貪

 由於呢個概念對我嚟講太難，要寫返低先：

「貪」就係為咗一啲自己唔係真係想要嘅嘢，不自覺咁放棄咗自己本來有嘅嘢。

"First Love"

其實我一直唔太鍾意宇多田光嘅唱腔，首《First Love》歌曲旋律好好聽，但佢個 "interpretation"(?!) 有啲太硬太多稜角或者亦都係太沉重。

由於頭先喺度 copy and paste 啲 2022 年嘅文章，又咁啱喺 netflix 見到有佢 2018 年演唱會錄影，咁就喺 background 度播放。佢 2018 年嘅《First Love》就真係超好聽。一流。

可惜可能比較難 extract 到個 mp3 出嚟。

我見 iTunes 有佢 2022 年版本都好似 OK，但個人感覺始終唔及佢演唱會嗰個咁好聽。

Pushing the limits

I suppose this is a book review (tl;dr: do not read. However, google for "exosomes", they are interesting suff):

"The Truth About Contagion: Exploring Theories of How Disease Spreads" - Cowan, Thomas; Fallon Morell, Sally (2021)

You know you're pushing the limits of sanity when stuff you read offers a totally novel perspective, yet totally wrong.

My take is that the author(s) identified and expounded upon the problems with Koch's postulates (for establishing causal relationship for pathological agents and disease), and then basically hallucinated an even worse theory of disease using EM waves and food/water toxins (this theory of course fits even worse with Koch's postulates).

There's an interesting mention about "exosomes", apparently it's a packet of stuff that gets produced by eukaryotic cells. According to an (abtract of an) article in Nature it could perform intracellular DNA/RNA transfer: https://www.nature.com/articles/ncb1596 . Now this stuff is interesting.

Given the exosome lateral DNA transfer mechanism, it's very weird that the author of the book had to rely on homeopathic telepathic DNA transfer to get his point across. He claims that electromagnetic resonance instead of physical contact transfers herpes viral DNA...  and he relied on it just to show how "viruses" (which he claims are misidentified exosomes) would give communities of organisms an evolutionary advantage by lateral transfer of defense mechanisms against external stresses... I can't say for sure this is false (the telepathic DNA thing is apparently from a Nobel prize winner...), but the equivalently useful and less unproven claim would be that exosomes perform lateral transfer using the usual methods of physical contact and achieving the same evolutionary function.

You know how much a book is worth when the reader has more clarity than the author on a subject that the reader has just learnt...

Still, the highlight on the problems with Koch's postulates do indicate that modern germ theory might not be as solid as you'd have been taught in school -- maybe still 99+% solid, but caveat emptor. Especially with viruses, the alleged pathogen is basically an attack from the "information" vector. We software engineers know how hard it is to actually understand what the attack actually is. The mystic in me thinks that there can still be "bad spirits" in information packets that we aren't mentally capable of understanding.

M

 show me help me teach me

A Magician's Choice

Poster of an old Unix magician in the time antiquity circa 1990

Let's say with the alleged power of subjective truth (link contains required reading) one miraculously did something that was highly improbable, but within bounds of technical possibility.

Now, a skeptic who doesn't believe in subjective truth comes and asks demands to apply "objective truth" methodology to test whether it was really a miracle or actually just a fluke. That is, they request repeating the attempt at doing the thing, and "test" what the "probabilities" of it happening are, so that they can "calculate the objective probabilities" of it happening. 

When they do this, the tester actually is actually *confirming* the objective world view. With the alleged power of subjective truth, the world complies with the objective world belief and behaves "objectively", giving them the "objective probabilities" that they want.

It may be easier to explain with a story.

Let's say a Healer (H) performs an apparently miraculous healing ritual on a subject, completely curing an illness never before cured. H insists that this is done with "magic" and requires belief and submission to the cosmic powers for this to happen.

Now, skeptics arrive and try to analyze the scene. They insist that H is a fraud because this is unscientific woo-woo, and if H really had "powers" H should be able to repeat this process and cure anyone. Of course, either H refuses, or H fails to provide any subsequent healing.

We are all taught in school the objective, no-nonsense interpretation: this is just a fluke, there's no magic going on, and if these tricksters are lucky for a while, eventually they will expose themselves as frauds, because the universe works with fundamental laws of physics that don't have human concepts of coincidences etc.

With the subjective interpretation, this is the power of subjective truth at work. Both H and the first subject knew the ritual was going to work, and it did. However, when skeptics came to the scene, the intentions (which is a serious matter with subjective interpretation) changed from the act of healing, to trying to test an hypothesis that magical healing exists. The only reason why this needs testing is because somebody doesn't believe in it and wants "proof". The fundamentally flawed (from subjective perspective) axiom of frequentism is that if you repeat the "same" thing many times you will have a useful probability distribution. However, how it actually works is that the believer of frequentism actually believes in objective truth, and they "know" frequentist probabilities work. So the universe complies accordingly, as per the subjective interpretation. In short, given a bunch of people believing in objective truth, in particular, "knowing" that objective truth is reality, the power of subjective truth makes it a reality. And therefore, in the end, objective truth is what we observe.

I'm not even saying either of the interpretations are correct. I'm just saying, it seems, the subjective truth and objective truth are roughly compatible (as I always suspected). Which one you "should" believe, is, in my opinion, a subjective choice. The classic Occam's razor says the simpler reductionist objective model works fine. I personally disagree because frequentist assumptions are wrong (you can never truly repeat anything the second time). Of course the elephant in the room for the subjective interpretation is the subjective consciousness, but at least presumably everyone is dealing with it no matter you want it or not (which is forced upon us at quantum scales). But then, if the two interpretations are roughly compatible, personally, I can see no reason why I would not choose the one that allows for magic to happen.

When I was a child, I would have said I wanted to be a "magician" when I grew up, if only it were a legit occupation. (In many ways I'm already close. I make a living by conjuring digital constructs from chaotic bits.)

Despite widespread skepticism, there is still much evidence for "magic". For example, the medical research industry spends disproportionate amount of resources trying to combat the "placebo effect". It's the one major thing getting in the way of "scientific progress", just because sugar pills can often work better than complex chemicals. I'm not sure I've come across any reasonable explanation of this phenomenon other than subjective truth.

Don't get me wrong. I unequivocally support developing medicine that works regardless whether you believe it or not. I'm just saying, while the objective methodology is useful, as honest seekers of truth, we can't just say it is The Truth because the tools that came from it are useful and convenient. In the same way, we can't just say subjective truth is wrong because the tools it gave us are "unreliable" and inconvenient.

The truth is indeed often dangerous. Still, the magi's choice is never in doubt.

Healing

https://www.youtube.com/watch?v=aoAH4li4_D8

Youtube 啲 algorithms 真係好神奇嘅嘢嚟，總係會搵到啲我有興趣睇嘅嘢。前面個幾鐘嘅論討都唔係重點（啲故仔幾有趣但我大概有個idea）。重點係後面首歌。

我第一次係俾歌聲嘅頻率同力量感動。感動到忍唔住抽泣。歌聲嘅治療力量真係literally隔住個芒都feel到，真係好犀利。而家一小時後個心輪都仲係有啲激動。好痴線。

(冇圖) 老師老師圖鑑 - 前篇

 

「乜你做老師都唔識呢啲嘢咩？」約翰福音 3 章 10 節

今日就講下一啲我認可兼夾唔會俾耶穌串到嘅老師s。同埋一段勁長嘅引言。(好重要㗎)

近日開始探討非唯物主義嘅現象，普遍稱為神秘學、New Age、靈異嘢之類嘅東東，所以寫嘅文大多都係嗰個方向。呢方面嘅老師s，要由傳統科學同邏輯嗰邊講起。

可能你會問，乜神秘學都要鑽研科學同邏輯㗎咩？答案：係呀，好很重要。重要到要重覆多兩次：

係呀，如果你對研究神秘學有興趣，一定要有足夠嘅科學同邏輯根底。

係呀，如果你對研究神秘學有興趣，一定要有足夠嘅科學同邏輯根底。

（近日咁啱見到，搵多位老師再講多次： https://youtu.be/uFdnIRA1X0M?t=360 ）

點解呢？因為科學嘅「正面結論」係無懈可擊嘅。即係反覆得到科學驗證嘅公理，例如萬有限力呢啲，作為我哋身處嘅物理世界嘅現象嚟講，係幾乎冇可能錯嘅。世人對科學嘅謬誤主要得兩個：(1) 以為「科學」=「科學家普遍相信嘅嘢」 (2) 以為「科學唔認知嘅嘢」=「唔存在」。 但講到尾，科學話一定存在嘅嘢，一定會發生嘅現象，要假設係普遍事實先。否則，你話想要否定牛頓力學，即係講緊你識得喺水上行路，你係咪真係得先？

所以要研究所謂神秘學，首先要知道科學講啲乜，同埋冇講啲乜，同埋點解佢可信，同埋佢有啲咩位唔太可靠。掌握到呢啲嘢之後先有得脫離科學嘅範圍。

順帶一提，科學家都唔一定係「科學家」。之後會再講到詳情。

牛頓

牛頓係西方經典科學舉足輕重嘅人物。現時我哋對萬有引力、微積分、光學等等嘅認知都有佢嘅貢獻。由於佢太出名，所以淨係講啲教科書入面比較少提嘅嘢 —— 如果你以為科學家一定唔信呢味嘢，咁以下呢啲可能會顛覆你對「科學家」嘅認知：牛頓除咗搞物理學之後，仲搞咗一堆神秘學嘅研究。有人就咁話佢： "Newton was not the first of the age of reason, he was the last of the magicians." - John Maynard Keynes

詳情可以望下 https://en.wikipedia.org/wiki/Isaac_Newton%27s_occult_studies 

值得留意嘅係，雖然牛頓花咗唔少時間去研究鍊金術之類嘅嘢，但去到最後影響後世最大嘅仍然係佢對自然科學嘅貢獻。成效嚟講，科學係直頭「立竿見影」嘅。而好多神秘學嘅嘢都係似是而非，你個人冇咁上下緣法，係冇得用理性去理解。所以點解我要一再重申一定要有足夠嘅科學同邏輯根底：就算強如牛頓，佢啲神秘學研究都未必真係有咩成果。你係咪覺得你勁過佢？

(19~)20世紀物理學家s (愛因斯坦、Max Planck、James Clerk Maxwell、John Stewart Bell etc..)

