Thursday, October 09, 2025

為何英國高層政界痛恨香港BNO移民?因為香港搶了英國的飯碗:倫敦接棒香港成國際金融中心遺址

 

圖一:兩年前香港被內地網友稱「國際金融中心遺址」,今年IPO集資規模全球排名僅次於首位的美國,反觀倫敦市場全球排名跌至廿三位,比墨西哥、非州阿曼更低。(圖片來源:經濟一週,香港經濟日報)

據報英國財政大臣李韻晴(Rachel Reeves),九月初與投行巨頭高盛(Goldman Sachs)的國際聯席行政總裁古特曼(Anthony Gutman)聯手主持閉門會議,旨在向各大企業尤其是科企高管招手,說服他們在倫敦上市。記者引述其他投行的觀點,認為單一投資銀行直接參與由政府主導的會議,實屬罕見,尤如免費給予高盛吸客機會,暗示做法不公。

英國之所以這樣做,是因為倫敦作為傳統國際金融中心的地位正受到嚴峻考驗。數據顯示今年八月,倫敦市場的首次公開招股(IPO)融資額創三十年來最低,全球排名更跌至廿三位,比墨西哥、非州阿曼更低。反觀兩年前被不少內地網友戲謔是「國際金融中心遺址」的香港,今年IPO集資規模已達一千五百億港元,全球排名僅次於首位的美國。會計師事務所德勤四月時估計,二〇二五年香港新股全年集資總額介乎一千三百億至一千五百億港元,近期再度上調預測,由六月中旬預的兩千億,大幅上調到兩千五百億至兩千八百億,接近二〇一八年水平。

前有線新聞助理總監兼有線財經台台長顏寶剛認為,倫敦交易所的上市地位急速下滑,一是因為脫歐後對跨國企業的吸引力大減,二是倫敦交易所對上市公司高層領高薪極為抵制,三是工黨政府聚焦財赤,忽略刺激經濟與金融市場的措施。

三十年河東,三十年河西,不過兩年光景,倫敦已取代香港成為國際金融中心遺址。回想起來,英國當年對中共友善,願意交回香港主權,是為了貪圖海外人民幣的金融市場,好似二戰之後可以處理海外美元金融一樣,到時中美兩大國的海外金融都由英國代理,英國繁榮無比,甚至大手放棄工業,變賣勞斯萊斯公司給德國。可惜中共的人民幣國際化計劃中途停止,限於在香港做國際金融,故此香港在主權移交中共之後,無意之中搶奪了前宗主國英國的飯碗。英國賠了夫人有折兵,經濟衰到今日;至於一直當中俄及一帶一路那些古怪國家的白手套的香港,經濟繁榮至今,實在諷刺。為何英國的高層政界痛恨香港人?時不時就要拿BNO 5+1的移民來出氣?也許就是因為香港搶了英國的飯碗吧。


Source: 陳雲

https://www.patreon.com/posts/wei-he-ying-guo-140801132

Sunday, October 05, 2025

半身入土紫霞仙,國慶華服費金錢

 

圖一、畫面好靚,但沒有意義的紫霞仙子。圖片來源:橙新聞


美學小評。不講藝術,只談時代。在中國大陸仍是封閉鎖國的時候,香港是英國殖民地,但更加是國際看中國的窗口,當時擺這種中秋裝飾,肯定招徠無數歐美遊客的目光,來香港看中國。

現在中國改革開放幾十年,看這些,連帶什麼華服漢服,去大陸豈不更好?

以時代看來,與其說是慕華,不如說是戀殖。

時哉時哉!儒家的學問,最難的就是個時字。能得其時,謂之時中。這些話,今日在華夏大地,幾個能懂?

至於術數吉凶,貧道也透露半句吧。天可憐見,美人半身入土了......[1]


日前在面書寫了美學短評,只談藝術的時宜——是否切合當時的需要,不談藝術本身。國慶用華服和華服的亮燈裝置,表面上看,是合乎時宜,然而從中共的方面看,這種跳躍——或者回歸傳統,需要一個解釋。為什麼繼承辛亥民國革命的人民共和國、為什麼推行簡體字和拼音中文教學的中共要回歸漢唐衣冠呢?當初的中華蘇維埃、社會主義和國際主義的、共產黨版本的普世主義何處放呢?


更無一個是男兒?

展出的華服公仔,都是女人,沒有男人,那麼孫中山的驅逐韃虜,還我中華的反清復明有何用呢?因為滿清是容許女子着明朝衣冠的,剃髮易服,有所謂十不從,例如男從女不從,老從少不從,生從死不從。女人、少年人和死人不必更換滿洲衣服,僧人、畫像的古人、奴僕之類也不必。

五代十國時期,後蜀國主孟昶後來沈迷於酒色,以致國事日非,後蜀廣政三十年(西元九六五年)孟昶投降趙宋,花蕊夫人在亡國之後,寫下《述國亡詩》:「君王城上豎降旗,妾在深宮哪得知,十四萬人齊解甲,更無一個是男兒。」此詩甚受宋太祖趙匡胤賞識。

中共意圖恢復漢服和儒學,以新時期的民族主義統治和籠絡華僑,我早得知,然而這種轉變,雖然有利,也該有個說法,不能先做了再算。否則清末之後的幾次革命,有何用哉?

藝術而言,華服裝置的主角——紫霞仙子(見圖一)是主色是慘藍和淡紫,紫色不是正色,淡紫更加不是,都是收斂的顏色,缺乏國慶或中秋需要的艷麗和光芒,仙子的表情也是收斂的,所謂含羞答答、含情脈脈之類,不夠大方,而且周身的配襯都是自然界的花,並無人工的金銀珠寶,近乎水妖、山妖,並不是位列仙班、居於天界的仙子。當然,這是取材自當今大陸的動漫造型和設色,但在國慶或中秋採用,當略作改變的,否則就不合時宜——不切合當時的喜慶場合需要。


圖二、滿服女子,皇陵出土復原圖。圖片來源:橙新聞

顏色慘淡,寓意前景慘淡?

其他的滿服女子造型正常,設色也略為吉利,但面目死板,有樓蘭女屍復原圖、清皇陵格格女屍復原圖的觀感。嫦娥仙女的造型正常,但紅蓮的顏色也偏向慘淡,難免令人覺得前景慘淡——這倒是切合時宜的。


圖三、瘦面嫦娥,月上無啖好食?圖片來源:橙新聞

整個藝術裝置,就是媚俗和堆砌,德文的藝術評論說的Kitsch,專事於模仿、複製現存風格,缺乏原創性與藝術反省,為了迎合商業需求或政治需要而製作,過目即忘,俗不可耐。

要略為改造是可以的,就是放棄嫦娥與滿服少女,只是做紫霞仙子的各種造型,講出一個故事來。這樣就有主題,好可惜,講故事(story-telling)是要有意義的,而講出意義是目前香港政治、甚至中國最為顧忌的事情。還是堆砌幾個無法關聯的零碎片段安全好多——紫霞仙子與滿服少女與嫦娥,關聯在哪裡呢?沒有關聯就沒有意義,最安全。


附錄:新聞簡報

臨近國慶及中秋節,紫荊文化集團、香港中旅集團、民建聯與香港文化藝術界慶祝中華人民共和國國慶籌委會,周一起(9月29日)一連四日在西九文化區藝術公園海濱東草坪舉辦「迎國慶.華服嘉年華」,展出「香江仙子」巨型人物花燈組,其中主燈是高12.8米半身的「紫霞仙子」,為香港展出最大單體人物花燈,另設多個華服體驗攤位、華服國潮時裝秀及表演,打造「白天遊園加夜間賞燈」全時段體驗模式予市民及遊客參與。

嘉年華最為矚目的「香江仙子」花燈,燈組設在沿維港岸線旁,由四個巨型人物配荷花造型的花燈組成,運用國家級非遺自貢花燈技藝,其中主燈「香江仙子」高12.8米,為香港首次展出最大單體人物花燈,晚上時段將會亮燈,部份人物花燈更會做出眨眼、扭動身體等動作。

場內亦設多個體驗區域讓市民參與,包括「華服遊園會」的霓虹街拍擂台提供摺扇、宮燈等道具打卡;「華裳制衣坊」體驗製作手絹及香囊等手工藝品;「非遺體驗區」展示蘇繡、福聯、漆扇等非物質文化遺產;以及「國風市集」設中港兩地機構,展示傳統文化特色的產品。


Source: 陳雲

https://www.patreon.com/posts/ban-shen-ru-tu-140493570


Tuesday, September 30, 2025

政府製造忌諱,目的是令人民貧窮

 

圖、Woke culture就是令普通人得到裝備去批鬥窮人。


我之前拍片講過風暴期間有勇士在海濱端坐,無懼風浪。然而,政府也以疏忽照顧兒童的罪名拘捕了一些家長。事後政府也說考慮立法對付,當然我認為這只是應付民意,並不真的會做。

在台灣,所謂博愛座也經常引起爭吵,甚至打鬥,例如短片的情況。


《道德經·第五十七章》講過一句,「天下多忌諱,而民彌貧」,以前不大了解,這幾年容易明白了,就是政府或政府背後的財團勢力故意製造社會禁忌,目的是令人做事顧慮太多,被犯規的焦籠罩,恐懼周邊的人群的注視目光和公審壓力,於是膽小畏縮,無法改變自己的現狀,大部分都是安於貧窮,甘願被剝削,遇到大財團的交易合約和app的下載合約,明明是不公平也是啞忍,希望大財團針對的不是自己,只是在保護禁忌的時候才顯得勇猛起來,於是窮人舉報窮人,樂此不疲。普通人很容易動員起來去罵窮人違反女權,但富豪頻頻更換女模特兒,在媒體面前自誇,窮人卻從不想到去譴責。

舉凡Woke culture、DEI或香港與台灣那種關懷型政府(caring government),都在製造社會禁忌,在經濟衰退的時候更加熱衷於立法強制關懷,例如立法禁止獨自爬高山峭壁、風暴期間觀浪等等。


附錄:

原文:

以正治國,以奇用兵,以無事取天下。吾何以知其然哉?以此:天下多忌諱,而民彌貧;人多利器,國家滋昏;人多伎巧,奇物滋起;法令滋彰,盜賊多有。故聖人云:「我無為,而民自化;我好靜,而民自正;我無事,而民自富;我無欲,而民自樸。」

語譯:

