information-processing psychology (or cognitive science) the authors develop a series of artificial-intelligence programs that can simulate the human thought processes used to discover scientific laws. The programs—BACON, DALTON, GLAUBER, and STAHL—are all largely data-driven, that is, when presented with series of chemical or physical measurements they search for uniformities and linking elements, generating and checking hypotheses and creating new concepts as they go along.
Scientific Discovery examines the nature of scientific research and reviews the arguments for and against a normative theory of discovery; describes the evolution of the BACON programs, which discover quantitative empirical laws and invent new concepts; presents programs that discover laws in qualitative and quantitative data; and ties the results together, suggesting how a combined and extended program might find research problems, invent new instruments, and invent appropriate problem representations. Numerous prominent historical examples of discoveries from physics and chemistry are used as tests for the programs and anchor the discussion concretely in the history of science.
- Gottweis, J., Weng, WH., Daryin, A. et al. Accelerating scientific discovery with Co-Scientist. Nature (2026). https://doi.org/10.1038/s41586-026-10644-y
這篇發表於《自然》的論文介紹了 Google 團隊開發的 Co-Scientist。這是一個基於 Gemini 的多代理 AI 系統，由生成、反思、排名、演化、鄰近和元審查六個專業代理組成。不同於傳統工具，它能在非同步框架內透過錦標賽制的演化過程與自我對弈辯論，生成前所未有、可證實的新穎科學假說，且其假說品質隨測試時計算量的增加而持續提升，未見飽和。
- Aygün, E., Belyaeva, A., Comanici, G. et al. An AI system to help scientists write expert-level empirical software. Nature (2026). https://doi.org/10.1038/s41586-026-10658-6
這篇發表於《自然》的論文介紹了 Google 團隊開發的 ERA（實證研究助理）。這是一個將大型語言模型與樹狀搜索結合的代理 AI 系統，能自主生成、測試並迭代改進科學軟體，有效解決了過去需要專家花費數年勞動才能創建特定領域軟體的瓶頸。
ERA 的核心機制是透過大型語言模型改寫代碼以提升可量化指標，並利用樹狀搜索引導探索與回溯。該系統在六個科學基準測試中達到專家級水準：在單細胞 RNA 測序分析中，其生成的方法超越了公開排行榜上的既有做法；在 COVID-19 預測上，也擊敗了美國 CDC 的集成模型。研究更發現，ERA 有能力透過重組現有演算法來開創全新策略。這項突破展示了 AI 在多個高風險領域同時創建科學軟體的能力，也引發了未來如何部署與治理這類系統的深遠討論。//