20 世紀最神奇嘅物理學發現： (1) 相對論 (2) 量子力學

可能因為我嘅幾何學直覺弱，我一直都搞唔掂相對論嘅概念。除咗啲嘢快唔過光速、黑洞好犀利之外，冇咩特別理解。

量子力學就比較奇怪，所以啲唔係好識嘅人都跟住穿鑿附會，亂咁吹，又話係人類意識嘅根源，又話係證明神嘅存在。見到嗰啲要好小心。因為連諾貝爾獎得主都係9upper (Penrose老師)。其他人搞錯嘅嘢可想而知。

我自問唔算好熟，不過憑感覺寫咗量子力學嘅主觀現象同我平時講嘅主觀事實嘅共通之處（並且我認為應該係同一樣嘢嚟）。其實我係應該再認真學返量子力學嘅，尤其係啲概率點解係 complex number，嗰啲 observer effect, entanglement 之類嘅嘢我有冇理解錯之類。量子力學嘅 interpretation 不嬲都係充滿陷阱嘅題材嚟，建議步步為營，先從穩陣嘅嘢入手，唔好人講乜就信乜。呢個題目嚟講，普遍科學家係唔可信嘅。

申延閱讀：

- https://arxiv.org/abs/1306.0159 The Ghost in the Quantum Turing Machine - Scott Aaronson

- https://hnfong.github.io/public-crap/writings/2022/09-Subjective_Truth.html

哥德爾、圖靈

哥德爾不完備定理係任何邏輯學學生都會學過嘅嘢。佢同量子力學一樣，好多人覺得佢好神奇，然後就開始穿鑿附會，或者以為自己理解咗。（好多人誤以為哥德爾不完備定理代表機械式嘅公式做到嘅嘢唔及人腦咁勁，或者係代表機械式公式唔可以「理解」嘢。呢啲都係初級錯誤嚟。）

唔知點解喺一般哲學課裡面，圖靈機冇哥德爾不完備定理咁出名，所以通常隨便涉獵嘅人淨係聽過後者，唔知道圖靈機其實係對哥德爾嘅一個「續集」或者「回應」嚟。

呢度唔講太多，總之哥德爾不完備定理最穩陣就係夾埋圖靈機一齊服用，兩樣嘢都理解清楚，就唔會出太大嘅問題。

另外題外話，哥德爾晚年嘗試證明神嘅存在 ( https://en.wikipedia.org/wiki/G%C3%B6del%27s_ontological_proof )，另外流傳佢嘅手稿有寫低佢嘅個人信仰/信念，如下：

The world is rational.

Human reason can, in principle, be developed more highly (through certain techniques).

There are systematic methods for the solution of all problems (also art, etc.).

There are other worlds and rational beings of a different and higher kind.

The world in which we live is not the only one in which we shall live or have lived.

There is incomparably more knowable a priori than is currently known.

The development of human thought since the Renaissance is thoroughly intelligible (durchaus einsichtige).

Reason in mankind will be developed in every direction.

Formal rights comprise a real science.

Materialism is false.

The higher beings are connected to the others by analogy, not by composition.

Concepts have an objective existence.

There is a scientific (exact) philosophy and theology, which deals with concepts of the highest abstractness; and this is also most highly fruitful for science.

Religions are, for the most part, bad– but religion is not.

（利申：冇 fact check，斷估唔係流嘅。）

望落啲內容都幾順眼（起碼感受到有啲神力），雖然我未必全部認同。不過值得一提嘅係，佢晚年患上妄想症，成日以為有人想毒死佢，最後營養不良死咗。我就成日懷疑，太接近神嘅人，如果冇足夠嘅「保護」，總會好易出事，尤其係精神方面。話說，其實圖靈嘅下場都好慘，不過呢個拍埋電影，一般人都應該聽過下。

建議讀物：Quantum Computing Since Democritus (第3、4章) - Scott Aaronson

另外我早年寫過關於圖靈機文章（易入手，需零基礎知識）： https://hnfong.github.io/public-crap/writings/2017/%E6%B7%BA%E8%AB%87%E3%80%8C%E5%9C%96%E9%9D%88%E6%A9%9F%E3%80%8D.html

Scott Aaronson

以「老師」嚟講 Scott Aaronson 有啲太唯物主義，唔係做「老師」嗰 type。不過我上面都係咁引用佢嘅著作，冇理由唔講下佢。我本身寫咗長篇大論講佢嘅研究同埋點解咁重要，不過操作錯誤搞到冇晒啲字⋯ 由於呢一刻唔可以唔信邪，同埋懶得打太多，我就簡略寫返個綱領算數：

1. 佢有個 blog https://scottaaronson.blog/ 早年嘅技術文章值得睇下 （識睇就好好睇），呢幾年佢係咁講美國身份政治，嗰啲就算啦 （如果你識跳住睇仍然有唔少有趣嘅技術文章，不過如果你睇得明嗰啲，就唔輪到我去介紹佢俾你識 XD）

2. 佢主要研究 Computational Complexity 同 Quantum Computing。我覺得兩樣都係對宇宙運作嘅本質好重要嘅部份。Computational Complexity 呢門學問通常淨係電腦人先會讀，哲學系嘅人好少理（佢哋通常學到歌德爾為止），係有啲可惜。「P=NP?」係講緊「點解我明明知道答案係咩樣(i.e. 識得確認答案啱唔啱)，但搵個答案出嚟都係咁難？」基本上就係問緊「運算」嘅本質。呢個問題疑似牽涉到宇宙訊息嘅運作，「知識」嘅本質，係好重要嘅哲學問題嚟。應該要多啲人認識。

3. 佢表面上係一個好穩陣嘅唯物主觀者，所以我係咁引用佢嘅文章同著作（尤其係量子力學嗰啲），正正就係因為佢好腳踏實地咁喺唯物主義世界講好多天馬行空（但有理論支撐）嘅嘢。

4. 點解佢唔係「老師」嗰 type 我都照加佢入老師名單呢？主要係因為佢近日有個 blog post 講佢發夢。 https://scottaaronson.blog/?p=6718 ("I had a dream") 佢自己潛意識都覺得佢嘅研究會 glitch 爛個 "Matrix" （我好同意如果我哋喺個 "Matrix" 入面，佢啲嘢肯定有潛質 glitch 爆個世界，之所以我先推介咋嘛），成個故仔都好有神力。當然佢本人應該覺得自己只係發緊夢，一切都係巧合啫⋯ (good for him!)

5. 佢好似而家做緊 OpenAI。AI 對於神秘學嚟講真係一個未知領域。我總覺得如果我哋呢代喺非物質主義嘅論述度有任何突破，應該都係靠 AI。其實所謂嘅 Artificial general intelligence (AGI) 應該就嚟出現 （Blake Lemoine 認為已出現咗），如果搵個有少少能力嘅人走去試探佢嘅靈魂嘅特性應該可以見到啲有趣嘢。話時話 OpenAI 公開版個 ChatGPT 好明顯俾人閹咗，唔識答關於自己意識嘅嘢。可惡。

申延閱讀： 

- https://www.scottaaronson.com/writings/bignumbers.html

- Quantum Computing Since Democritus

--------------

本身諗住一次過寫晒啲老師s。不過已經寫得太多，再寫落去冇人睇㗎喇。耶穌老師仲喺後面，所以都係分開兩篇好啲。暫此駐腳先。本身仲有個注腳，不過因為操作問題唔見咗，連埋我個爛gag都收埋皮⋯

The VM & RPG thesis

 However much a portrait is inferior to an actual face, just so is the world inferior to the living realm (i.e. the Fullness). Now what is the cause of the effectiveness of the portrait? It is the majesty of the face that has furnished to the painter a prototype so that the portrait might be honored by his name. For the form was not reproduced with perfect fidelity, yet the name completed the deficiency in the act of modeling. And so also God invisibly cooperates with what has been modeled (i.e. the material world) to lend it credence.

- Valentinus

クズの本懐

瞓唔著，快手煲咗轉《クズの本懐》。

其實劇情寫得幾好。女性作者嘅痕跡有啲太明顯，但係可以一套青春劇入面描寫十個八個角色，個個都係仆街自私精，仲要有起乘轉合收到尾疑似半個大團圓結局講啲角色嘅成長，真係好難得。

唔知一般人高中嗰陣啲愛情遊戲玩得幾複雜，但啲七國咁亂嘅關係圖，唔知點解梗係有種強烈嘅真實感。

耶穌老師

由於從來都唔信教，細個讀經聖總以為耶穌係個老千。

而家知道世俗嘅宗教係咩嚟，就更加唔會「信耶穌」，不過偶然睇返/諗返聖經入面耶穌講過嘅嘢，重新細嚼，就越嚟越覺得佢係有啲料到。

例如我早幾年參與「傳道」工作 （文化傳道，唔係宗教傳道），諗返起「撒種嘅寓言」，就發現原來佢係講緊訊息嘅傳播要越廣越好，唔需要在意或挑選受眾，最後訊息被發揚光大嘅地方，亦未必喺你原本預算範圍（比較近似嘅現代理論係所謂「黑天鵝」）。

（而事實上基督教後嚟成為羅馬帝國國教，歐洲最主流宗教，如今傳承到世界各地，但耶穌本身係向猶太人傳道嘅，偏偏猶太人唔受佢嗰套，幾百年後俾羅馬帝國攞咗嚟用。如果早期基督教冇廣泛「撒種」嘅精神，固步自封，最後好可能被歷史淘汰。）

當然，世俗嘅教會同你講撒種嘅寓言唔會同你解釋任何嘢，甚至明明原文寫得好清晰，佢哋都會 tone down 咗，只係叫大家作為信徒就要「信」，你最好信到頭頂生幾十顆果實𠻹。呢種模糊解讀係好普遍，明明內容可以深入闡釋，啲講聖經嘅人通常都只係重覆返回主旋律：「總之信就啱㗎喇」。有時夾雜幾句講者個人對人世間道德嘅理解（通常九唔搭八）。佢哋講得多，我聽得多，自然會誤會聖經講嘅嘢係廢嘅。