以無為、清靜之道去治理國家,以奇巧、詭秘的辦法去用兵,以不擾害人民而治理天下。我怎麼知道是這種情形呢?根據就在於此:天下的禁忌越多,而老百姓就越陷於貧窮;人民的銳利武器越多,國家就越陷於混亂;人們的技巧越多,邪風怪事就越鬧得厲害;法令越是森嚴,盜賊就越是不斷地增加。故此有道的聖人說:「我無為,人民就自我化育;我好靜,人民就自然富足;我無欲,而人民就自然淳樸。」


Source: 陳雲

https://www.patreon.com/posts/zheng-fu-zhi-zao-140082527

Sunday, September 21, 2025

什麼是無為而治?最深刻的解答是……

 

圖、傳說孔子曾經問禮於老子,見於《莊子》及《孔子世家》。圖片來源:網絡


讀者問及,何為無為而治?這個是道家修煉自己和治理事情的原則,也同時是儒家歌頌上古聖王的統治方法。

先解說簡單的儒家,儒家的無為而治是歌頌舜帝的先王政治,見《論語·衛靈公》篇:  

子曰:「無為而治者,其舜也與?夫何為哉,恭己正南面而已矣。」

語譯是:孔子說:「不發號施令就能治理好天下的人,看來只有舜吧?他做了些什麽呢?他只不過是嚴整自己之後,端正地坐在天子座位上而已。」


孔子隱去不講,子思講個明白


孔子是用設問、反問的語氣講的,他不提道、天,但說無為而治的,大概只有上古的舜帝了。合道、合天、合虛這些我們今日看起來依然含糊的概念,孔子也不講出來,他是不好意思講,不屑講出來。後面呢,他說舜帝可以給人家看到的作為,就是修己、正身。這樣坐在天子位上,就可以治天下。提到舜帝,孔子連修己治人的治人也不講。若是講治人就是對舜帝的不敬,因為舜帝根本沒這想法。到了孔子的孫子思,從春秋去到戰國,子思在亂世就在《中庸》寫了修道而令天地萬物可以化育的方法,因為時候無多,必須傳出方法,講個明白了。《中庸》開首第一段:「天命之謂性,率性之謂道,修道之謂教。道也者,不可須臾離也,可離非道也。……喜怒哀樂之未發,謂之中;發而皆中節,謂之和;中也者,天下之大本也;和也者,天下之達道也。致中和,天地位焉,萬物育焉。」這就解釋了聖王和君子如何領受天命之後修道,之後治理天下。這與《道德經》的講法甚為相似,故此我不在這裡講《中庸》(復興神壇有講座講授《中庸》),跳去講《道德經》(復興神壇也有講座講解)。

道家的老子,其實比起孔子更為多言的,更講得清楚,這是我們學道家的時候必須知道的:道家比儒家容易學啊!


道家的無為,是什麼回事?


老子並無提到無為而治這個詞,但他很多地方都講了這個概念。引《道德經》兩段給大家看,之後我再稍為講一下。

第一段引文是《道德經·三十七章》:

「道常無為而無不為。侯王若能守之,萬物將自化。化而欲作,吾將鎮之以無名之朴。鎮之以無名之朴(樸),夫亦將不欲。不欲以靜,天下將自正。」

語譯是:道恆常是順任自然而無所作為的,卻又沒有什麼事情不是它所作為的。侯王如果能按照道的原則為政治民,萬事萬物就會自我化育、自生自滅而得以充分發展。自生自長而產生貪欲時,我就要用道來鎮住它。用無名的、道的樸來鎮服它,就不會產生貪欲之心了,萬事萬物沒有貪欲之心了,天下便自然而然達到穩定、安寧。

這句引文的重點是:無為不是人的事情,是道的事情!聖王、君子、侯王還歸於道、合乎道,天人合一了,就可以行使道的方法,令萬物自己演化。到了萬物有了感情和慾望,不合乎道,聖人又要用樸來鎮住他們,令他們自己矯正過來。萬物有了感情和慾望,偏離了自然而化,變成為自己是感情和私慾而化,聖人就不能用道,而是等而下之,用樸了!樸之後呢,是沒法了。天下大亂了!

怎樣去到無為呢?我們看第二段引文。《道德經·四十八章》說:「為學日益,為道日損,損之又損,以至於無為,無為而無不為。取天下常以無事,及其有事,不足以取天下。」

語譯:學習是日益增長的,而修道則是日漸減少的。減少再減少,最終達到無為的境地。能夠無為,能達成一切,什麼都做到了。欲取天下,應該保持無事發生的狀態,當有事發生時,便不足以取天下。


是無為,不是不為


《道德經》的無為,先從語言來講。「無為」,不是毋為、不為、勿為、莫為、弗為,後者都是否定的,而「無為」本身是一件東西,不是「為」或「有為」的否定。這點明白了,就可以通道家的學問。「無」是什麼?「無」不是做了事情、也不是不做事情,而是處於與道合一的狀態,模模糊糊的入定的狀態,之後做出的事情就是無不為,做了你也以為我沒做,這就是「無為而無不為」。


為何「有生於無」?「無」是什麼?


雖然我在復興神壇的道德經講座講過,無是什麼。不過,這裡也透露一點吧,等大家看了文章,依然有收穫。《道德經·四十章》:「反者道之動,弱者道之用。天下萬物生於有,有生於無。」講出「有生於無」的道理。

要讀懂這句,千萬要注意章句!道祖老子說「天下萬物生於有,有生於無」,而不是「有生天下萬物,無生有」!道祖講的「······生於······」,是描述一種已經發生的狀態,而不是預告一件沒有發生而等待發生的事情。不懂得這樣讀,永世也不必學道,不必旨意明白《道德經》!「······生於······」是沒有時序概念的,是有與無、無與有,恒常在互換、對流,好像雌雄頻繁交媾的動態之中,由於來回往復,好像前後的動作疊影了,疊加一起了,故此你也可以說是靜態,沒所謂的。修道去到極致,入了道了,我與道、天與人合一的時候,就可以見到這種疊影,而普通有一點點修為的人就只能看到先後的次序,或看到不動的靜態。疊影在物理上是存在的,但要用言語來講,就必須有先後,道祖的做法,唯有用中文的句法來講,用「生於」來講:「天下萬物生於有,有生於無」。

雖然講了兩個面向(aspect),其實是一個面向,還有其他面向的,那麼就要留待復興神壇的《道德經》講座再談了。


附錄:與讀者的問答。

SLK:晚生有惑,對無為而治的想像是這樣的:「 經過綜合分析後,調整環境內的某些因素以達致環境的循環和平衡,得一勞永逸的效果。」

晚生才疏學淺,對無為而治的理解,得到類似 「 自動化 」 的結論。晚生的疑惑,正是 「 無為而治 」 與 「 自動化 」 的區別在哪裡?如果有人說 「 一個環境的管治成功,在於能不能使它自動化 」。撇開用詞不當,是不是就是「 無為而治 」 的意思?

晚生對自動化有 「 侵蝕就業環境 」 、 「 過度發展 」 和 「 資源過剩 」 等偏見。老師若能為學生解惑,定必受益匪淺。

陳雲:自動化在工業是將既定的生產方法自動化,依然依照之前的環境,是限制了可變因素的。在地鐵行駛方面,可以自動化,因為軌道封閉,飛機飛行可以在某些情況自動化,找個飛機師看住,如爬升到上空之後。

政治的環境可變因素太多,而且要顧及百姓感受,故此難以自動化。

無為而治是道家的觀念,是順天應人,是華夏的神學來的,並非是普通研究道家的人說的,順應環境而選擇什麼都不做,或不去胡亂做。我有機緣的時候再講啦。

SLK: 感謝老師!晚生自知兩者有別,然仍想嘗試融會貫通,把無為而治應用到管治或政治以外的範疇。目下還是不要相提並論,期待老師的講解。


Source: 陳雲

https://www.patreon.com/posts/shi-mo-shi-wu-er-139389876

Thursday, September 18, 2025

拜登佈局烏克蘭,陷害德、法不翻身!美國前高官揭秘:烏克蘭如何受到NED和索羅斯的擺佈而觸怒了俄羅斯,澤倫斯基身不由己!

 

圖一、被美國捧上台的澤倫斯基,俄烏戰爭之後叛變美國。圖為澤倫斯基在二〇二五年的新年講話剪影。圖片來源。


俄烏戰爭的關鍵原因,是澤倫斯基執意要加入北約,令俄羅斯失去烏克蘭這個中立區的屏障,北約有機會直接碰觸俄羅斯邊界。然而,為何澤倫斯基會「找死」,令烏克蘭陷入戰火呢?一旦知道真正的原因——原來出自美國情報部門的威迫,會令香港人大吃一驚!國際局面果然是弱肉強食,幸好澤倫斯基沒有聽從拜登老爺的指示,在俄軍攻入烏克蘭之後流亡去波蘭,而是負隅頑抗,守住烏克蘭西部的防線,苦戰到今日,恐怕俄羅斯自己反而突進到北約的波蘭邊界,直接威脅到北約國家而面臨更大的反撲,令德國和法國遭受更嚴峻的戰爭打擊。

在講出答案之前,要提到前日在Charlie Kirk的追悼會上,美國副總統萬斯正式向索羅斯的開放社會基金會和福特基金會開炮,說它們享受美國稅務優惠,卻散播左翼思想,破壞美國二百年來的基業。萬斯如此放言抨擊NGO,與美國前助理國務卿Mike Benz九月十日在youtube的訪問揭露,互相輝映!

令人意想不到,美國國務院前高官揭露,迫使烏克蘭陷入戰火的,正是美國!行凶的是USAID(前美國對外發展署)屬下的NED(美國民主基金會),而背後操縱的,不難估到,是CIA(中央情報局)。

這要從顏色革命講起。從東歐到中亞,多數所謂的「顏色革命」背後,都有NED 的身影。它們提供資金、訓練、媒體管道,成為政權更迭的重要推手。NED在一九九〇年代全面進駐烏克蘭,並深度介入基輔的橘色革命與廣場起義。NED的錢是不夠的,另外還靠索羅斯,他是民主黨最大金主,他的基金會(開放社會基金會)往往與 CIA、NED 的行動同步。他能提前知道某國即將政局不穩,先布局貨幣市場,藉「內線」獲利。根據Benz 的說法,開放社會基金會並非純 NGO,而更像索羅斯投資基金的行動部門。基金會推動的「改革」,往往與索羅斯的投資組合利益高度一致。一次政變成功,索羅斯進帳萬千,之前他的捐款可謂一本萬利。

澤倫斯基是怎樣被美國操縱的呢?烏克蘭媒體高達 85–90% 由 USAID供養,特朗普回朝,砍掉USAID的援助,之後曝光,連最大型的《基輔獨立報》都是由USAID資助。透過對NGO的大筆資助,掌握這個國家公民社會的每一個面向,對烏克蘭形成類似殖民地治理的型態。司法,甚至包括教育,都與USAID深度綁定,美國最具影響力的教師工會主席與烏克蘭當地的教師工會合作,教導孩子們關於他們自身歷史的敘事、並徹底清除任何與俄羅斯的親近感,或烏克蘭、俄羅斯人民之間正面的互動痕跡。

藝術文化領域也是,背後的目的也在徹底清除任何與俄羅斯有關的遺產痕跡,抹除烏克蘭曾經存在俄語,或曾經與俄羅斯共享歷史的事實,將烏克蘭「歐洲化」,亦將烏克蘭的天然氣與稀土資源納入控制。

澤倫斯基當選,是靠民望高,民望高是媒體營造出來的。二〇一九年,澤倫斯基以七成選票當選,政見是與俄羅斯和平相處(頓巴斯和平、遵守明斯克協議、允許烏東居民使用俄語)。但上任一個月,「烏克蘭危機媒體中心」便出現了。這個組織的資金來源包括:美國國務院(透過美國駐基輔大使館)、USAID、北約,以及其他一些長期與 CIA 有關的外圍組織,例如 Chemonics。Chemonics 是 USAID 裡規模排名前五大的補助承包商之一,該組織集合了七十個 NGO, 聯名對澤連斯基發出著名的「紅線備忘錄(Red Lines Memo)」,警告他不得就與俄羅斯和平解決的事情舉行公投、不得在西方夥伴缺席下單獨與俄羅斯談判、不得對俄羅斯採綏靖政策做出妥協滿足其要求、不得拖延破壞或拒絕加入北約與歐盟的方向⋯⋯等二十條。違反的話,澤倫斯基將面臨「動盪(instability)」。意思是說,美國如何扶助澤倫斯基上台的,就如何推他下台。

如此,澤倫斯基被逼申請加入北約,令俄羅斯進攻烏克蘭。然而,當拜登授意澤倫斯基逃亡波蘭的時候,澤倫斯基留在烏克蘭死戰,這可以說是澤倫斯基的叛變,挽回烏克蘭的命運,也間接短暫救回了歐洲。這是英雄式的一局,至於他如何貪污腐化,如何鎮壓反對者,是另一回事。俄國入侵烏克蘭一個月,遇到抵抗,本來想和談撤軍,奈何拜登不容許澤倫斯基與普京在土耳其談判,這是德國前總理施羅德揭露的,我在本欄寫過。

俄烏戰爭是拜登布的局,拜登灑下的網,現在由特朗普破局,特朗普收網。唱黑白臉的,都是美國。至於萬斯在Charlie Kirk的追悼會上破口大罵索羅斯,是深層國家要換過一套玩法了。


Source: 陳雲

https://www.patreon.com/posts/bai-deng-bu-ju-e-139186787

Wednesday, September 10, 2025

When Deepware meets Wetware: the uncomfortable truth about responsible AI

There is a memo from IBM dated 1979 that should be mandatory reading for every CEO racing to deploy AI. It states simply and profoundly: machines cannot be responsible.

IBM note, 1979


Fifty three years later, we have built what I call “deepware” — the layered neural architectures of AI and machine learning that sit atop our traditional software and hardware stack. Yet somehow in our breathless rush to innovation we have convinced ourselves that this deepware can shoulder the burden of responsibility that our wetware — our human brains — seems increasingly eager to abandon.

Let me be clear: responsible AI does not start with better algorithms or more data. It starts with humans who remember how to think.


The four layers of our digital reality

We have always understood technology through layers. Hardware provides the physical substrate. Software gives us the instructions and logic. But now we have added deepware — these probabilistic, pattern matching systems that generate increasingly convincing simulacra of intelligence.

The critical layer though remains wetware: the human brain, with its capacity for judgement, ethics and crucially responsibility. Yet this is precisely the layer we are systematically deactivating in our AI deployments.

Consider the lawyer who recently submitted AI generated legal briefs to court, complete with fabricated cases and fictional precedents (Morgan 2025). Or the Melbourne trial delayed because no one verified the AI outputs (Editorji 2025). These are not stories about technological failure. They are stories about wetware choosing to go offline at the exact moment it is most needed.


The 7% that reveals everything

MIT research recently exposed something chilling: when identical medical symptoms are presented to AI healthcare systems, female patients are 7% more likely to be told to “manage at home” (MIT 2025). Add a typo, use uncertain language like “maybe”, format your message imperfectly — and the deepware decides you deserve less care.

This is not a bug in the deepware. It is a feature of how we have trained it. Our wetware, in its rush to deploy, forgot to ask critical questions: Whose patterns are we matching? Whose biases are we encoding? Whose voices are we amplifying or silencing?

The deepware is doing exactly what we programmed it to do: replicating patterns from data. The failure lies in our wetware’s abdication of its responsibility to think critically about what patterns we choose to perpetuate.


The 11.8 billion pound illusion

The United Kingdom’s AI industry contributes £11.8 billion to the economy, growing 150 times faster than traditional sectors (UK AI Sector Study 2024). Yet here is the paradox that should keep every executive awake: 95% of generative AI investments are producing no measurable value (MIT Project NANDA 2025).

How is this possible? Because we have confused deploying deepware with thinking. We have mistaken automation for intelligence. We have treated AI as a substitute for human judgement rather than a tool that requires even more rigorous human oversight.

The organisations seeing real returns? They are the ones where wetware remains fully engaged — where humans apply critical thinking, systems thinking and what I call “consequence thinking” to every AI deployment. They understand that deepware without active wetware is just expensive pattern matching.


The shadow economy of human intelligence

While corporations pour billions into generative AI initiatives that fail, something fascinating is happening in the shadows: employees are independently using consumer AI tools to achieve real productivity gains (MIT Project NANDA 2025). The difference? Personal accountability.

When an individual uses ChatGPT to help write an email, they remain responsible for the output. They verify, they edit, they think. Their wetware stays online because they know their name is attached to the result.

But when organisations deploy AI at scale, something strange happens. Responsibility becomes diffused. “The AI recommended it” becomes the new “I was just following orders.” We have created systems where no human feels accountable for what the deepware produces.


Going slow to go fast

There is wisdom in the observation that sometimes we need to “go slow to go fast”. In our current AI gold rush, we are doing the opposite — going fast to go nowhere.

We are deploying first, thinking later if at all, and then acting surprised when our deepware amplifies every bias, mistake and oversight we failed to catch.

Real innovation in AI is not about being first to market. It is about being first to understand the systemic consequences of what we are building. It requires wetware that is fully engaged in both critical thinking and systems thinking, understanding not just what AI can do, but what it should do, and more importantly, what it absolutely should not do.


The great abdication

What we are witnessing is not technological revolution; it is responsibility abdication. CEOs chase innovation metrics while ignoring impact assessments.

Developers optimise for speed while overlooking safety.

Organisations celebrate deployment velocity while their wetware atrophies from disuse.


The simple truth is this: when professionals who should know better skip basic verification, when medical systems encode bias without oversight, when legal documents get submitted without review — we are not seeing AI failure. We are seeing human systems in collapse.


The responsibility stack

Here is what nobody wants to admit: as our technology stack grows more complex, our responsibility stack must grow more robust. Every layer of deepware we add requires an exponential increase in wetware engagement.

Think of it this way:

  • Hardware requires responsibility for physical safety and environmental impact
  • Software requires responsibility for functionality and security
  • Deepware requires responsibility for bias, fairness, and societal impact
  • Wetware must orchestrate responsibility across all layers

We cannot delegate ethics to algorithms. We cannot outsource judgement to models. We cannot transfer accountability to machines that, as IBM reminded us in 1979, cannot be responsible.


The path forward: reactivating human intelligence

The solution is not to slow AI development. The solution is to reactivate our wetware and practice responsible AI by following regulations like the AI Act that recommend that human oversight is not optional; it is essential.

This means:

  • Verification before deployment: every AI output requires human review
  • Critical thinking over speed: better to be right than first
  • Systems thinking over isolated innovation: understanding ripple effects before creating them
  • Accountability at every level: from the developer to the CEO, everyone owns the outcomes


The choice that defines our future

We stand at a crossroads. Down one path lies the continued abdication of human responsibility, where we blame algorithms for our biases, hide behind AI for our decisions, and gradually forget how to think critically about the world we are creating.

Down the other path lies something more challenging but infinitely more valuable: a future where deepware amplifies human intelligence rather than replacing it, where AI serves as a tool for enhanced thinking rather than an excuse for not thinking at all.

The uncomfortable truth is that responsible AI has nothing to do with making machines more responsible. It has everything to do with humans remembering that they are and must remain the responsible party in every equation.

Your wetware is the most sophisticated technology you possess.

In the age of deepware, it is also the most important.

The question is not whether AI will transform our world — it will.

The question is whether we will remain conscious, critical, and responsible enough to guide that transformation.

The next time someone tells you about their AI initiative, ask them not about the technology, but about the humans: Who is thinking? Who is verifying? Who is responsible?

Because in the end, no matter how deep our deepware becomes, it is the wetware that determines whether we are building a future worth living in.


Source: Patrizia Bertini

https://pat-bertini.medium.com/when-deepware-meets-wetware-the-uncomfortable-truth-about-responsible-ai-245842121749

Monday, September 08, 2025

習主席和普京都談延壽長生,但如何長生呢?《黃帝內經》的想法,上古醫書怎樣讀懂?