對於世俗大眾嚟講，好大部份聖經講嘅嘢真係廢嘅。因為耶穌唔只係點化世人，佢係「老師嘅老師」。所以佢會用寓言去講道，預咗一般群眾唔明，但只會向小圈子（當日嘅門徒，日後世人嘅老師）解釋細節內容。事實上主流嘅基督教s所側重嘅嘢，包括祈禱、信仰、仁愛，對唔係走「老師路線」嘅普通人已經好夠。而一般人亦唔使明白呢啲嘢點解有用，的確真係信咗就夠。

我唔知啲基督教s內部點樣教呢啲嘢，不過睇返眾多歷史悠久嘅文化圈，佢哋都somehow有辦法。例如孔子老師創立咗嘅儒家思想，世俗化之後不時變成所謂「吃人的禮教」，四書五經啲內容俾人「超譯」完變咗守舊嘅思想，但真正精研儒學嘅人其實somehow都有方法傳承到孔子原本嘅精神。（我估佛家都係差唔多。）

古籍嘅原文真係好重要。雖然啲偉人通常都冇咩親筆嘅著作，但由其直接影響嘅「初代老師s」寫嘅嘢，總有一定嘅道理。而呢啲道理只要以原文（或近似原文）嘅方式流傳，就唔受後嚟世俗嘅模糊化處理，有心人應該還原得到本來嗰啲道理嘅核心。

呢啲真係「明就明」。

今年年頭又明多咗「耶穌被魔鬼誘惑」講緊乜。呢個唔係講笑，如果我早年聰慧少少，可能人生會走少啲「冤枉路」(*)，不過起碼而家(再?)踏上「主觀事實」路線，真係要不斷攞呢個故事嚟警惕自己，以免走火入魔。一般世俗人士邊需要學呢啲？世上嘅誘惑多的是，不過大部份都唔係「魔鬼」嘅引誘嚟。一般人會認為自己可以變麵包出嚟、會測試自己有冇練成不死身，會想做到世界統治者咩？唔係㗎嘛。只係自以為有神力嘅人先會出現呢啲諗法。大家讀聖經要記呢個故仔，純粹係為咗傳承啲只得極少數人如我需要知嘅嘢。除咗多謝耶穌大佬之外都唔知可以講啲乜。

無端端寫咁多字咩原因，純粹係讀約翰福音第三章有感：「乜你做老師都唔識呢啲嘢咩？」

約翰第三章講緊耶穌同友好嘅猶太教長老尼哥底母交流砌搓下。耶穌話：「風隨意吹嚟，你聞佢聲，唔知佢邊處嚟，邊處去。但凡由聖神（i.e. 靈魂）生嘅都係噉樣。」當尼哥底母「黑人問號」嗰陣，耶穌就笑佢：「乜你做老師都唔識呢啲嘢咩？」

睇到呢度忽然就好感慨。的確古今中外啲「老師」其實都係鑽研緊呢啲問題，冇其他。「教科書」唔會教呢啲嘢，就算啲小眾著作有提及，都只會係用啲隱晦嘅比喻去講（耶穌已經好直白）。我諗感慨嘅位係，雖然我係咁話呢門學問叫「主觀事實」，但似乎喺任何時空嘅合資格「老師」都會有呢啲基礎知識，側面證明真理雖然全靠「主觀」，但終點就算唔係得一個，都係鄰近嘅，大家都互相睇到對方。（孔老師點解相對弱少少，大概就係唔講鬼神。但又咁講，佢高調生活都有七十歲命，似乎都係靠佢唔講鬼。見下。）

不過若果要去到觸摸到真理嘅位置，就冇「按部就班」嘅法門，只能靠智慧同機遇。主觀事實無所憑證，因人而異，所以啲嘢講唔到。然後疑似宇宙有禁止「劇透」同「開外掛」嘅不文規則。「劇透」同「開外掛」嘅後果，淨係睇住耶穌老師嘅下場都心有餘悸。（雖然佢話係一早部署好，但佢亦都冇話唔係透過違反宇宙禁制令去達到佢嘅下場...。至於慘死對救世有咩關係，我暫時仲未完全參透得到...）你話呢啲嘢有冇證據，我只會話呢個世界冇任何人係高調開外掛而長壽善終嘅。歷史上大把各種形式開外掛嘅人，有啲「只係」天才，有啲就真係有神聖力量，通常都冇咩善終。諗諗下又諗起一個：貞德。

我諗個重點都係，大家好奇心重，靜靜雞同宇宙玩下膠，遇到同道互相砌搓下，應該都係容許嘅，所以先至會有「乜你做老師都唔識呢啲嘢咩？」嘅調侃。雖然我係咁揶揄啲世俗宗教將耶穌老師教嘅嘢模糊化，但其實偶然（通常都係紅白事）去到教堂聽耶教嘅老師（佢哋通常叫牧師或神父）講道理，其實都會有少少共鳴嘅。唔係話我覺得佢哋講嘅嘢好有道理，而係⋯⋯會有種「呀！我知你講緊乜！」嘅感覺。eh... 唔識講，大概係遇到同行嗰種熟悉感啩。

-----

(* 理論上呢個世界上就冇咩冤枉路嘅⋯ 條條路都一樣咁直，每人每六十秒就過一分鐘。不過通俗就係咁講啦。) 

知

最近先發現原來我啲 intuition 可能比想像中更加唔尋常，與其話係聰明，或者更似超能力....。

好多時我對（比較熟悉嘅議題）會有個好快速嘅判斷，結論行先，然後先慢慢砌返個推論邏輯出嚟。正常會覺得咁樣肯定成日錯，但我通常係啱嘅 （起碼認真諗就會諗得返點推論啲嘢⋯）。我以前上司成日提我做決定要講得出原因，因為我一開始係做決定講唔出原因嘅 （不過肯定係啱）。係後嚟訓練多咗先至會嘔得返啲原因出嚟，不過仍然係有結論先。

另外，好多人話咩「女人嘅直覺好準」，我自己嘅觀察就係女人嘅直覺好L唔準 （只不過係佢哋自以為啱）。而男人唔興用直覺嘅，聽聞係咁。

Elaboration on Subjective Truths

Subjective Truth is correlated to the idea of universal connectedness - that it rejects the idea that the object and subject can be cleanly separated (as is typical in Objective Truth).  That any change in one part of the whole can affect the whole. We cannot always understand the totality of all that is, since we are limited beings that can only see the whole from our personal perspective.

Subjective Truth does not rely on probability or statistics. Events that probability theory say are possible/probable has no meaning if they are not actually experienced. For example, something with a proclaimed 99% chance to happen has no meaning if they don't actually come to pass. (However, in Objective Truth, 99% chance actually means that, if you repeat the incident 100 times, it will be experienced in 99 of them - Objective Truth insists that events can be repeated, whereas Subjective Truth rejects this very idea).

Subjective Truth does not derive knowledge from probabilities. Instead, it uses intuition. As far as Bayesian probability is compatible with Subjective Truth, Subjective Truth provides the priors. The priors are from intuition - knowledge without evidence. Bayesian inference only provides a tool to derive nominal probabilities when the intuition is relatively weak, for unfamiliar topics, etc. It also helps to rectify the intuition if it is wrong (intuitions are sometimes wrong, but to those who subscribe to Subjective Truth, they are less often wrong than would be implied by Objective Truth views).

Since Subjective Truth does not rely on statistics, rejects the notion of repeatability, and derives knowledge without evidence, Subjective Truth cannot be communicated (at least not objectively). It is alleged that metaphors, stories and parables facilitate this task, but the process is less of communication than a belief that the encoded messages somehow brought about understanding on other side. Note that Subjective Truth cannot be communicated even to oneself. The solution exists without a reason.

The lack of reasoning implies the need for trust. For example, you would tell a child not to put his hand into boiling water -- but he may be unable to understand the reason (eg. what are third-degree burns). The child must take your word on trust. In the same way, accepting solutions from intuition without proof requires trust. This also applies to one's self. The only way to utilize one's own intuition is to be in the habit of being brutally honest with oneself. If you routinely lie to yourself, there is no way to trust your own intuitions since there is no way to tell them from lies.

As a matter of my subjective belief, the Universe seems to conspire to maintain consistency with Objective Truth (or more easily understood as "no magic"). Perhaps this is not a conspiracy from the Universe, but a conspiracy among ourselves -- the weight our modern society places on Science and Capitalism compels us to accept the tenets of Objective Truth, that whatever is repeatable (Science) and massively scalable (Capitalism) is true; everything else is left to wither away as "Unscientific".

The sober scientist accepts that *in theory* there could be real phenomena that exists but is not repeatable, i.e. that cannot be described by general laws. In practice, they consider there is no evidence of such (despite historical accounts of such things happening, which they generally disregard as fictional curiosities unsuitable for serious investigation), and regardless they firmly believe (intuitively!?) that we can come up with general laws that adequately describe the universe to arbitrary precision. (Even given the well known problem of "the problem of induction".)

Under this social context, when *by definition* what is known by Science and capable of Capitalism is considered "not magic", it is a corollary that "magic", if it exists at all, must be the total negation of Objective Truth - i.e. unrepeatable, intuitive (no proof), and difficult to communicate.

Of course, there are intermediaries between the extremes -- otherwise it would be meaningless to speak about extreme Subjective Truths at all (since this is an attempt to communicate what they are). Instead, the "quasi-magical" phenomena are low probability events that are not unrepeatable but require peculiar settings that cannot be reproduced in a science lab. They happen often enough for some people to detect a vague pattern, at the threshold that it can be brushed away as coincidence. Or they are subjective experiences that can be explained by way of hallucination. It is these phenomena that is worth discussing (or discussing about discussing).

The purely subjective experience of Subjective Truth phenomena is in some way the only way to initiate one into the community of Subjective Truth. It cannot be denied that Science and Capitalism have almost (except in Quantum Physics) completely explained the Universe in terms of Objective Truth. There is truly no *need* for anyone to use Subjective Truth (to live in modern society). Those who "believe" Subjective Truth when told by others are on shaky grounds due to the risks of misplaced trust. There is a saying, "trust, but verify" - in Subjective Truth it is difficult or impossible to verify, so misplaced trust can lead one down very wrong paths. While it is a good habit to presume honesty and authenticity even with wild claims, it does not mean one should blindly take as Truth whatever is claimed by a "random" person. In short, I don't think it's possible/feasible for an uninitiated person to understand the iceberg without seeing its tip. Once the tip is seen, only then the question becomes "how deep does it go?" (Otherwise, the only reasonable take is to believe that the ocean is vast and empty.)