 

圖一、路透社無意中發現國家機密了。圖片來源:RNews


九月三日,中共在天安門廣場閱兵,紀念抗戰勝利。時候一段小插曲,變成熱談,那就是在天安門牆邊散步的時候,習近平向普京談延壽,說現在可以用器官重生激活等醫學技術,輕易延壽到一百五十歲。「過去人們活不過七十歲,現在七十歲還算孩子。」一名男子用俄語轉述習近平的話給普丁聽。當被塔斯社記者問及上述言論時,普京說實有其事。然而,路透社被中共電視台發律師信要求下架該談話片段,路透社由於失去轉載的授權,唯有照辦。然而,若干轉載的媒體顯然不理會呼籲,保持上載該片段。有網媒主持人探討為何該片段可以在三十秒的延遲之後依然直播出去,就是央視的審查員認為醫療隊伍為了領導而研究保健和延壽,從毛澤東以來就是規定做法,而國家主席向來賓誇耀中國的醫學成就,並無不妥。

人體結構最為複雜,單是人腦已經是宇宙最複雜的物體,即使可以激活細胞負責複製的端粒體、激活細胞負責新陳代謝的粒腺體、均衡飲食適當運動社交愉悅等等,加上器官移植或器官自體重生來輔助,也因為綜合維護肉體生存的機制出錯而引發衰老或死亡,故此將肉體壽命在本世紀限制在一百五十歲是合理的,也很難再有突破。


肉身有限,念力與神氣無限:前現代(pre-modern)的方法

現代醫學用的是身心分離的現代科學方法,即是十七世紀末之後的笛卡爾的身心二元論,認為物件的現象要與人的意念分開處理。這方法開啟了現代科學的因果關係(causality)的研究,因果關係必須要限制在可以客觀觀察、測定數量和用實驗證明或否定的範圍之內,離開這些範圍,科學就無法處理,也不能評論。要解釋到因果,也必須限制觀察對象為可以測定的個體,另外必須要有結構性質(structural)的解釋(這個來自古希臘的亞里士多德)。例如說,人體吸收陽光就可以健康和愉悅,不能只是籠統地說陽光就是陽氣,而人體需要陽氣,陽氣令人振奮,容光煥發,但不能過度,否則陽氣也會燒焦人,這是前現代的(pre-modern)思考方法,歐洲中世紀也是用這種想法。要去到陽光在皮膚之內起了什麼作用:促進血流、促進維生素丙和酵素活動,多方面增加人的健康,這才是結構性質的解釋。雖然我們現在知道陽光含有多種輻射線,會引起皮膚表面的熱力、靜電、電磁場和細胞功能的微細變化——即所謂中國古代說的氣的某些可以觀察的現象,但現代科學不處理這個,因為這些變化難以孤立出來量度,即使量度出來也沒有可以觀察的效果。

華夏古人看壽命與健康,是用另一個方法,這個身心重疊、身心交互影響、身心互相轉化的方法——我們說身心合一,就包含這三重意義!在歐洲古代和印度、西藏之類也通行身心合一的說法,只是用的名稱或假說不同。這裡只看華夏醫家及道家如何解釋衰老的原因。由於無法證明或證偽,故此只是假說,不是理論。


圖二、黃帝內經素問,唐朝王冰註解本。圖片是台北圖書館的元朝刊本。


人為什麼會衰老?先看《黃帝內經·素問第一卷·上古天真論篇第一》,黃帝與天師(天子之師)岐伯關於今世的人未老先衰的問答:

昔在黃帝,生而神靈,弱而能言,幼而徇齊,長而敦敏,成而登天。乃問於天師曰:「余聞上古之人,春秋皆度百歲,而動作不衰;今時之人,年半百而動作皆衰者,時世異耶?人將失之耶?」

岐伯對曰:「上古之人,其知道者,法於陰陽,和於術數,食飲有節,起居有常,不妄作勞,故能形與神俱,而盡終其天年,度百歲乃去。

今時之人不然也,以酒為漿,以妄為常,醉以入房,以欲竭其精,以耗散其真,不知持滿,不時御神,務快其心,逆於生樂,起居無節,故半百而衰也。

夫上古聖人之教下也,皆謂之:虛邪賊風,避之有時;恬淡虛無,真氣從之;精神內守,病安從來?

坊間有白話語譯,大家可以參考,但要解通這篇醫學古文,先要弄清楚一些詞的意思:動作、道、陰陽、術數、形神、精、真、神、真氣、精神,再看章句的語氣,才可以完整解通!讀古文要這樣讀,先識字,後解句,再明章。我相信現在讀中醫學的朋友,很多都沒讀懂《黃帝內經》的本事,因為註解的書都很差,沒有很好的訓詁。


中醫學生也讀不懂的《黃帝內經》開頭第一篇!

沒人執手而教,醫書或道書是很難讀通的。首先是顯示年青活力的「動作」一詞,所謂「動作不衰」是兩種事:動和作,因為後面有寫「動作皆衰」,故此動、作是分開解的,不是一個詞。動是指平日的舉動和特別的作為不會衰退,平日的舉動就是行走飲食之類,作是特別的作為是房事、耕作、工藝、祭禮之類。動作在這裡,類似英文的move和act。這樣說,大家會容易掌握到。(動是日常的活動,作是興起、振作,做某些特殊行動,例如合乎社交禮儀的行為。《左傳·襄公三十一年》:「法行可象,聲氣可樂,動作有文,言語有章。」)衰老就是身體不能動得自由,也不能做有意義、有職業反應和人倫反應的工作。黃帝是聽聞古代的人可以百歲而康樂自在,但現在的人為何五十歲就衰老的呢?古代的人可以度百歲而動作不衰,黃帝只是聽聞,不是實在的啊!這裡的百歲,依照唐朝王冰的註解,是一百二十歲的意思。《尚書.洪範》提出「一曰壽,百二十歲也」,《左傳》則有「上壽百二十,中壽百歲,下壽八十」之語。

歧伯的回答很妙!他說上古那些人,曉得道的,所謂明道之士,才可以度百歲而去,不是所有人都是這樣啊!曉的道的人,道生陰陽,故此要取法陰陽,用術數來調和,術數是日夜時辰與四季二十四節氣與人體的對應。食和飲有節制,飲特別是指飲酒。起居有定時,勞作不過度也不亂作,所有可見的形(身體/精氣)與不可見的神(元神/神氣)有交流地在一起。「形與神俱」的俱,是偕同、陪同的意思,即是英文的 in the good company of/ well accompanied的意思。有了這些條件呢,古代理解道的人,才可以好好地活完上天賜予的壽數(「盡終其天年」),度過一百歲而死掉。

歧伯講古代的明道之士,只是用了形、神兩個概念(「形與神俱」),是很模糊的,但到現在的人的時候,他換了詞彙,用了精、真、神三個詞。那就是形與神脫離之際,就出了精、神的概念!岐伯說,現在的人不同了啊,將酒當作日常飲料(漿是飲料的意思),將酒當作水來飲,亂來當作是常態,酒醉的時候行房交媾,用慾念來用乾了精,精是睾丸的精液、腎的腎氣和全身來自先天父母的元精(元精的概念有些似法國哲學家伯格森說(Henri Bergson)的Élan vital (法文發音: [elɑ̃ vital]) ,見他在一九〇七年出版的《創造的進化》(L'Évolution créatrice)。此假說的思想來源可以追溯至古希臘斯多葛派哲學家波希多尼提出的「生命力」假說,即太陽將生命力帶到地球表面的所有生物上。)三者,這三者在交媾的時候用乾淨了,就是耗散了真,真就是賴以生存的真氣,啟動人體的真氣,在沉睡之後將身體喚醒的真氣,在病倒或昏倒的時候令人振作而起的真氣。之後岐伯說這些人亂飲酒、亂作為、交媾的時候不留神而縱慾,就是不知道如何保住好的健康,不懂得按照四時節氣來駕馭神(陽氣、陽神),例如在暑熱的時候交媾,只是令縱慾的心快動起來。「逆於生樂」,一般的醫書註解都是將「逆」訓為違背的意思,即是違背養生之樂,在身體快樂的時候,元氣是耗散中的,尋歡的人違背這道理而去尋歡作樂。這樣解,頗為曲折,其實逆就是迎的意思(依《說文》),「逆於生樂」即是純為了迎接人生在世的樂趣,於是起居無節制,五十歲就衰弱了。

最後一句頗為詭異:夫上古聖人之教下也,皆謂之:虛邪賊風,避之有時;恬淡虛無,真氣從之;精神內守,病安從來?


《黃帝內經》的岐伯教學生,竟然留一手?

這是什麼意思呢?之前說「上古之人」,為什麼忽然變了「上古聖人」的呢?「聖人」這個詞出來了啊!「聖人」就是仁人愛物的有德之士,是有心教導人的那些。故此才會「教下」。「教下」教那些等而下之的人,不才的人,說不好聽的,就是教我們這些普通人,不是知識貴族。故此這種教法是遷就能力較差的人的。千萬不要以為就是好方法啊。故此這些聖人就籠統地講一些,隨便教教:這是「皆謂之」的意思,帶一點輕蔑、不屑。要避開那些不依照時節而來的、令人虛弱的邪氣(例如暑熱季節的涼氣、寒冷季節的熱氣)和從門戶縫隙穿透進來的急風(賊風),按照時節的情況來避免(如夏天不要貪着吹涼風、冬天不要貪着吹熱風),心裏不要妄想亂來,那麼真氣就跟隨不貪戀慾望的心,(精)神就守在形(身體)的裏面,病就無法來囉。岐伯這裡說的聖人是教老百姓防病的,不是教養生,更不是教延壽,《黃帝內經》也沒有修煉成仙的想法。

上古的明道之士,他們是「形與神俱」的,形與神合一,樂也融融的,熙熙和和的,不必特別去學的,不必修煉的。為什麼呢?因為「其知道者」的「知」字,是曉得,是相熟,也是主宰之意,有三重意思。「其知道者」就是那些曉得/熟悉/主宰「道」的人。故此,即使大家是聽我講道經和內丹,即使不去修煉,也有延年益壽之效。上古聖人教的下等一輩,只是氣跟隨心,神內守於神,而不是氣與神交融,不分你我,不分內外的。故此只能無病,不能長生延年!即是說,不是聖人教凡人是留一手,不是師父留一手不教徒弟,而是凡人的承載能力有限,只能教他們一點點。只這一點點,他們也未必做到啊。

例如我上面我展示的古文讀法,是我將各位當作平等人那樣來解釋的,和盤托出,我不是當大家是傻子,隨便說一下,糊弄過去就算。以前在大學跟隨我讀古書的,我都將他們當作平輩來論學,他們有些感到汗顏,重新學過中文,但也有些甚至動怒,成為我haters的。因為我令他們感到無知,無地自容。

至於如何養生延壽,修煉成仙?就看道家。下次談。


Source: 陳雲

https://www.patreon.com/posts/xi-zhu-xi-he-pu-138418115

Saturday, September 06, 2025

Incorporating environmental considerations into water procurement could deliver better outcomes

There is “an opportunity to improve procurement processes in such a way that could have saved tens of millions of dollars while keeping water quality irreproachably high – all while keeping due diligence costs low,” writes Basel Kimani.

----

One of the best bosses I ever had would, if something went wrong, calmly analyse what led to the error, and then, without pointing fingers, figure out how to prevent the mistake from happening again. Every failure was an opportunity to learn and improve.

This photo shows a warehouse in Yuen Long, where thousands of bottles of drinking water linked to a scandal-hit government contract are stored, on August 20, 2025. Photo: Kyle Lam/HKFP.