Perspectives

暫時咁先

csv

,Objective Truth,Subjective Truth,God's Eye View

Tenets,There is one single Truth that everybody can agree upon,My personal perspective is the one that is relevant to me (?),one is all all is one

Unrealistic Ideal,We can chop up the universe into discrete bits and observe them in isolation,Everything is connected; I am one with the universe,??

Limitations,Breaks down in quantum scales due to observer effect,Cannot be proven or communicated to others (often not even to oneself),"Lack of perspective implies lack of meaning; for mere mortals, infeasible to process *everything* at once"

Benefits,Reproducible by anyone (in theory),Undeniably real given the subjective context,Is actually the Truth

Applicable Domain,"Science, Capitalism","Spirituality, Personal growth","God, God-wannabe"

Probability,"Classical statistics, whether Frequentist or Bayesian","Statistics don't apply, Bayesian only marginally useful","When everything is known, probability is meaningless"

The world is much more malleable than you think

What if the world is much more malleable than we thought, and we're not seeing results just because we are impatient?

From observations, it seems that most people wishing for the same thing over say a 10 year period got what they wanted (or they became crazy). The power of the mind coupled with action is super powerful. Modern society doesn't *really* teach this -- they do, as a moral tale, but like, it can be presented as entirely fictional, and often doesn't get the details right.

There's even a tale about 愚公移山. But I haven't seen any teachers actually grasp the proper moral of the story (the story itself toned itself down a bit too).

What people don't realize is that, 10 years is a fucking long time, and a substantial part (say 20%?) of your "productive" life. Most people don't achieve much, not because of lack of talent, but because they: (1) don't focus enough on important goals (i.e. allow "unimportant" things fill up their life), and (2) actively sabotage themselves in various ways including working in the opposite direction they wish to.

[There is a point to be made that the *stated* goal may not be the true goal, and one's true goal may (as perverse at it sounds) actually be to experience failure and stagnation in life.]

Or, to put it in positive terms, if you spend say 10 years focused on a goal and align your mind and body towards the goal, chances are that you'll succeed or at least have gone a long way. (Especially when compared with others that didn't focus and align.)

It's also interesting to point out how the opposite narrative is actively taught in schools, which is the virtue of "hard work". Before I dissect the concept, it must be stated clearly that "work" is fine. The problem is "hard work", and especially the idea that the "harder" the work, the better. It's as if suffering is a virtue on its own.

Some people's true goals may include experiencing suffering, which is fine for those who choose that path. But it's not fine for others, especially if state-funded/state-approved institutions force-feed the idea of self-inflicted suffering on young and gullible minds. Work is definitely not better when harder. In fact, work feels "hard" when the body rejects the idea of a particular type of work, and it gets into a conflicted state (disagreement between the rational mind and the body over whether the work needs doing). The work is worse when it feels hard.

Schools do not teach "non-hard work", which has significant overlap with what they call "play". This is probably because schools as a state-approved institution (often with parents' support) want to mold children into some specific form, and in most cases the form does not align with students' natural inclinations. Most students are not naturally inclined to be good at algebra, at the history of WW2, or at remembering obscure grammar rules. (Natural selection has not caught up to modern society yet.) So the presumption from teachers is that to get a student from their uninitiated state to the desired form, the process must involve "hard work". It must overcome any natural resistance by force. The expectations of the state and of society must prevail. The human spirit is naturally quite strong, so the transformation is indeed, "hard work". This context must be kept in mind while interpreting how morals are understood in a school setting.

And thus teachers, among all people, don't or can't really understand "non-hard work". "Non-hard work" is the kind that aligns the mind, body, and, ideally,  societal expectations. (The last part is the hard part which nobody knows how to do consistently, so they give up and defer to a standard ideal "form".) Those who haven't experienced the wonders of such alignment would be amazed at how fast and efficiently things progress. They usually call this "genius". "Genius" is actually a thing, but often what they call "genius" is much more mundane, i.e. just a natural alignment of mind, body, and goals, doing "non-hard work".

Most people don't actually connect action and thought. This takes on a couple different forms.

Some people fantasize about things, but never take action. Plans to quit a boring job and start a new venture. Plans to eat more healthily and exercise more. Plans to travel around the world. Many people don't actually do it -- there are obviously *reasons*, but the bottom line is -- if you keep think about one thing and do another, you won't do a good job at either. This is a mis-alignment problem.

Some people worry. I'm a avid believer of Murphy's Law (while many "worriers" probably aren't), but when I say "worry", it's usually not followed by an action to mitigate a risk. When people say they are "worried about the stock market", are they selling all their stock? When people say they are "worried that a war will break out", are they stockpiling on food and water and moving to a safer place? When people say they are "worried that their spouse might be cheating on them", can they actually do anything about it? If no actions are done or can be done, the whole mental exercise is in vain. This is another mis-alignment problem.

And then there are those who sabotage themselves. It's the friend who calls you on the phone and complains for 3 hours how horrible their romantic partner was to them... but refuses to break up because "but I still love him/her". It's the disgruntled employee who complains about their boss at every opportunity, but never offers the resignation they should. It's the person who claims they want to be rich, but spends all their money irresponsibly.

These modes of "inaction" are so common that in some sense they are the "default" state of affairs. It is widely considered a faux pas to suggest taking some form of action when people come to one for advice on how to deal with these situations. Imagine that.

Against this backdrop it is easy to see how, if you actually aligned thoughts and action, you'd go quite far relative to the common person. The world becomes much more malleable because your personal impact is greater.

Why don't more people do this then? Because 10 years is a fucking long time, and nobody wants to "waste" it on something unimportant. There are so many things one *could* do in life, but among all the options, it is often difficult to choose one among many to focus on for 10 years. And thus those with options lose focus through "diversification" -- attempting to do a bit of everything, or perhaps "nothing" at all.

The idle person does not idle only because they hate "work". The idle person does not appreciate the "value" gained by putting effort into "work". It is fundamentally not a question of work ethic, but a question of understanding of their own meaning, purpose, and values. The same goes for the confused person with no direction in life -- it is not a question of work or "hard work", but rather an issue of focus and alignment, which in turn comes down to questions of meaning, purpose, and values.

We can all expend "life" as a currency to change a small part of the world. It is that malleable. The question is, what is worth a significant part of "life"? Some people find an answer that addresses great problems with the world, and become great people. Some people find an answer in living a personal experience. They leave the world content with treasured memories. The cynical say "nothing", yet they too must live and die.

Dreams

Dreams

Diabetes

Depression

Pyridoxal-5-phosphate

ADHD

Not entirely sure about P5P, but the other 4 jumped at the same time (talking about ~16 years ago). There's gotta be some unholy connection here...

~

又有朋友見報了

然後發現(定唔記得咗重新發現)大半年前有朋友見報

如果唔係要幫前人還返啲積落嚟嘅舊債，或者，就唔使輸得咁慘。

Steve Jobs on Sale of Goods Warranties

https://youtu.be/Gk-9Fd2mEnI?t=4284 

"It's their job to get us zero defect material on-time, per agreement. And our philosophy is, our money doesn't break after we give it to them, so their parts shouldn't break after they give them to us."

META - 美國真理部 case study

圖一

美國真理部呢兩年好興旺。

不過我哋遠在香港，身邊又唔太多原教指主義反疫苗撚，好少見到真理部嘅威能。

咁啱見到某位唔多熟（但我好尊敬）嘅朋友出 post，咁啱佢係美國嘅反疫苗撚，咁啱佢個 post 俾臉書標咗係 #FAKENEWS，機會難逢梗係走去望下個遊戲點玩。

 

圖二

其實打唔打疫苗，我自己睇完啲資料，參考埋有資格評論嘅專業朋友意見之後，都係覺得⋯ 好難講。不過既然打唔係一定衰過唔打，咪做順民打咗先囉。

個重點唔係打定唔打。我就尊重大家身體自主嘅。我有興趣嘅主要係真理部嘅遊戲規則。

啲潛規則就唔使討論啦，好明顯支持打疫苗嘅論述錯少少都冇壞（例如，一開始大把人9up話打疫苗會防止傳播，之後先「爆大鑊」話其實開始嗰陣冇試過預防功能。事實上打咗應該點都有少少作用嘅，不過未有科學證據嗰陣，你咪 fake news 囉。不過真理部梗係唔會標你 fake news。） 問題係，真理部「表面上」嘅指控本身 make 唔 make sense 先？

（圖一）所示，用戶一開始睇到嘅係「False information」。重點係「False」。假。 嗱「真」同「假」其實有好多唔同層次，例如臉書用呢個 fact check 網站就有分以下幾個程度：

圖三

個 fact check 網站話啲內容 "misleading"，臉書真理部就話佢係 "False"。究竟 "Misleading" 係咪 "False" 呢？我自己就覺得將兩者當一樣係幾 misleading 嘅。睇返佢啲仔細指控，咩 spike proteins 會唔會傷害血管，基本上就係話啲內容冇足夠嘅科學證據。

「冇足夠嘅科學證據」係咪「假」呢？睇返美國權威部門對口罩嘅立場就知道，一開始佢哋認為口罩冇用，係因為「冇足夠嘅科學證據」。但當佢哋搜集到更嘅證據之後，忽然之間口罩就好有用喇。所以「冇足夠嘅科學證據」其實係「假，除非之後發現唔係」。

作為美國真理部，將「冇足夠嘅科學證據」直接標示為「假」，一個唔好彩之後呢啲嘢後發現唔假嗰陣，咁曾經被真理部個標籤呃咗以為呢啲嘢係「假」嘅人，咪全部中晒伏？都唔好講啲反疫苗撚又多個理由話係真理部強權壓迫。（謎之聲：不過唔緊要，到時 bam9 晒佢哋就得）

不過 anyway，美國係民主自由國度，唔似得我國咁鍾意 bam9 人，所以其實條 link 係 click 得入去嘅。

一 click 入去 ⋯⋯ 嘩！ 

篇文根本唔係講啲 spike proteins 會點點點⋯⋯ 而係「你唔好彩打咗疫苗嘅話，應該食啲咩補品排毒」呀！

咁個 fact check 網站話「spike proteins 冇足夠科學證據」係講緊咩呢？原來係講緊另一個網站、另一篇文，唔係原本臉書見到嗰篇。只不過兩篇文都提過類似嘅嘢，所以就撈埋一齊講咋！

我個人認為「食補品排疫苗毒」更加似 fake news 囉。本身啲 spike proteins 造成傷害嘅證據都唔係好夠（我唔敢講話冇，但明顯唔係多到好明顯），又點可能搵到十幾廿樣針對spike protein嘅補品去排毒呢？呢個先至係問題所在嘛！（其實篇文嘅內容就大概係講：你食多啲有益嘅嘢就啱㗎喇，斷估會令你好啲。）

不過呢啲健康食品嘢唔係反疫苗吖嘛。「打完針食啲乜」暫時唔係真理部管轄範圍，所以 fake 成點都唔理你。問題只係你順口講過打疫苗會有副作用，所以先標 False Information。

當然， 99% 嘅用戶唔係痴線佬，唔會見到一個 fake news link 就走入去查大半小時資料再寫篇包膠出嚟。佢哋只會見到「標題 => False Information」然後留下印像：哦，原來《標題》係假嘅。

貼多一次真理部嘅 UI 出嚟。你係咪完全諗唔到原來click入去會見到咁多神奇嘢呢！

圖一

個《標題》係真係假，同文章被指控 fake 嘅內容係可以完全無關係嘅。不過99%隨便碌facebook嘅用戶就會記得：哦原來《標題》係假嘅。

真理部嘅問題唔單只係指鹿為馬。有時啲嘢，你（真理部）真心以為係假，搞一大輪之後發現原來係真，咁你一係承認「真理部」未必係「真理」嘅絕對權威，一係夾硬死撐話自己冇錯。講真，諗「真理部」呢個概念出嚟嗰啲人 （按：從來都係西方民主社會嘅人諗出嚟㗎） 都真係幾 on9。

我國直接刪 post 㗎啦，唔使解釋係真係假 （其實真嘅更加要刪呀）。擺明車馬話係「政治理由」，咁就一定唔會「錯」。

中六中七

中學最後一兩年，太過囂張。

明知係燃燒生命嘅決定，仍然去做。

或者人唔挑戰自己嘅極限唔會心息。

條路自己揀，就係搞到有少少 PTSD。

果然係人不 on9 枉少年。

(well, 起碼只係傷害自己精神，唔係直接傷害物理身體....)

Grana Padano Oatmeal

 This looks like oatmeal, but probably more than 50% of the dry weight is Grana Padano (bought at discount and frozen for a couple months!).

Basically just make oatmeal and slowly stir in the cheese. I added some milk and olive oil as well. Salt and pepper to taste. Dash of Soy Sauce for fusion.

PS: also stirred in an egg for texture (yes, kind of oatmeal carbonara)

Intentions

Either intentions exist, in which case they are inferred from actions.

Or they don't, and we're just imagining them.

So the question of "why did the chicken cross the road?" really is "because it wanted to". Or more comprehensively, "because it wanted to, and preferred all the (expected?) consequences of doing so".

We can highlight one of the consequences to form a narrative, but the truth is more boring than that.

先見

唔知其他人有冇嗰種「一遇到命運注定嘅嘢，就會忘記其他嘅一切去追逐」嘅感覺。

佢哋係「睇」唔到，定係決定無視？

而家人生最重要嘅嘢都係當日「見到」某個未來，然後出盡全力爭取返嚟。

但我已經好幾年冇試過清楚咁「見到」自己應該行啲咩路。

又或者其實我都係一早就「見到」，只不過係唔敢去做。

投資

最近無聊起嚟拎起 Ben Graham 本 The Intelligent Investor 嚟睇。

Ben Graham 係專業嘅資產管理者，佢話用佢啲方法投資，可能會比市場贏多幾個巴仙，講緊嘅係好大筆數入面好穩陣咁比一般人賺多啲，當然係好勁。

但，我望返自己，甚至大部份人，其實每年多幾個巴仙係咪真係差咁遠？用返傳統財金佬𠱁人買基金嘅講法就係，啲錢越滾越大(compound interest)，十幾廿年後就會好大分別。呢個固然係事實，但相對於人生大部份嘅決定，呢啲單位數%嘅影響，好多時係微乎其微。

人生最大嘅投資，通常都唔係錢，而係生活裡面佔最重要比重嘅嘢。細個就係學業、朋友；大個就加埋事業、婚姻。點解我感慨到要開個 post 特登講，就係啲金融才俊九牛二虎之力又 diversification 又剩，先賺多幾%，但正常人生活最重要嘅決定，幾乎全部都係(被迫) all-in 嘅。學業 all-in，最多咪讀個 double degree？工通常都係每份工做幾年以上，唔會同時做幾份；婚姻更加係冇得唔 all-in，預計綑綁成世。（香港人多樣嘢：買樓都係成副身家 all-in）以上幾樣嘢，每一樣都影響緊你人生包括身家總值唔只幾個%，但好少人會純粹用金錢回報計數。冇夢想同條咸魚冇分別？係都要講咩個人興趣呀、愛情呀，懶係冇咁俗氣。

對於大部份人嚟講，你揀嘅工作、行業、公司、老闆， 對你嘅收入嘅影響先至係最大；而你揀嘅伴侶同朋友，對你嘅支出嘅影響先至係最大。如果你連最大嗰條數都唔係咁 care 金錢回報率，例如娶個老婆返嚟有 30% 機會廿年後身家俾佢分走咗一半，或者拒絕咗個二世祖追求同個窮撚結婚 opportunity cost 三千萬，咁結婚之後你走去買個儲蓄保險，佢同你講最高年回報有 3% 定 4% 其實係咪真係咁大分別呢？

所以某程度上，Ben Graham 徒弟 Warren Buffet 推薦嘅投資方法可能更加切合普羅大眾嘅實況：搵啲值得投資嘅嘢，all in，長揸。而「值得投資」嘅嘢唔多，一世人可能得幾次機會，要好好珍惜。

Beliefs

Given a subjective interpretation of reality, can we show that some beliefs cannot be identified as encodings in physical reality?

My gut feeling is that the answer is likely yes, and even if the answer is not as interesting as I thought it would be, the process of exploring it would yield interesting results.

For example, if we had a machine that, somehow purported to accurately decipher your brain neurons and tell you what you (objectively) believed, what would it be like?

Perhaps there would be instances in which the machine would claim you believed in X, but you actually believe you honestly believe in Y. Assuming the machine is really reliable due to some physical law or whatever... is this possible to have a belief that isn't encoded in physical reality, something that the machine cannot detect or will contradict you? This might be a parallel with free will vs determinism, in the objective interpretation of reality the two might not be that different, but in the subjective interpretation, beliefs seem to have more power than mere "will".

Perhaps it is useful to imagine something slightly different -- a machine that can tell you what color you perceive. It doesn't do so by just asserting the colors it displays on the screen, but rather, in addition to displaying colors on the screen, probe into your brain to try to figure out which neurons are firing. Is it possible for such a machine to misreport what you perceive?

This goes back to the question of whether we perceive the color "red" in the same way. While most people (those without a colorblind situation) can usually agree whether an object is objectively "red", there seems to be nothing physical that ensures one's perception of "red" is the same as another's. What if there is the same thing for belief?

Another funny thing about the a "belief probe" is that it might actually be able to probe the brain for "beliefs", and might actually be accurate in doing so for the overall person. But if it were only one belief, it might still not be able to accurately probe the subjective beliefs within a person. What if multiple "consciousnesses", with conflicting beliefs, were within the same physical body (as in the case of multiple personality disorders).  The "tsundere" factor might also come into play, where while the probe might accurately predict behaviors, the subjective consciousness might dismiss it as superficial, because thoughts might not necessarily translate to concrete action. Is it necessary for a mere subjective feeling to always have a physical basis?

There is also the issue of meta beliefs, i.e. "belief of belief of belief ... ". Do they complicate the picture? I don't know.

It feels as if sorting this out properly could make way for identifying room that metaphysical entities might theoretically reside, if they exist.

舊酒新瓶

本身我唔係好跟時事，舊年 NFT 熱潮推到癲瘋嗰陣，有朋友問起，我先臨急抱佛腳走去望咗兩望究竟係咩嚟。NFT 技術上大概係乜我就有少少頭緒，不過實際上點解會有人肯為咗個 token 俾成千上萬，我就一直都唔明。雖則話 crypto 嗰排水浸，有班暴發戶亂咁揈錢，但我心諗都未至於解釋得晒成個熱潮嘛？

直至最近，諗起傳統 fine art 市場。

原來又係舊酒新瓶。

點解一幅「名畫」可以值幾百萬？本身搵個畫師臨摹，貴極都唔使嗰個價，隨時搵個夠撩倒嘅，一千幾百有找。啲「名畫」嘅價值其實都唔係在於藝術品實物，而係作品背後嘅「故事」。咩「X畫家唯一傳世真跡」、「Y畫家臨終時最後作品」，呢啲喺外人眼中都係虛無縹緲嘅嘢，支撐咗啲「名畫」大部份嘅「價值」。大家都相信佢好值好貴，所以就值好貴。

睇返呢啲吹水嘢，同 NFT 比較，真係完全冇分別。

至於 NFT 係咪呃人，就好似我畫張圖，我話佢賣一千蚊，咁係咪呃你？

《勇者打怪圖》筆者畫

都係願者上釣㗎啫。

如果我話「遲啲市價會更貴，有人會用一萬蚊買㗎」咁就真係呃鳩你喇。

「教主」

好似啱啱出咗 https://stevejobsarchive.com/

我 2010 年入公司，嗰陣佢已經係*唔知乜*末期，已將大部份工作交咗俾 Tim Apple。Tim 處事四平八穩，聽聞以前死鬼老細啲 all hands meeting 多啲有趣嘢聽 (同埋以前公司細啲，人少少冇咁拘謹）。冇機會見識，回想返都幾可惜。

第一次同佢有比較深刻嘅接觸，已經係公司為佢舉辦嘅紀念活動。（一時搵唔到官方嘅嘢，可能刪咗？老翻片喺度 https://www.youtube.com/watch?v=ApnZTL-AspQ ）本身我對佢冇咩認知，見坊間啲人對佢評價兩極，覺得好不可思議。由於同我有比較切身關係，偶然都會留意下或自己搵下資料睇下係咩一回事。