I am not inclined to call the recent errors in water procurement a “scandal.” Chief Executive John Lee’s description of the Government Logistics Department’s “failure to do its job” seems fairer.

Yes, avoidable mistakes were made, but to err is human, and I don’t see anything particularly scandalous about being human. 

It’s more constructive to see what can be learned to prevent similar issues from arising in the future. The Audit Commission is grinding through that process now.

And in the spirit of constructive feedback, I do see an opportunity to improve procurement processes in such a way that could have saved tens of millions of dollars while keeping water quality irreproachably high – all while keeping due diligence costs low.

Here’s the trick: incorporate environmental considerations into the procurement process.

We know that as of today, cost and value-for-money are two of the most important factors in selecting a new supplier. What if environmental considerations are also baked in right at the start?

Bottled water is trucked from water plants, so carbon emissions are embedded in transportation. The water bottles also generate large amounts of waste: the plastic carboys themselves are reused but not infinitely, and the plastic seals on the carboys contribute even more to the waste footprint.

If we could identify a source of water that avoids all that carbon and waste, while maintaining water quality and taste, surely that should count for something? And if it were cost-competitive, that should make it a top contender.

Let me introduce a high-quality, very affordable, great-tasting water supplier: Hong Kong’s tap water. Not only does it tick the boxes for cost and quality, but due diligence should also be easy.

The Water Supplies Department enjoys the imprimatur of the Hong Kong government, which, I would hope, holds some water during the government’s due diligence process.

I know many Hongkongers flinch at the thought of drinking water straight from the tap. Some of this is a legacy from the days when corroding pipes in older buildings would make the water appear an unappealing rusty colour.

But modern buildings use better pipes, so that really isn’t an issue today. Many Hongkongers still boil or filter their tap water before consuming it. Strictly speaking, this is unnecessary from a water quality perspective, but if an extra filtration step makes people feel more comfortable, sure, why not?

If civil servants in government buildings really blanch at the thought of drinking water straight from the tap, perhaps the procurement process could have considered installing and maintaining water filters instead of trucking in bottled water. 

I’m not saying all this to provide an advertorial for the Water Supplies Department or to besmirch the reputation of bottled water suppliers.

Rather, I’m suggesting that if, right at the start of the supplier selection process, the questions “How much CO₂ is emitted? How much waste will be generated?” had been asked, then perhaps the problem of how to hydrate thirsty civil servants might have been solved with a higher quality, cheaper, and – as a cherry on top – more environmentally friendly outcome.

Every failure is a learning opportunity. I hope that in this case, the procurement department, whether in government or in the private sector, can learn that incorporating environmental considerations into decision-making may even result in higher-quality procurement at lower cost.


Source: Basel Kimani

https://hongkongfp.com/2025/09/06/hong-kong-govt-should-incorporate-environmental-costs-into-drinking-water-procurement-process/

Tim's Take on New Education Guidelines

There was news last week of a new edition of the School Administration Guide, a regular publication from the Education Bureau. This handy document runs to no less than 313 pages, ensuring that few people will read it and fewer still will remember the contents.

Still, awareness is a good idea. These days things labelled “guidelines” have a way of being laws in effect and intention.

The guide mentions national security about 50 times. Clearly the Education Bureau is not at all discouraged by the suspicion that the most popular motive for emigration is to move kids out of range of its propaganda preoccupations.

Well, we live in interesting times. I cantered happily through the new rules on external events, speakers and reading exercises, which must not endanger national security. I take it, from a recent last-minute cancellation, that this includes not allowing your pupils to participate in debates if the judges are going to be retired democratic politicians.

This is a shame. Former democratic politicians have a lot of time on their hands these days. And I can say, as a very experienced debate judge, that one really does not consider the merits of the motion as an idea, because that is not chosen by the contestants. If the luck of the draw puts you in defence of Adolf Hitler you defend him as best you can and should be judged on the quality of your efforts, not on the judge's opinions of Adolf.

Anyway, onward and upward, as we used to say, until we get to the innovations in the primary area, where I found something a bit shocking.

Pupils will, apparently, be expected to acquire a basic knowledge of national defence, the national security law, and the Hong Kong People's Liberation Army (PLA) garrison.

Of course, there can be no objection to an introduction to the law, to national defence, or the PLA, at least from me. I chose military history as my undergraduate special subject, did a Master's degree in War Studies and then by way of penance spent three years in a Peace Research programme, which is just war studies from the other direction. National defence is a worthwhile subject of study, though perhaps an ambitious one for primary schools.

I object, though, to the classification of this kind of thing as “humanities.” This label is sometimes abused as a catch-all term just meaning Not Science and Not Social Science. This is an error.

Humanities properly applies to the study of philosophy, religion, history (if like most historians you think it is not a social science) and the arts: performing, visual, and dramatic. This is not the way the word is used by the Education Bureau, which offers humanities for primary pupils in six flavours: Health and Living, Environment and Living, Financial Education and Economics, Community and Citizenship, Our Country and I, and The World and I.

Some curiosities here. Who decided, I wonder, that every item should include the word “and”? One might also consider the merits of replacing My Country and I with My Country and Me. Just a suggestion.

More seriously, most of these items – at least health, environment and economics – are not humanities at all. What seems to have happened is that the old general studies subject was divided. One part became Science and the other part became Not Science. Calling it “humanities” is lazy.

It is also likely to lead to disappointment. One of the attractions of studying the humanities is supposed to be the encouragement of critical and creative thinking. But we're not really looking for that any more, are we?


Source: Tim Hamlett

Hong Kong Free Press Members Newsletter 

Friday, September 05, 2025

AI Is Disrupting How Young Brains Grow

Remember when the internet blew our minds?

I was 14 years old when we got it in our house, and I still vividly remember the urrr EEEE NNGG CRRrr keeee nnn ding ding sound of dial-up as I waited to chat on ICQ.

Life changed fast. No more flipping through Britannica at the library–we had the World Wide Web at our fingertips.

The internet changed our lives, but it also changed us as humans. The internet has changed how we think, focus, remember, and relate to others — we scan instead of reading and skim headlines instead of absorbing meaning. Our attention spans have shrunk, and our reliance on Google means we store less of what we learn.

For today’s generation, that same mind-blown moment is happening again — but with AI.

AI isn’t just writing for us — it is thinking for us. But how will this new phenomenon change us as humans? Especially our children, who are still learning how to think?

As a neuroscientist and mother of two, this honestly frightens me. Childhood is a time for getting messy with learning: for wrestling with words, piecing together meaning, and solving problems using multiple trial and errors. That is how strong, flexible minds are built.

But now we are robbing them of this mess— we are letting AI do the thinking for our kids. The consequences are showing up fast: studies reveal drops in school performance, and, most worrying of all, less active, less engaged brains.


Our brains are built, not born

Our brains aren’t naturally programmed to read and write; we invented these skills as humans. This means, when a child learns these skills, they have to borrow circuits originally designed for vision, language, memory, and movement and rewire them.

And this is anything but simple!

When we read, our visual brain region (the occipital lobe) scans the shapes of letters, then the language systems (located across the temporal lobe) kick in to decode those squiggles into sounds and whole words. At the same time, the frontal lobe (the thinking part of the brain) steps in to make sense of it all.

Writing is even more complex: We need to plan our ideas (frontal lobe again), find the right words (language areas), and then guide our hands to write (motor cortex). This is a huge team effort with all parts of the brain working together.

So when my daughter insists that “w” is just an upside-down “m” and won’t acknowledge it is a real letter, this is literally her brain rewiring itself to learn something new.

But what we are really building here is thinking — when we read, our brain predicts and fills in gaps, it compares what we read to what we already know, and builds mental models to understand concepts faster — as shown by decades of research by the late Walter Kintsch, PhD, who was a professor of psychology at the University of Colorado. Writing pushes us even further by having to organise thoughts, choose the right words, and hold ideas in mind as we get them down on paper.


Why this effort matters

When children rewire their brains for reading and writing, it is like reorganizing a kitchen into a home office — the fridge becomes a filing cabinet, the toaster holds pens, and the stovetop balances the laptop.

It is a lot of effort, but it is necessary fuel for growth — it forces new neurons to grow, rewires the brain, and improves the functioning of existing neurons. Stanislas Dehaene, PhD, a cognitive neuroscientist at Collège de France and author of the book Reading in the Brain, coined this process as “neural recycling.” It is when the brain takes areas that normally recognize objects and trains them to recognize letters and words. This rewiring is so significant that it is allocated its own name, the Visual Word Form Area.

Learning to read and write also trains the brain’s CEO — the frontal lobe, which is home to executive functions such as focus, working memory, planning, and self-monitoring. Research shows that key components of executive functions become strengthened when children are challenged to predict and reason while reading a story. I can see that when my daughter sits down to write a story, she has to plan the plot, hold multiple ideas in working memory, shift between thoughts as the story unfolds, and self-monitor her work, all while trying to stay focused as her little brother throws popcorn at her head. She is getting the ultimate mental workout.

If we rob our children of this crucial brain workout, they risk missing out on the very process that builds a resilient, flexible, and creative mind — a mind that is capable of navigating complexity and thinking independently. Which, ironically, are the very skills they’ll need to thrive in a world powered by AI and technology.


How AI robs children of the learning struggle

Children’s brains are under construction for the first 25 years of their life, and early experiences shape how brain regions grow and link together, building the networks they will later use as adults for essential skills and thinking.

When we let AI do our children’s thinking, we strip their brains of the workout needed to uniquely wire their brain and develop their unique personal way of thought. It’s like giving a child a calculator before they can count — they’ll get the answer but never understand the logic behind it.

This becomes most detrimental during critical periods of development — when the brain is most hungry to learn, grow, and refine itself. These are the years when effort literally shapes the architecture of the brain. Once those windows close, the chance to rebuild itself becomes much harder.

We can already see the price of AI — not just in declining test scores, but in brain scans that show crucial regions of children’s brains failing to turn on:

Test performance: In a study of more than 1,000 high school students, researchers at the University of Pennsylvania found that students using a ChatGPT-style tutor became 127% better at solving problems. But when the AI was removed, their performance dropped, falling 17% below the students who never used AI. This shows that while AI can provide instant results, it may threaten the development of critical thinking skills.

This happens for a simple reason: the brain is a muscle. Effort strengthens it; shortcuts weaken it.

Brain activity: In an EEG study at Cornell University, students who used ChatGPT were found to have the lowest brain engagement during essay writing. Their essays were coherent, but described by English teachers as “soulless.” When they were asked to rewrite their essays later without help, they barely remembered what they’d written. It was like trying to recall a dream someone else had for you.