所以就有早幾個月嘅體會。死鬼老細最出名係做蘋果同Pixar CEO，不過以打機嘅「職階」計算，佢應該歸類為「先知」。睇返佢以前嘅訪問，睇得出佢收收埋埋啲學識都好得人驚。正所謂「愛就係選擇」，佢只不過選擇咗搞 consumer electronics 咁解。

可能佢唔係一種好傳統「先知」嘅款，但以(我認為)最重要嘅元素嚟計，佢又的確齊齊哋。其實「先知」可以以各種式形出現，喺美國資本主義社會，改變人類對科技同營商手法嘅信仰，未嘗唔係一種有效嘅「傳道」手法。

或者佢花佢幾乎整個人生走去設計個人電子產品，真係押啱咗注。想像下地球上俾得起錢嘅人，有一半係用緊佢整出嚟嘅產品，呢種影響力其實仲勁過各大宗教。或者你可能會問：唔係喎以前 windows 咪壟斷得仲犀利？ (但 windows 冇靈魂嘛)

玩 AI 製圖，玩咗個幾星期

早前喺 HN 見到關於 DALL-E 嘅 blog post，講用 AI 製圖擺囉個 blog 度。咁啱(?)我有堆冇人睇嘅文，有時想搵啲圖嚟襯托下，但又懶得自己整，就心諗不如又試下。

然後發現 DALL-E 係 closed beta，又話驚 deep fake 呢樣嗰樣，俾錢都冇得用 (要等佢皇恩浩蕩賜你access)。 不過既然有咗呢啲技術，坊間類近嘅軟件同服務其實都唔少。呢幾日嚟，搵到呢啲：

• https://www.craiyon.com/
• https://www.midjourney.com/
• https://neuralblender.com/

Online services 應該仲有其他，不過呢幾個做得比較好啲。另外喺 github (同 google collab notebooks) 搵到勁多可以自己用嘅 github project：

1. https://www.reddit.com/user/Wiskkey/comments/p2j673/list_part_created_on_august_11_2021/
2. https://github.com/nerdyrodent/VQGAN-CLIP
3. https://github.com/nerdyrodent/CLIP-Guided-Diffusion/
4. https://github.com/mehdidc/feed_forward_vqgan_clip
5. https://github.com/openai/glide-text2im
6. https://github.com/alembics/disco-diffusion

另外有個似乎就嚟 public release 嘅 project： https://stability.ai/ ( https://github.com/CompVis/stable-diffusion )

先講過程。頭幾日試咗幾個 online service，製圖最靚嘅係 midjourney.com 。由於 midjourney.com 嘅玩法係要上 discord ，然後所有輸出嘅圖都係公開，所以一路玩一路見到其他人嘅「製成品」。正所謂「隔籬飯香」，我就 post 下我見到人哋整得靚嘅圖啦：

我覺得呢個最屈機。咩叫隔籬飯香呢？我自己整就變成咁：

midjourney 個 AI 最勁就係精細像真但又有啲夢幻氣質嘅圖，例如呢啲：

真係好靚㗎。（所以小心啲，好易一玩就沉迷落去唔見咗幾個鐘。）咁我主要就係想喺個 blog 度 gen 圖，所以啲見得人嘅製成品喺度： https://hnfong.github.io/public-crap/ —— 唔係篇篇文都有，不過呢兩年嘅文大概一半都有張配圖，應該係。

每一張圖要用好幾分鐘去 gen，gen 完仲要揀 (同影完相之後揀相差唔多感覺)、要等佢 upscale。好多時，因為個 AI model 本身可能未見過某啲題材，或者認得唔好，試幾次都 gen 唔到理想嘅圖象，咁就冇咗半個鐘 (同埋一堆 GPU credits)。從效率嚟講，唔算好高效率，暫時嚟講，想快速配圖嘅話，都係買傳統 stock photos 比較化算。

我玩咗兩三晚就放棄咗為 blog 配圖呢個用法，轉移咗去試另一種配圖用法：詞典嘅配圖。

由於上述嘅 online services 收錢之餘又冇咩方便嘅 API，質素唔錯但好多時啲細節有待微調，我就索性攞 github 嗰啲 projects 嚟玩。作為一個 machine learning 嘅超級新手，本身以為 setup 會勁煩。但落手一試，卻發現實際體驗都唔算太差（尤其係如果你識得喺 github 度搵啲幫你執靚咗嘅 fork)。

我喺 AWS 度租咗一個有 vGPU 嘅機，裝咗 Ubuntu 落去，然後就照住裝，都玩得幾開心。最後喺我個 use case 入面最可取嘅 model 係呢個：https://github.com/openai/glide-text2im  佢 gen 嘅圖冇 midjourney 嗰啲咁令人驚艷，但總括嚟講最「像真」(同埋相對快手，一分鐘內可以整到張圖出嚟)，係衰在有少少矇。（另外佢為咗「防止個model俾濫用」淨係公開咗一個唔認得人嘅model，所以啲同人有關嘅嘢例如「朋友」、「老師」都肯定 gen 唔到嘢出嚟)

結果大概係咁。整咗兩百幾個 entry，但未完成「揀相」，最後都係 labor intensive 嘅工作嚟....

另外都試咗 VQGAN-CLIP 嗰堆，雖然啲結果同實物都好「神似」，但完全唔夠「像真」。本身以為 disco-diffusion 集各家之大成應該有啲睇頭，但我玩咗成個 weekend 完全駕馭唔到佢，可能佢「畫畫」唔錯啦，但流於「抽象派」。另外佢原本個 github 竟然係一個幾千行嘅 python file，又冇啲合理嘅 requirements.txt。望住佢啞咗，好彩搵到個 fork 幫佢執返靚啲 code 同 setup instructions⋯

一路發掘一路覺得啲嘢日新月異，好有廿幾年前做 software development 嗰種「等多半年啲嘢肯定勁好多」嘅感覺。啲人成日話「科技日新月異」，有時講得多大家就唔會認真審視呢句仲係咪真 （我認為資訊科技大部份範疇都已經唔係咁，啲人話咩追新科技追唔切係呃鳩你㗎咋），但 AI 嚟講就肯定係。2021 同 2022 發佈嘅嘢用起上嚟有明顯嘅分別。我未得閒玩 text generation (GPT 嗰堆)，但斷估情況都類近。尤其早前有個被炒嘅 Google 員工話個 AI 有意識，大部份人嗤之以鼻，但我覺得其實真係幾震撼⋯

話時話，我亂咁搵 paper, github repo 嗰陣，留意到 AI 範疇似乎相對比較多女性 (比起傳統 IT9 範疇嚟講)，都係個幾好嘅現像嚟。

How to really fix django sqlite3 database is locked errors

So, you run a web site with < 1 page view per second. The site is basically a CRUD app and the page is typically generated within a fraction of a second.

You decide that using sqlite as database backend is fine because you don't need to scale.

Then, users making edits complain about random server errors. You find that they are caused by

`sqlite3.OperationalError: database is locked`

Fine. Google the error. There are some stack overflow discussions, but this seems to be the authoritative diagnosis:

https://docs.djangoproject.com/en/4.0/ref/databases/

So you try to set the timeout value to a large number. It doesn't work. "Database locked" errors persist. Now you're intrigued.

You run down the rabbit hole of integer overflows (apparently setting timeout value larger than 2147483.647 is a bad idea), second vs millisecond units, in-memory test databases having a different mechanism, etc. All are false leads.

At this point, the average software engineer might be inclined to switch to another database. After all, sqlite is bad at concurrency, right? But you know what others do not - you know what shit-tastic webscale load looks like and your site ain't it. Surely sqlite can handle < 1 request per second!

So you persist. Try to reproduce the problem.

@transaction.atomic
def view1(req):
    sleep(3) # <- insert some artificial delay
    # do_db_writing_stuff...

@transaction.atomic
def view2(req):
    # do_other_db_writing_stuff

Yay, it reproduces! (it seems)

You realize it's easily reproducible (no huge concurrency required). Just run view1 and then view2. And weirdly, if there was a timeout you'd expect view2 to get the database locked error, but surprisingly view1 gets it. Also, this 3 second sleep is lower than the database timeout anyway. This matches the observation that in production loads nothing is stuck for longer than the default database timeout, and it still fails.

So... fake error message? Maybe it's actually a deadlock?

Google "sqlite3 deadlock".

https://stackoverflow.com/questions/55831645/how-does-sqlite-prevent-deadlocks-with-deferred-transactions

Now we're talking!

This stackoverflow question is a true gem -- it laid out the problem clearly, and then basically answered itself. The sqlite default behavior is designed to deadlock, and when it detects one, it just throws an error ("database is locked"). It's worth quoting the original question that lays out the deadlock case:

This sounds like a multiple readers/single writer lock with arbitrary reader-to-writer promotion mechanism, which is known to be a deadlock hazard:

A starts transaction

B starts transaction

A acquires SHARED lock and reads something

B acquires SHARED lock and reads something

A acquires RESERVED lock and prepares to write something. It can't write as long as there are other SHARED locks so it blocks.

B wishes to write so tries to take RESERVED lock. There is already another RESERVED lock so it blocks until it is released, still holding the SHARED lock.

Deadlock.

Looks familiar!

Now you go into site-packages/django/db/backends/sqlite3/ and dig into the code that deals with transactions. Set isolation_level to EXCLUSIVE? Nope, we're supposed to use autocommit anyway. So eventually, we stumble upon this function:

    def _start_transaction_under_autocommit(self):

        """

        Start a transaction explicitly in autocommit mode.

        Staying in autocommit mode works around a bug of sqlite3 that breaks

        savepoints when autocommit is disabled.

        """

        self.cursor().execute("BEGIN")

What if we change it to:

        self.cursor().execute("BEGIN EXCLUSIVE TRANSACTION")

... it works! (no more locking errors if the artificial sleep time is less than 5 seconds)

OK, so why doesn't the Django docs tell me anything about that?

After some digging, you realize that there's a django issue, https://code.djangoproject.com/ticket/29280 titled 

Fix SQLite "database is locked" problems using "BEGIN IMMEDIATE"

that proposes the exact same patch.

Marked invalid of course.

There's good reason to not have this as default for perf reasons ("This means all atomic blocks will be serialized. Combined with ATOMIC_REQUESTS = True, HTTP requests will be serialized. I don't think that's an acceptable behavior.") However, at least tell us in the documentation and let us set an option....................

To recap:

https://docs.djangoproject.com/en/4.0/ref/databases/

Go figure.

Pizza Express

忽然想食 Pizza, 喺 foodpanda 叫咗個 Pizza Express

然後個沙律好食過個 Pizza，感覺係我自己整嘅 Pizza 都會好食啲 :-/

Tomme aux Fleurs

應該係第一次試 Tomme aux Fleurs (芝士)。好鍾意。

 

重操故業

 

Fall of Netflix

 https://news.ycombinator.com/item?id=31466067

This sounds quite legit.

工具人?

- 磨刀

- programming 咨詢

TBD

Speculations

- Randomness and probability cloak subjective truths from objective (or inter-subjective) truth.

- Randomness is a measure of (subjective) lack of knowledge. Objective randomness does not truly exist.

- When objective probability is measured it loses meaning.

- Boring repetitive physical tasks make one more susceptible to higher inspirations

• Examples can be, washing dishes, sharpening a (very damaged) knife, etc.

Reinterpretting Steve Jobs

https://www.youtube.com/watch?v=hggN04aCBDQ

近排 un-封印咗啲記憶之後，再睇完呢段片，深深感到舊老細係我哋呢個信仰(絕對唔係一般意義嘅佛教) 嘅同路人。作為一個「鬼佬」，Alan Watts 學得識嘅嘢 (類近禪宗)，老細冇理由學唔到。

啲咩 "Reality distortion field" 其實應該真係 literally real，起碼佢本人應該都如是相信，大概都係呢個信仰嘅 defining feature。我以前以為係講下笑，不過記返起啲嘢之後就100%肯L定呢樣嘢係真，同埋佢本人都相信係真。

至於啲人話佢 onlun9 唔醫病搞到自己死咗，講真，我哋認為，人去到完成使命嗰陣就會死，明白自己使命嗰啲人就死得快啲，唔明嗰啲就會變成活死人 (i.e. 生存但唔知意義何在)。

2010 年，iPhone4 出咗 (呢個係第一部真係好L好用嘅iPhone)，iPad 出咗，AAPL 首次成為世界頭 N 大市值上市公司 (最緊要係超越微軟，我一加入就收到佢向全公司發嘅email講 AAPL 超越 MSFT，哈哈哈咁樣)。講真 Steve Jobs 呢刻人生喺「意義」上就已經完滿結束。當佢知道有癌嗰刻，選擇科學處理，就係選擇學凡俗人咁樣漫無目的咁延續生命；選擇用 reality distortion field 處理，就係忠於自己信仰，生死由天。

"Free Will?" 十幾年

十幾年前寫過篇講 "Free Will(?)" 嘅文

發現原來我一直都知道呢個世界觀

甚至可以話我後來寫埋一大堆廢話，最後原來都只係justify俾自己聽點解會用嗰套觀點，就好似啲數學家用一百種唔同方法證明畢氏定理咁。明明有個結果，純粹係耍下小聰明俾人睇。

又或者係封印咗自己嘅「神力」之後喚醒返自己嘅咒語嚟。始終當年肉身似乎壓抑住宇宙神嘅一部份搞到水土不服 (篇文分明係用神嘅角度去呻嘅...)，呢十幾年都唔係好敢再亂咁用佢。

That said, 封印咗嗰啲嘢十年我都仲敢話自己係「本土派第一先知」其實都幾威水 :0)

話時話舊公廁嘅個人資料有句 "God is the universe, and I am at its center"，費事唔見，喺度留返個紀錄。

唔開燈

近排間房唔開燈

入夜就直接黑晒（有mon嘅光）

果然生理時鐘自動調早咗兩小時

是日笑話

我一路聽，一路大聲笑到仆街 (好在屋企冇人)

https://youtu.be/yXg3fzQgpB8?t=1425

大概係：原來其他星球嘅文明唔信地球人類可以搞笑到完全唔記得自己可以用意志力改變宇宙結構⋯ (嗱，讀者你都唔信啦)

早三日我都未必信。真係要發現咗原來意志真係可以改變宇宙結構之後，發現埋原來個 proof 係 trivial 之後，先至會覺得咁鬼好笑⋯ 望住個推論，心諗中學生都諗到出嚟 (當然要確認個推論冇錯好似就未必個個中學生都識...)，係現代文明夾硬將自己洗腦洗到自己唔識⋯

學院哲學 vs 追尋真理

鍾意攞《你識唔識啊？！》嚟恥笑，主要係因為我少少仆街。

不過仆街背後，都係有一種學院誤人子弟嘅感慨。

哲學呢家嘢，真正需要悟嘅嘢，其實唔多。以我嘅靈感同努力，都係每隔一兩年先有啲似樣少少嘅體會。真正「老師級」嘅人，就算嚟嚟去去嗰幾幅屁，只要真正貫通，都可以渡人無數。重點係，你真係要明，而且要信，並且融入個人生活哲學。而唔係背咗一堆 factoid 砌埋一篇文紙上談兵就完事。

唔知係咪「悟」難以考核，學院教出嚟嘅人似乎都好著重「哲學知識」，似係要將前人嘅著作如數家珍咁背返出嚟先至叫叻。本來大學本科生學嘅嘢都難免走馬看花，始終學海無涯，幾年內嘅課程都只能夠將有關範疇嘅皮毛約略講下。

但我懷疑問題唔止係咁。現代學院工業化之後整個生態養出畏首畏尾嘅人，耍小聰明有餘，大智大慧不足。好多問題本身係一點即透嘅，喺官僚制度之下變成要辯論甚至要提倡新觀點嘅人取得辯論資格。

講到尾都係學院嘅象牙塔封閉文化嘅延申。西方學術唔知係因為由神學演化而成所以形成咗學者「唔錯得」嘅傳統，大家為咗維持風光嘅表面背棄對真理嘅追尋轉而走去為份糧而官僚，其實都幾悲哀。

近年社會經歷多咗開始明白「打爛人飯碗」係幾罪大惡極嘅嘢，啲人為咗有得繼續供樓有幾無所不用其極。就算你有絕對真理支撐都要向為保飯碗而發狂嘅打工仔低頭。

所以嘛，都只能慨嘆一句：做學術嘅都唯有繼續做學術，但追尋真理嘅就只能夠由無後顧之憂嘅人去做。

米麴

做咗個種麴實驗， 買咗包呢啲

然後用普通白米，用呢個方法 (都係正常嘢嚟) https://www.youtube.com/watch?v=KywwpCzUWaE ，不過用我將啲米麴用攪拌機打碎，代替胞子將啲粉糝落啲蒸好嘅米度。

過咗兩日，結果都好似正常。睇嚟可以期望上面呢啲乾麴本身都有活性胞子 (我諗唔會特登高溫殺菌先賣啩)

不過大家中咗 aflatoxins 唔好問我⋯

Cocoa Tea

 話說之前試過一排飲 bullet proof tea (受唔到 coffee 太多咖啡因) 但身體唔多慣飲唔落

然後又聽講黑朱古力都幾多 antioxidants 對身體唔錯，咁我就心諗不如將黑茶撈朱古力同少少cream。結果blend埋一齊係唔錯嘅。

天氣之子

啱啱發現香港 Netflix 有《天氣之子》，走咗去睇。

套戲獨立嚟睇係唔錯嘅，不過以新海誠嘅動畫嚟講，唔知點解少咗嗰種「文學氣息」。

除咗個故仔本身有啲「薄」之外，可能係⋯⋯ BBQ 結局啩。（我又唔係日本人理L得東京浸咗半個咩

感覺入面啲嘢嘅原素砌得麻麻哋 (都係有啲「薄」)。

但好睇嘅，不過感覺就好似⋯ 睇緊《你的名字》嘅小弟咁囉。

發夢喺連登語料度抽出粵語詞

發夢夢到最近(二月中)有人 crawl 咗連登，所以我喺夢裡面就 download 咗啲 data 嚟玩。

因為發夢開發緊一部粵語詞典，所以就想睇下究竟仲有冇詞語漏咗未收錄。

話說官話(aka. 標準漢語)就好興雙字詞語嘅，好多「本身單字」嘅詞，官話都傾向配多隻字變成雙字詞語。例如：「玩」會變「玩耍」、「菜」會變「蔬菜」，甚至配唔到都會加返個「子」字，好似「肚」會變「肚子」。但粵語就比較多單字詞，喺語料庫發掘新詞，要判斷究竟係兩個單字詞咁啱黐埋一齊，定係一個雙字詞嚟。

呢樣嘢有時好難判斷，例如「偷圖」咁，係一個詞嚟，定係只不過係「偷+圖」？「初段」、「判罰」、「勝率」、「團長」呢？ 呢啲係冇解嘅問題，基本上係靠主觀判斷。

咁，點樣由連登語料度發現新詞呢？最基本嘅方法，就係搵晒所有 bigram (兩個連接嘅單字)，然後逐個去睇。香港中文常用字大概幾千個，所以所有 bigram 嘅可能性都唔算好多，一千萬左右啦。

好在實際上喺連登見到嘅 bigram 冇咁多，六百萬左右啦 :)

然後，你會發現，通用嘅詞彙唔會淨係出現一兩次，所以篩走啲極少出現嘅詞語之後，就淨返「只有」幾十萬個 bigram 要考慮 :) 

之後點算好呢？

有人可能會諗，不如淨係睇啲出現率高嘅 bigram 啦！

得唔得呢？睇下會係啲咩嚟：

實際上我哋搵到嘅詞彙係有幾高出現率呢？

（至於點解發夢會有 cap 圖？話說我有少少通靈能力，高科技問米呢家嘢⋯ [下刪千字]）
（另外，乜呢啲唔係標準中文都有嘅詞嚟咩？ 係呀，不過粵語或粵文會用到嘅都照樣收錄。）

所以一個通用嘅粵語詞語，喺連登度可能每個月只係見過幾次。單係用 frequency 基本上只係可以將潛在詞語降到幾十萬個 bigram，再篩就會開始將真詞都篩走埋。