Meanwhile, students who wrote without AI showed higher brain activity. This showed up as better memory, greater originality, and a stronger sense of ownership in their work.

Other EEG studies have shown that writing by hand activates more brain areas and strengthens connectivity between regions compared to typing. It’s not just about forming letters, it’s about feeling them — the pressure of the pen, the scratching sound of pencil against paper, the visual tracking of letters as they are formed on the page — all of this gives sensory and motor regions a workout.

One child in a study put it beautifully: “When I write by hand, I can see what I’m thinking.”

Ying Xu, PhD, assistant professor of education at Harvard University, put it perfectly when she asked the critical question: “Are they actually engaging in the learning process, or are they bypassing it by getting an easy answer from the AI?”

Real learning only happens when kids wrestle with ideas, make mistakes, and figure things out for themselves.

If we shortcut that struggle, we shortcut their brains.

And no computer can ever match the limitless potential of a child whose brain is well-connected, creative, and endlessly adaptable.


Strategies to avoid the AI brain drain

I’m not suggesting AI is evil. It’s revolutionary, and here to stay. The goal here isn’t to ban it, but to reframe its role– AI should be a support, not a substitute.


The first strategy that comes to mind is simply: “Let kids try first on their own, then turn to AI for help.” Sounds good in theory. But I still remember how my kid-brain worked — if I knew the answer was coming, my first attempt would be half-baked. Why wrestle with the problem when a shortcut is waiting?

So, putting my child psychologist hat on, here are some potential ways around that trap — giving kids the struggle back.


1. Become a detective: “Can you trust this?”

Teach kids that AI doesn’t know things — it predicts them. It looks at all the data on the internet and then uses probability to guess what word or answer will come next. This means it doesn’t understand what it is saying, which makes errors highly likely.


How to do it

When your child gets an AI answer, try saying, “Awesome. Now our job is to be a detective and see if this is a good answer, and how we can make it even stronger.” Here are a few questions to ask:

  • Should we trust this? Where can we check if this is true?
  • What’s the source? How could we find an expert or a reliable book/website to back this up?
  • What’s missing? Is there another side to this story or an important detail the AI left out?
  • How can we make it better? What information do you need to add to make it clearer or more interesting?

Why it works

Cautioning that AI is not always right flips the brain from passive to active. This means higher-order executive functions get a workout, like analyzing, evaluating, and reasoning. Research shows that when students are prompted to question and justify information, they deepen their comprehension and problem-solving skills.


2. Use the Feynman protocol: “Teach it back to Teddy.”

Nobel physicist Richard Feynman’s principle was simple: You don’t truly understand something until you can teach it.


How to do it

Let your child use AI and say, “Ask AI to explain the concept to you. Once you think you have all the information you need, close the screen. Your mission is to help someone else understand the concept– you can choose me, the dog or your teddy, we are all great listeners.”

For an additional brain workout, you can say “If you get stuck [which they will], find out more information to fill in the gap.”

Why it works

Explaining something strengthens memory more than rereading ever could. It also exposes what they really know, not just what they’ve skimmed. And it flips motivation: The goal is no longer “finish the assignment,” but “master it well enough to teach.”


3. Protect the analog brain: “Unplug and play with me”

Make space for activities AI can’t replicate: board games that train strategy, unstructured play that builds creativity, physical books that train focus, and family dinners that nurture empathy and conversation.


How to Do It

The idea is to make it fun and enticing (never give away your hidden agenda). To do this, try:

“I was about to set up [Suggest a specific game]. I haven’t played it in forever and I need a partner in crime/need someone to destroy me. You in?”

Or

“This game is hilarious when people get competitive. The person who loses has to wash the dog next.”

Why It Works

Brain imaging studies show that face-to-face conversation, reading offline, and handwriting light up more regions of the brain than typing or screen time.

---

As parents and teachers, it’s our duty to safeguard the sacred work of childhood: the messy, but always beautiful, process of learning how to think.


Remember, brains are built, not born.


Source: Dr CJ Yatawara - Tiny Brains expert

https://medium.com/wise-well/ai-is-disrupting-how-young-brains-grow-c1f30196ac63

Wednesday, August 20, 2025

健康碼就是電子軟禁

 


健康碼是什麼?是電子軟禁(house arrest through electronic device)。

不給你綠色的健康碼,你連家門都無法出去。更甚者,你連家住的大廈門口都不能進去,要住到政府為你預備的、香港醫護為你親自搭建的社區臨時居住設施——集中營、再教育營,你要在裡面學習愛國教育,直至檢疫滿意變成綠色碼。大廈看更同情你也無法為你開門,因為他沒有你的DNA認證來為電子門鎖下指令。

電子軟禁?好新奇嗎?不是。我們香港人演練過的,而且官民合作愉快。今年三月二十五日,口罩爭論的時候,政府同時推出的電子手環,用來隔離懷疑感染的機場入境者。

當時偶然有些佩戴電子手環的人出街,又很巧啊,在天氣還冷的手環,這些人會吧長袖衫捲起,露出手環,但以香港人那種息事寧人的態度,是不會有反應的,但恰巧又會有議員助理看見,熱烈舉報,在KOL的區塊瘋傳。出街的未必是政府安排的人,但舉報的,肯定大部分是得到政府歡心的人員。大家看看當時熱烈舉報手環出街的黃絲議員,看看他們當時那副嘴臉,就知道港共政府玩弄的把戲和人員安排——他們都是你投票的,你用稅金來供應的,而他們是踏在去年街上的屍體而當選的人血饅頭議員。

香港人那種驕傲和愚昧,那種恃勢凌人的刻薄,是上天賜予共產黨的美食。在香港,也只有陳雲懂得怎麼保護你,而他受到很多報紙和論壇排斥,當然,那些也是共產黨安排的媒體和KOL,而香港人民熱烈跟隨的。

各位可以在下面貼一下,當時捉電子手環的新聞和人物。 

Source: 陳雲

https://www.patreon.com/posts/jian-kang-ma-jiu-40635129

Tuesday, August 19, 2025

7 Visualization Hacks Every Data Analyst Should Know (But Most Don’t)

Last Tuesday, I was presenting quarterly sales insights to our C-suite when the CEO stopped me mid-sentence. “Hey, this chart is… confusing. Can you make it tell a story?”

That moment hit me hard. I’d spent 40+ hours analyzing customer behavior patterns, uncovered a 23% increase in retention after our product update, and built what I thought was a comprehensive dashboard. Yet my visualization failed the most basic test: clarity.

Here’s the uncomfortable truth — 67% of data analysts can crunch numbers like wizards, but their visualizations look like rainbow spaghetti threw up on a spreadsheet. I learned this the hard way after 8 years in analytics roles across fintech and e-commerce.

Today, I’m sharing 7 visualization hacks that transformed how I communicate data insights. These aren’t textbook theories; they’re battle-tested techniques that helped me go from “confusing presenter” to “data storytelling expert” in 6 months.

Image - Dribble.

Why Most Data Visualizations Fail (And It’s Not What You Think)

Before jumping into the hacks, let’s address the elephant in the room. Most data analysts approach visualization like they approach SQL queries — technically correct but missing the human element.

I used to create charts that answered every possible question. Color-coded by region, segmented by time, filtered by product category. Technically impressive? Yes. Actionable for business decisions? Not really.

The problem isn’t technical skill. It’s empathy. We forget that our audience doesn’t live and breathe data like we do. They need guidance, context, and most importantly, a clear path to action.


Hack #1: The “So What?” Test for Every Chart

Every visualization should pass this simple test: A busy executive should understand the key insight within 5 seconds.

Before: I created a complex multi-line chart showing website traffic trends across 12 months for 8 different channels.

After: I highlighted the one insight that mattered — organic search traffic dropped 34% in Q3, directly correlating with our competitor’s aggressive SEO campaign.

Implementation: Add a single sentence annotation to every chart stating the main takeaway. Use tools like Tableau’s annotation feature or Python’s matplotlib text() function.

# Example: Adding context to a matplotlib chart

plt.annotate('Organic traffic declined 34% due to competitor SEO push', 

             xy=(7, 15000), xytext=(8, 20000),

             arrowprops=dict(arrowstyle='->'))

Impact: My presentation time dropped from 45 minutes to 20 minutes, and stakeholder follow-up questions became more strategic instead of clarifying basic trends.


Hack #2: Color Psychology for Data Impact

Colors aren’t decoration; they’re communication tools. Most analysts use default color palettes that convey zero meaning.

The Framework:

  • Red: Problems, declines, urgent attention needed
  • Green: Success, growth, positive metrics
  • Blue/Gray: Neutral data, benchmarks, historical context
  • Orange/Yellow: Warnings, moderate concerns

Real Example: When presenting customer churn analysis, I colored churned segments in red, retained customers in green, and at-risk customers in orange. The executive team immediately focused on the orange segments — exactly where we needed intervention.

Pro Tip: Never use more than 4 colors in a single visualization. Your brain can only process so much before it gives up.


Hack #3: The Data-to-Ink Ratio Revolution

This hack alone improved my visualization clarity by 60%. Remove everything that doesn’t directly support your insight.

What to Remove:

  • Unnecessary grid lines
  • Redundant legends
  • 3D effects (seriously, stop this)
  • Multiple y-axes unless absolutely critical

Before/After Example: My original sales dashboard had 47 visual elements. After applying data-to-ink principles, I reduced it to 12 elements. The result? Stakeholders could identify trends 3x faster.

Implementation in Excel:

  • Remove chart borders
  • Lighten grid lines to 25% opacity
  • Delete redundant axis labels
  • Use direct labeling instead of legends


Hack #4: Progressive Disclosure for Complex Data

When you have complex data stories, don’t dump everything at once. Guide your audience through a logical sequence.

The 3-Layer Approach:

  1. Overview: High-level trend or summary metric
  2. Zoom: Segment or time-period focus
  3. Details: Specific data points or outliers

Case Study: Analyzing user engagement across our mobile app, I started with overall monthly active users (layer 1), then segmented by user acquisition channel (layer 2), and finally highlighted retention patterns for each channel (layer 3).

Result: Instead of one overwhelming dashboard, stakeholders could absorb insights incrementally, leading to more thoughtful discussions about each layer.


Hack #5: The Comparison Anchor Technique

Humans are terrible at interpreting absolute numbers but excellent at understanding comparisons. Always provide context.