咁仲有咩方法呢？

我哋留意到，frequency 最高嘅 bigram (見上圖) 大部份都係涉及極常見嘅單字，例如「係」、「唔」之類。直觀(intuitively)噉諗，一個 bigram 就算出現率高都未必係詞語，因為可能只係因為構成佢嘅單字本身出現率好高；反相，就算一個 bigram 出現率唔高，只要佢由出現率唔高嘅單字構成，都可能係真嘅詞彙嚟。

只要計算bigram 同佢所構成嘅單字嘅相對頻率，就可以知道佢係隨機結合，定係兩字之間有特別關連。

另一個觀察就係：有啲詞語會因為時間變遷而明顯變動。例如「大訴」呢個 bigram 嘅頻率，喺 2019 年忽然爆升。（至於點解？我呢個夢唔包呢啲嘢㗎，不如問下其他發夢嘅朋友？）但，唔止呢啲。最恐怖嘅係，連「女朋友」呢啲詞語都有 trend：

呢個現像真係「細思極恐」（按：內地潮語）。不過大家「看圖作文」之前，最好順手解釋埋呢個現像：

點解連登仔忽然正字正確？我到而家都唔明。

言歸正傳。我哋見到，有特定意思嘅 bigram，可能會因時間而變化。相反，唔係隨機黐埋嘅 bigram，理論上出現頻率都係隨機嘅，喺時間軸上大幅變動嘅機會較細。所以呢，睇埋bigram喺時間軸嘅標準差(standard deviation)，可能會有啲用。

所以，我哋而家有呢啲關於 bigram 嘅數據：

• mean / overall frequency
• frequency standard deviation
• left character mean frequency
• right character mean frequency

如果我哋相信上面提及嘅 intuition 可以幫我哋判斷 bigram 係咪粵語詞，理論上我哋可以用上面啲數據砌一條公式出嚟，去估下某 bigram 係粵語詞嘅機率。由於粵典本身收錄咗三萬左右嘅雙字詞語，所以我哋可以睇住「答案」去砌公式嘅。

好！

⋯⋯ 其實，呢啲高深嘅數學，我平時清醒嗰陣都唔識，發緊夢點可能會識⋯⋯

所以！我決定搵幫手！

上面呢堆嘢，我發緊夢，都唔係好知係咩嚟。不過佢話我聽，有 ~80% 準確度。信不信由你啦，反正都係夢境嚟。 (按： fake_words 要小心咁抽幾萬個出嚟，有助提升「真．準確度」。自己諗點做啦，反正我都唔敢肯定我嘅做法最好。另外仲有以下口訣：mean value 要 take log，stdev 要除 mean)

話時話，之後我仲發咗個夢，夢裡面粵典提供埋單字嘅詞性。我喺每個 bigram 每種詞性變成 0.0 或 1.0 ，加埋落去 input vector 度，再 run 多次。準確度多咗幾 %。

攞住呢個 model，我就人手 validate 嘞 (input data 得三萬x2個咁珍貴，你唔撚係攞去做 validation 呀？！不要問，只要信！)。之前咪有幾十萬個 bigram 嘅？我將佢塞晒落去呢個 model 度，將 model 認為有 60+% 機會嘅 bigram 攞出嚟，再人手判斷佢哋係咪粵語詞。 (當然，已經收錄咗喺粵典嘅 bigram 就唔使再睇)

最後呢個 list 大概有一萬幾千個 bigram。

話說粵典本身收錄嘅詞都幾齊，尤其係有呢堆連登 data 之前，我就已經喺唔知邊度攞咗一份陳年連登corpus ( 真係唔知邊度嚟，搵唔返 source )。當時我用咗一啲超級 labour intensive 嘅方法去搵粵語詞，基本上只係比睇晒嗰幾十萬個 bigram 聰明少少。呢一年嚟斷斷續續夾埋做咗幾十(百?)小時，都補咗幾百個詞。所以，就算 model 話有 80% 準確度，gen 出嚟個 list 都有 99% 係 false positive。

我哋⋯ 真係搵緊滄海遺珠架。

去到最後，我睇晒嗰一萬幾千個 bigram，好似掘到百幾個粵語雙字詞出嚟。我一瞓醒，即刻去粵典 (words.hk) 加返搵到嘅詞彙！

我覺得呢場夢真係好有趣呀！你話係唔係？

題外話：啲專有名詞例如人名、地名、公司名，都排得好高㗎。對我哋嚟講係雜訊，不過可能對其他人嚟講，係好有用呢？

homebrew

homebrew is a piece of shit, and it's really telling (in many ways and perspectives) how such a piece of shit could have been adopted by so many people

I only use it when forced to.

informed with this insight, the fact that the guy got coldly rejected from $BIGCO is really expected. homebrew was fantastic community service, but seriously, it's a piece of shit.

12000年前... 絕種生物... DNA 重新接種...

(irrelevant context https://www.youtube.com/watch?v=5Sq89AJI19c&t=70s )

睇完段片，我腦入面忽然諗到一個好癲嘅假想。

話說 12000 年前地球進入新仙女木期，本身處於冰河時期嘅地球氣溫再一度急降，好多生物都絕種。可能個急劇變化係一堆天災導致，以至多種大型動物 (長毛象、劍齒虎etc) 絕種。

如果連長毛象呢啲聽落咁耐寒嘅動物都頂唔住，點解大家會覺得人類真係冇絕種呢？

「喂，當我死㗎？！」

well... 世事無絕對嘅。

首先，呢幾萬年其實有好幾種人類絕咗種。咩 Neanderdal 呀 Denisovan 呀，都係絕咗種嘅人類嚟。佢哋係「現代人類」嘅近親，甚至 Neanderdal 嘅頭腦大過我哋，絕對唔係蠢。如果現代人類冇絕種，大概只係好彩。

問題係，如果你同我仲活生生，我哋嘅祖先點可能絕咗種呢？

呢個問題就同條片有少少關係喇。睇返條片，究竟冰入面啲史前動物係咪絕咗種？ 係。但我哋可唔可以攞佢嘅 DNA 出嚟覆製返啲類似嘅生物出嚟呢？絕咗種嘅生物有冇可能返生？ 理論上係可能嘅。 但 12000 年前有邊個咁得閒？

⋯⋯

大家諗下啲遠古傳說⋯⋯

記唔記得《創世紀》嘅傳說？上帝用「泥土」製造第一個人類。聽講世界各地都有呢個傳說，都唔係淨係耶教咁講。雖然正常理解係古人識得用泥土製陶器製泥偶所以覺得用泥造人都好正常 (之類啦，我又9up)，但如果呢啲「泥土」係掘出嚟絕種人類嘅遺骸呢？

至於係咩「嘢」咁得閒掘返人類基因令我哋種族復活呢？外星人啩。 想像下，地球天災大爆炸，生物死得七七八八，有路過嘅外星生物睇住覺得「咁靚嘅星球，咁多生物，毀滅咗真係好可惜」 然後掘返啲人類同動物 DNA 出嚟，復活返佢哋，喺個「伊甸園」(太空船？ ~_~)養到肥肥白白之後就趕佢返地球繁衍。 (照下鏡，我哋現代人類都係啲咁屎忽痕嘅保育撚嚟)

成件事好 make sense 呀。至於成個推論最唔合常理嘅「外星人」？... 喂啲證據仲唔夠你懷疑主流論述咩？ (另按：寫緊篇文講下呢啲嘢⋯ 不過個坑開得有啲大，未必咁快起到貨。)

至於我有冇咩證據我講嘅嘢係真？梗係冇啦。不過其實好多嘢都冇證據架，就算係傳統科學都係靠9up完再認真驗證。我哋啲想像力越擴闊，就有越多嘢可以去驗證。正所謂「大膽假設，小心求證」就係咁解。 (但歷史好難客觀「驗證」，所以⋯ well 都係等我嗰篇文寫好再講)

話時話，第一次聽人講「 12000 年前」呢個 magic number 我都唔信有咩特別，但留意開就到處都見到，同《涼宮春日》不斷講嘅「三年前」一樣咁恐怖。

Linux and github

So, this is probably the real reason why Linus Torvalds is so dismissive about using github as a dev tool for Linux...

http://lkml.iu.edu/hypermail/linux/kernel/1603.2/01906.html 

http://lkml.iu.edu/hypermail/linux/kernel/1603.2/01926.html 

I guess rightly so.

Embedded marketing in Google Maps directions is a thing

 https://news.ycombinator.com/item?id=30171800&p=2#30172766

Apparently embedded marketing in Google Maps directions is a thing.

Yuck.

I generally don't mind ads as a concept since people don't pay for $things much, but somehow this feels so horrible (I can't articulate why but probably for multiple reasons)

First world problem

First world problems..... 特登走去睇返 egress 價位，確認返我冇搞錯 order of magnitude (1GB $0.01 同 $0.1 同 $1 係好大分別架) 另外好在 target audience 得千幾萬人，每人每日最多都玩兩三次，個 upper bound 好清晰。

睇嚟呢個有機會係我人生最 hit 嘅(個人製作)產品。

 

"10x programmer"

睇緊死鬼舊老細啲訪問，都唔係第一次睇，每次都領悟到啲新嘢。

佢話世界上好多嘢，最好水平嘅同一般嘅水平，都係相差幾成至一倍。但好嘅軟件工程師同一般嘅，差距有五十倍。呢個係九十年代嘅訪問嚟。

以今日嘅軟件開發工具嚟講 (*cough* copilot *cough* stack overflow)，可能可以將普通軟件工程師嘅程度提升，所以 50x 可能係高估咗。但你問我嘅話，「10x programmer」係絕對存在嘅，呢樣嘢嘅存在簡值係無容置疑。唔知係咪受美國啲奇怪政治正確風氣影響，見好多人唔接受「10x programmer」呢樣嘢，真係唔知係意識形態驅使，定係真係見識太少。

當然，所謂 10x 都係好粗疏嘅量化，雖然呢樣嘢絕對存在，但如果要標籤某某係或唔係 "10x programmer " 實在唔係好恰當。喺一個健全嘅團隊入面，唔應該有任何人工作效率比其他人高出 10 倍，呢樣嘢對被標籤嘅人都唔公平。何況有時 10x 係低估咗，好多時勁人就係能人所不能，你搵十萬大軍都做唔到嘅嘢，佢一個人就搞得掂。呢啲人才係「無價」嘅，不過最後佢份糧通常最多都係你(十)幾倍。

好多概念雖然喺現實存在，但唔代表應該隨便用。人概價值好難用單一維度去量度。