Instead of: “We acquired 2,847 new customers this month” Try: “We acquired 2,847 new customers — 23% above our target and the highest in 6 months”

Visual Implementation:

  • Add benchmark lines to show targets or historical averages
  • Use small multiples to compare similar metrics
  • Include percentage change annotations

Python Example:

# Adding benchmark line to show context

plt.axhline(y=target_value, color='gray', linestyle='--', 

            label=f'Target: {target_value}')

Impact: When I started adding comparison anchors to our KPI reports, decision-making speed increased by 40% because stakeholders could immediately assess performance relative to expectations.


Hack #6: Interactive Filtering for Stakeholder Engagement

Static reports tell one story. Interactive dashboards let stakeholders discover their own insights.

Strategic Implementation:

  • Add filters for time periods, regions, or product categories
  • Enable drill-down capabilities from summary to detail views
  • Include hover tooltips for additional context without cluttering

Real Success Story: I built an interactive sales performance dashboard in Tableau where regional managers could filter by their territory. Suddenly, they were spending 30+ minutes exploring data instead of glancing at static reports for 2 minutes.

Tools Recommendation:

  • Beginner: Excel with slicers and pivot tables
  • Intermediate: Tableau Public or Power BI
  • Advanced: Python Plotly Dash or R Shiny


Hack #7: The Storytelling Arc Framework

Every great visualization follows a narrative structure: Setup → Conflict → Resolution.

  • Setup: Establish the baseline or normal state 
  • Conflict: Highlight the problem, opportunity, or change 
  • Resolution: Show the outcome or recommended action

Example Application: Analyzing customer support ticket volumes:

  • Setup: “Support tickets averaged 150/day in Q1”
  • Conflict: “Tickets spiked to 340/day after our product launch”
  • Resolution: “Implementing chatbot reduced tickets to 180/day within 2 weeks”

Visual Elements:

  • Use annotations to guide the narrative
  • Highlight the conflict point with contrasting colors
  • End with clear next steps or recommendations


The Career Impact: Why These Hacks Matter Beyond Pretty Charts

After implementing these 7 hacks consistently, my professional trajectory changed dramatically:

  • Promotion Speed: Advanced from Senior Analyst to Lead Data Scientist in 18 months
  • Stakeholder Trust: C-level executives started requesting me specifically for quarterly reviews
  • Project Success Rate: Data-driven initiatives I presented had 85% approval rate vs. industry average of 60%

More importantly, I stopped being the “chart guy” and became the “insights guy.” My visualizations weren’t just reporting data; they were driving business decisions.


Your Next Action: Pick One Hack and Implement It This Week

Don’t try to revolutionize all your visualizations overnight. Pick the hack that resonates most with your current challenges:

  • Struggling with stakeholder attention? Start with Hack #1 (So What Test)
  • Charts look cluttered? Apply Hack #3 (Data-to-Ink Ratio)
  • Audience seems confused? Try Hack #4 (Progressive Disclosure)

The goal isn’t perfection; it’s progress. Every small improvement in how you visualize data compounds into massive career advantages over time.

Remember: In a world drowning in data, the analyst who can tell compelling visual stories doesn’t just survive — they become indispensable.

What visualization challenge are you facing right now? Which hack will you try first?

----

Tired of wasting hours on messy data and interviews that go nowhere?

I created two resources that changed everything for me:

1️⃣ ChatGPT Prompt Bundle (15 Productivity Prompts)

🚀 Turn AI into your co-pilot — summarize data, build insights, and save hours every week.

2️⃣ Top 50 SQL Interview Questions

🎯 Real-world, scenario-based SQL challenges with clear explanations — built to help you ace any data role.

These aren’t cheat sheets. They’re career accelerators.

Make data work for you — not against you.

👉 Get them today and reclaim your time + confidence.


Source: Analyst Uttam

https://medium.com/ai-analytics-diaries/7-visualization-hacks-every-data-analyst-should-know-but-most-dont-e35954ef1102

Wednesday, August 13, 2025

台灣大罷免鬧出納粹黨風波,惹來德國在台協會譴責,何解民進黨陣營容納這種人物呢?

 

圖一、閩南狼披露419罷團運動的徽章來自納粹的老鷹。(圖:中天新聞)


這新聞觀察了很久,包括在我寫台灣大罷免的評論的時候。最近,在台灣宣傳大罷免的網絡紅人閩南狼和八炯,公開分裂,鬧出風波,令民進黨出醜,那麼就值得評論一下了。

閩南狼指控八炯,要走納粹路線,公開私訊對話,指八炯要學習納粹操控大眾輿論節方法,又用納粹鷹做大罷免集會LOGO。八月四日,德國在台協會表示,譴責所有任何崇拜、贊美或淡化納粹主義的說法或行為,八炯透過影片道歉,並表示只是私下自嘲,承諾未來會減少線下活動。早在四月十六日,德國在台協會譴責有關大罷免的宣傳行為,嚴正譴責大罷免運動其中的宋建樑此舉為「無恥行為」,並強調納粹象徵「對人類的鄙視與迫害」。

八炯利用納粹的老鷹代表419的造勢活動,而賴清德總統更高度肯定此場的造勢,被國民黨質疑:難道賴清德一直與納粹同行?


這是中共一向的做事方式了。在反對者的陣營安插最激烈的人物,強詞奪理,只求吸引群眾,按道理這些人物很快就會用完情緒動員力,然而給這些人的舞台、媒體報導和評論(包括極其激烈的抨擊和謾罵)不絕,只需要曠日持久,這些人就可以排擠疲憊的老敵人而䇄立敵營,令敵方不得不姑息甚至合作。這種方法結合現代媒體(以前的報紙雜誌電台電視台,現在包括互聯網和社交媒體),是現代政治技術,無以名之,可以稱為安插假敵之法,是間諜術的一種。

舊時在香港的民主中國論的締造者(來自台獨派的包錯石)、台獨論的締造者,近年港獨論的締造者,這些極端的言論締造者,都離不開這種套路。當然,有時候也有例外的、沒有政治背景的激進者,但他們只如流星,沒有持久的續航力,閃亮一時之後就離去,謀生要緊;那些可以持續取得資源、發布平台和輿論關注的,沒有什麼營生活動的,說他們沒有奇怪背景,是難以置信的。

圖二、今年四月十五日,台灣新北地方檢察署約談罷免李坤城案領銜人宋建樑(圖),宋建樑穿著象徵納粹的「卐」字臂章,手比納粹敬禮手勢,還拿著希特勒著作「我的奮鬥」進入新北檢複訊。圖片:經濟日報


近年在台灣和以前香港的所謂學術圈裏,樹立內亞史觀,說中國的文明器物和思想都是來自中亞內陸,美化蒙古等游牧民族的統治,並且持續出書,持續在網絡出帖文的,都是這種討論,目的是用這種排斥理性討論的激烈言論來吸引民眾,特別是年輕人的反叛者,掩蓋有意義的學術討論,也令外界認為反共陣營都是充斥這些思想偏狹的人物,敬而遠之。

這是筆者從事海外民運以來的觀察,也是我遊學德國的讀書所得。我在本土運動的時候,稱這些人為混入陣營的粘貼式炸彈,等候時機引爆,企圖將陣營消滅。


新聞詳情:

「閩南狼」於8月3日發佈影片,指控「八炯」曾私下研究納粹如何群眾動員「反共」,甚至想發明「新的手勢」、「把部分人當成猶太人」以及建立「納粹衝鋒隊」。

納粹政權領袖希特勒(Adolf Hitler,又譯希特拉)於1933年崛起,他奉行法西斯主義(fascism),推動以極端民族主義和反猶主義為核心的政策,「雅利安人」(Aryan)被塑造成「純正德國民族」的理想象徵。二戰期間,近600萬猶太人在大屠殺中遇害,其他少數群體也被迫害,包括吉普賽人(Gypsy)、同性戀者、共產黨人等政治異見者。

納粹衝鋒隊(Sturmabteilung, SA)是納粹黨的凖軍事組織,參與監控社會和政治鬥爭,用恐嚇和暴力手段對付被納粹黨譴責的對手。

影片還提及,「八炯」曾在4月19日的罷免集會上使用「納粹老鷹」的視覺符號,相關標誌在罷免運動中多次出現。


Source: 陳雲

https://www.patreon.com/posts/tai-wan-da-ba-na-136392138

Thursday, August 07, 2025

Mitschuldigkeit,一個百般沉重的德文字

(附上的短片是去年面書有人拍攝到,上水公園的外判除草工人在扮工,但用除草機傷害了樹根。面書帖文的連接一下子沒記錄,也許遲些找到。)


前日,同道在本欄問我:「這些本身無病的樹被風雨吹倒後,政府會重新栽種嗎?」

我的回答:「不會。即使原地有樹木種子生了樹苗,除草的工人也會將之打斷!此地實施的是極為恐怖的不生之政。

問題是:工人是可以放過樹苗的,但工人寧可將之用剪草機器打斷!這就是一般的香港庶民。」

我在德國遊學學時期,用了足足五年時間來理解納粹德國、東德、蘇聯和中共的政治運作。對於中共的政治運作,我也親自從民運圈子接觸的老幹部那裡學來很多經驗。

在德國的去納粹化的過程中,一個顯著的疑團是:平民、非納粹黨的人、不涉及納粹黨運作的人,是否無辜?

學界經過激烈辯論和案例研究之後,得出的一個概念,用德文來表示,就是:Mitschuldigkeit。中文沒有這種構詞,英文也沒有。德文的Mit是參與、與,schuld是罪,dig是形容詞後綴,keit是名詞後綴。Mitschuld就是同罪、共犯。在納粹德國,平民是共犯,不是無辜者。

參考的案例,就是柏林圍牆射殺案。德國統一之後,一九九二年,法官審判當年射殺越過圍牆的偷渡者的案件,開槍射殺的士兵被判徒刑三年,不准假釋,故意瞄不準而放槍警告的士兵無罪釋放。律師辯稱這些衛兵僅僅是為執行命令,別無選擇,罪不在己。然而法官西奧多·賽德爾(Theodor Seidel)卻不這麼認為:「身為警察,不執行上級命令是有罪的,但打不準是無罪的。身為一個心智健全的人,此時此刻,衛兵有把槍口拾高一厘米的主權,這是你應主動承擔的良心義務。這個世界,在法律之外還有『良知』。當法律和良知衝突之時,良知是最高的行為準則,而不是法律。尊重生命,是一個放之四海而皆準的原則。」(餐廳食飯要拍卡登記,餐廳員工是否要執行防疫惡法呢?——柏林圍牆射殺偷渡客的判案啟示 FEB 18, 2021 AT 12:38 PM https://www.patreon.com/posts/47696222)

基於此,我反對特朗普的政治顧問余茂春將中共與大陸人分開處理,我認為兩者是同一回事。故此在二〇〇三年本土運動時期,當大陸人在自由行初期搗亂香港市面秩序的時候,我在明報寫了一篇短文,說他們不是孤單的,他們拖着長長的帝國的身影。於此,梁文道在報紙與我激辯了一回,說大陸人也是鄰居,成為本土運動的一時佳話。

英文一句政治諺語Every country has the government it deserves或People get the government they deserve. 人民得到他們應得的政府。有什麼樣的國民,就有什麼樣的政府。此話雖然令人氣憤,也不大公道,因為有些政府得到境外資金和技術甚至軍事支援,用精銳的軍隊殺害大部分文人和反抗者之後,本國平民是無可選擇的。然而這個政府生存下來幾十年,而且壯大發展,那麼這些人民就有共犯的責任。

Every country has the government it deserves. 這句話我最早聽到的,是我從事德國民運的時候,有一次,大概一九九四年,去了瑞士的日內瓦開會,在火車站接車的是當地的老國民黨華祈石先生,是一名研究火車訊號系統的數學專家。他大概已經老邁甚至身故,故此可以寫出名字來紀念。在日內瓦街頭行走的時候,他拍了我的肩膀,說我不必為了中國民主來操心,說的一句英語,就是People get the government they deserve. 當時華老先生這句話,比起中國沒有憲政民主自由,令我更感到心情沉重。


Source: 陳雲

https://www.patreon.com/posts/mitschuldigkeit-135943860

Monday, August 04, 2025

AI Generated Feedback vs. Human Feedback

In July 2025, Canvas LMS announced integration with ChatGPT for instructor use. Features touted by the tech company include the generation of image descriptions, rubrics, and feedback for assignments.

It’s that last one that makes me pause. Many other educators, too.

Because how ethical is it to ask students not to use AI if instructors use it for feedback? Isn’t giving feedback part of the jobs we’re paid to do?

On the other hand, some instructors are bogged down with ridiculous student loads. AI-generated comments may be the best way to give timely feedback for formative assessments.

So what’s a teacher to do?

Research into this area is limited, but here are some findings to help you make an informed decision about relying on AI for feedback.


Human Feedback > AI Feedback

Steiss et al (2024) compared AI-generated feedback with feedback provided by trained instructors in five different areas: essay criteria, directions for improvement, accuracy, supportive tone, and prioritizing important feedback comments. Instructors scored better in four out of five areas. AI only scored better in criteria-based feedback. As much as AI has improved in essay feedback, human scorers still have the advantage in most areas.

A couple of things to note in this study. One, the instructors in the study received training. How many instructors receive training and professional development in giving feedback? How many colleges provide intensive work in feedback writing to their pre-service teachers? In my 25 years in education, I’ve received none outside of my own pursuits. This makes me wonder how a random selection of teachers would perform in this situation, not to mention the need for more education geared toward writing effective feedback to students.

Two, instructors outperformed AI in four areas, but AI wasn’t far behind. This wasn’t a slam dunk for instructors, merely a slight edge. This opens the door for other considerations, such as available time and student load. Timeliness is a key factor in effective feedback, but student loads of 150–200 students can take a teacher several days (or weeks!) to give fully developed feedback. Does the timeliness that AI provides outweigh the slight advantage instructors have in those other feedback categories?

Some teachers may decide the answer is yes.


Student Perceptions

We can’t forget the other key ingredient in this dilemma: students. After all, they’re relying on our expertise to guide their learning. What are their perceptions of teachers using AI feedback?

The results were mixed. According to Nazaretsky et al (2024), students generally preferred human feedback, even if those same students rated the AI feedback as higher in quality. In contrast, Zhang et al (2025) found that students preferred the feedback produced by AI or co-produced by AI with human modifications. Students in the Zhang study rated the AI feedback as less genuine after they learned that the feedback was given by ChatGPT. They didn’t lower “co-produced” (AI feedback modified by a human) ratings in the genuine category, though.

So genuineness is important. Students want human interaction.

To take advantage of both AI and human feedback, the answer may be what Zhang et al (2025) term “co-produced” feedback and Nazaretsky et al (2024) call “human-in-the-loop.” Instructors use AI to develop feedback for student work and then modify those comments by adding or deleting comments, prioritizing key suggestions, and adding encouragement.

Still, teachers need to be competent at feedback to effectively modify the comments that AI provides.

The research in these areas is limited. It was conducted with college students and instructors, not in secondary classrooms. Zhang et al (2025) also note that their findings may be different from Zaretsky et al (2024) due to increased time pressures that negatively affected the quality of human feedback.

The human component is a problem for any study. Many factors can affect the quality of human feedback, including time, training, experience, and stress levels. Different students will also prefer different approaches; some appreciate more encouraging feedback, while others want brutal honesty. These preferences will also affect how students rate AI and human feedback.

Knowing what’s best when using AI in student feedback is complicated. Both Zhang et al (2025) and Nazarestsky et al (2024) agree that instructors and schools need to consider the ethics of AI use in feedback.


Transparency in AI Use in Feedback

Instructors using AI for feedback need to be honest and explain their reasons for doing so. The need for a quick turnaround in feedback may be crucial for some assignments. Or a family emergency has taken over your life, and AI would provide better feedback than you’re able to give.

Otherwise, if we pass AI feedback off as our own, we’re just as guilty for using AI unethically. Students rely on their teacher's expertise to guide their learning, and we have a professional obligation to provide that guidance.

Like everything else in the AI and education world, opinions vary across the spectrum, and the best thing you can do is stay in touch with your principles and stay updated on the latest research on the effectiveness of AI.


Source: Melissa Pilakowski

https://medium.com/educreation/ai-generated-feedback-vs-human-feedback-639321d530b8

Thursday, July 31, 2025

一個國家有幾強大,看它能否保存傳統度量衡

 

圖、攝於上環市政街市的十進制推行廣告牌(圖片來源)

「出咗半斤力,想話攞返足八兩,家陣惡搵食,邊有半斤八兩咁理想!」「今個星期六,跑馬地有六化郎大賽。」「加時之後南華和精工依然打和,球證決定雙方互射十二碼!」這是一九七〇年代,香港的收音機常聽到的唐制和英制。

一個國家有幾強大,看它能否保存傳統,特別是傳統的度量衡。美國向伊朗的核地堡扔了B2鑽地砲彈,準確到出奇。這種計算,當然用米特制(metric system),即是用metre、kilogramme和litre來做長度、重量和容積的公制。然而,科學領先世界的美國,本地保存了英尺、磅和加侖的英制(略有美國的變化)。曾經稱霸世界的大英帝國也是一樣,科學用公制,民生用英制。(維基百科英文詞條解釋甚好。https://en.wikipedia.org/wiki/English_units)

中國大陸或台灣地區,就因為當年民國建國的一群低水平的知識精英而紛紛投靠源自法國的公制,毀滅本地的唐制(華夏度量衡)。中國大陸(共產中國)採取了德國的做法,在名號上容納傳統的名稱,但實質改為公制,德國一磅就是半公斤,共產中國的一斤就是半千克(公斤),唐制名存實亡。

唯一集合唐制、英制和公制的就是香港!科學計算用公制,買樓用英制(至今香港人最寶貴的資產依然用平方呎計算!),買金用両,買藥材用両、錢和分(揮發藥如麝香用分),街市仍有斤両或磅、安士,一九七〇年代,許冠傑的《半斤八両》唱通街。在足球上,香港慣常使用「碼」這個單位,如十二碼等。九七之前,跑馬有六化郎比賽。以前香港賽道以英制為單位,一個化郎(Furlong)是八分之一英里,二百二十碼,換算公制大約是201.168米。

此外,也有一些傳統量度單位是香港獨有的。新界的農田用斗種來量,即是一斗的穀種可以種多少的地。一斗種大約等於674.5 平方米或7,260 平方呎,少於一斗種的,就用幾多升種來計算。問題來了。以前的鄉民是不用尺去量度斗種的,全憑耕田的經驗來意會。


香港保存了唐制,是大陸及台灣所無

一九六〇年代,我讀小學的時候,小學的算術課本常有一個口訣:三斤等於四磅。英國治理香港,好快訂立英制與唐制的換算法例。據原香港法例一八八四年第廿二條,一斤為11⁄3常衡磅(即三斤等如四磅)。現時香港法律規定一斤等於一百分之一擔或者十六両,即0.60478982 公斤(參見度量衡條例。1988年第351號法律公告。https://www.elegislation.gov.hk/hk/cap68!zh-Hant-HK)。香港比起中國大陸更為保護傳統,依然沿用司馬斤。中國古代有個官職叫「司馬」,司馬主要掌管軍事,其中因為糧秣管理需要秤重,於是「司馬」就用來命名重量單位。清朝官府頒布的庫平制(清朝康熙年間制定發布的營造尺庫平制),中華民國在北洋政府時代沿用,但在英治香港則不採用,依然沿用民間的司馬斤。一司馬斤相等於604.79克,庫平制一斤相等於596.82克。

英治政府於一九七六年頒布《十進制條例》,香港逐漸部署由政府部門內部法例起步,將其所用的非十進制單位以國際單位取代,其後更於一九七八年一月一日成立「度量衡十進制委員會」(Metrication Committee),將十進制推至社會各行各業。然而「半斤八兩」舊有度量衡依然頑強,委員會於一九九八年一月一日以「完成使命」的名義宣布解散。

由於公制的推行,格外令人看出,傳統度量衡帶有神秘數理。公制是設計出來除得盡的,另外是互相化約的,如一立方米的純水就是一千公斤重。

然而,如果你用一米來換算英尺、用一公升來換算加侖、用一公斤來換算磅之類,多數都是無理數(如1 kilogramme = 2.20462262185 pounds),或餘數很多的(如1 meter = 3.280839895 feet)。

無理數是什麼呢,就是測不准!這正是最高的科學標準。要將之計算到有效數字,例如1 kg=2.2 pounds,是實際情況的折衷。如果你日常已經接收了無理數和折衷,你就是一個對於科學或技術抱住懷疑態度和勉強接受態度的人,即是說,你是一個理性的人,或者說,你是人,而不是機器。

傳統並非反科學,而正正就是科學精神之體現。我這帖文出來之後,恐怕中華大地的主政者會更努力去除傳統度量衡了。


---

改寫自筆者面書帖文2025-06-25 14:12 https://www.facebook.com/wan.chin.75/posts/pfbid0D82mP5hQ6FU4mUv8viGVummfUp8QjhtX1ynaPU8d3SHeQUwWGxd8EoG2whmSx9z4l


Source: 陳雲

https://www.patreon.com/posts/yi-ge-guo-jia-ji-135377912