Saturday, 18 July 2026

九龍書店雜憶

00年代末離開香港之前,我在油麻地讀中學。那時離學校最近的書店,應該是豉油街上,彌敦道和西洋菜街之間,行人隧道口旁曾有過的商務。店不大,但也有兩層。記憶中,我沒有在那裏買過書(畢竟家住沙田,有更大的商務),但打書釘是有的。

從豉油街商務出來,沿着那時還是行人專用區的西洋菜街北行兩個街口,便來到與奶路臣街交界,旺角地鐵站E2出口對出,當年相約時會以Body Shop、美歐皮鞋作記認的十字路口。這裏有好幾間我常去的書店。

在十字路口南側,鐳射(現蘇寧)對面樓上,沿着有綠色扶手的水磨石樓梯拾級而上,便是田園書屋。書店不大,格局也多年不變:進門右邊是收銀枱,正前方是旅遊書(包括風靡一時的閃閃書),轉左則是兩條走廊:後面那條放的是本地政治、社會、歷史;前面那條則以文學為主,但也有一整個書櫃放着似乎永遠都在的台灣學生書局版古籍(帶有現代白話文語譯,注音字母那種)。台版書七折、港版書八折的實惠定價,也是多年沒變。

十字路口的東側,曾經有樓高兩層,臨街營業,專賣新版教科書的漢榮書局。二樓售賣外國教科書,種類不少;在網購不如現在方便、資訊沒那麼流通的00年代,自有其市場定位。當年準備英國A-levels,便是在那裏買參考書。漢榮門口左邊的巷口有買賣二手教科書的漢記書局,我也曾經光顧。

漢記旁邊的舊樓樓上有樂文書店、學津書店;兩者看似入口不同,但記憶中是有鋪着紙皮石、牆壁油漆有些許剝落的走廊相連。我好像光顧過樂文,但沒有在學津買過書。

記得某天放學後,發現十字路口的北側、Body Shop斜對面某幢舊樓門口有一個藍底白字的招牌,上面寫着「序言書室,人文及社科書籍」,便沿着樓梯走到二樓,再搭電梯到7樓,然後再走半層樓梯一探究竟。是適合打書釘的書店:有沙發、有桌椅、有洗手間,還有免費Wi-Fi。左邊是中文書,右邊是英文書,選書甚合我心,又會舉辦討論會,也有讓讀者寄賣二手書的書架。我當即付了二十元年費做會員。後來序言有了貓,又沒有了貓。我的會籍時斷時續。差不多二十年便過去了。

序言樓下,曾經有專賣二手書的梅馨書舍。書刊雜誌堆放頗為隨意,但偶爾也會找到合心意的滄海遺珠。兩間書店雖然共用電梯內的「7」字按鈕,聽說東主政見卻稍有不同。不過,對讀者來說,兩家選書取向相異的書店在樓上樓下,也有方便擴闊閱讀視野的好處。有時序言人多,座位供不應求,我便會去梅馨的沙發打書釘。

從十字路口往北行,在亞皆老街街口(也就是曾經有售賣牛仔褲的Apple Shop那個路口)之前,有名為旺角城市中心的商業大廈,裏面曾經有好幾間書局。有主要販賣港台出版繁體書的榆林書店,也有00年代盛極一時,主打較廉宜內地出版簡體書的書店:文星圖書公司、博學軒、新思維書店,等等。記得我的第一本《挪威的森林》,便是在那裏買的林少華簡體譯本。當時覺得連「壽司」也要加上腳註,向讀者解釋含意,是頗為有趣的閱讀經驗。之後人民幣升值,去深圳越來越方便,加上港人閱讀偏好改變,這些簡體字書店也漸漸沒落。

疫情期間滯留香港兩個多月,發現旺角以北的深水埗、太子也開了好幾間書店。大南街有龐一鳴開的一拳書館,選書、鋪排、舉辦活動都甚有心思。該店也賣繪本。某年在一拳留意到不少本地出版的兒童繪本都由「木棉樹」出版,才知道小學時曾經訂閱的《木棉樹》兒童文學雜誌停刊後,負責人還在繼續做出版工作。自此以後,每次回港,我去一拳都會買繪本,送給有兒女的朋友作伴手禮。此外,一拳附近有舊居(沙田第一城)前區議員開的獵人書店,太子地鐵站附近有留下書舍,都是很有意思的店。可惜在這兩間店,我只是分別買過筆和T恤(我不是書店店員)。

從旺角往南走,在西洋菜街、登打士街交界,兆萬對面,曾經有中南圖書文具(後來搬到油麻地)。該店主打文具,款式相當齊全,中學時會在那裏買活頁紙、cue cards。

油麻地天后廟對面有中華書局,樓高三層。頂層面積頗大,選書側重文史類,有很多香港別處少見的內地作者作品。記憶中,中學時代在那裏買過茅海建《天朝的崩潰》。上次回港,也在那裏買了同一作者的《苦命天子》。駿發花園有百老匯電影中心,裏面的 kubrick 售賣不少電影、文藝書籍,頗為有名。

尖沙咀商務搬過幾次,在我讀書的年代是在星光行,後來搬到美麗華,現在又搬到九龍清真寺旁。都是面積頗大的店,中英文書的類別都多。商務離開後,星光行開過誠品,可以從店內透過玻璃窗看到天星碼頭,環境舒適,可惜已經結業。海港城以前有PageOne,主要售賣英文書,價格偏貴,當年只會瀏覽,甚少光顧。樂道曾經有售賣英文書的老字號辰衝圖書( Swindon )。做學生時,總覺得那裏氣氛過於嚴肅,有被拒於門外的感覺。後來在它結業前去過一次,在零落的書架上見過一本英文的文言文教材,可是沒有買下來。

上次離開香港,已經是年半前的事了。不知道下次到香港,又會去哪些書店呢?

序言書室、梅馨書舍所在大廈入口處,自攝於2011年9月11日

西洋菜南街、亞皆老街交界處,自攝於2004年6月12日

西洋菜南街、奶路臣街交界處,自攝於2004年12月13日

田園書屋招牌,自攝於2004年12月13日

一拳書館內售賣的木棉樹繪本,自攝於2024年12月26日

獵人書店,自攝於2023年12月30日

尖沙咀商務印書館英文部推介書籍,自攝於2024年12月29日

臉書連結

Friday, 12 June 2026

The value(s) of AI models

This started as a couple of comments under an ex-colleague's musings on AI on Linkedin, but I thought it was worth setting out in a standalone post. So here are my current tuppence on the value(s) of AI models: 

(The future is inherently uncertain. All views my own. Etc.)
  1. There is a real chance that many investments in AI models will turn out to be NPV-negative for investors. 
  2. This is not necessarily a bad thing socially. If none of the frontier labs achieve the dominant market position that, say, Google achieved in search, the result could be cheap, widely available AI capabilities as profits are competed away and consumer surplus is created. 
  3. Many leading Chinese models, as well as Meta's Llama and Mistral's models, are open-weights. Everyone with sufficient hardware can run them locally. Everyone can, in principle, tailor them according to their needs. For example, last year, Perplexity released R1-1776, an adapted version of DeepSeek-R1 that answers more freely on topics the original model treats more cautiously, illustrating how open-weight models can be modified for different markets and norms. 
  4. So, in theory, the amount that frontier labs such as Anthropic can charge is constrained by (i) how much better they are than the open-weight models; (ii) the costs of operating them locally; and/or (iii) how much the open-weight labs charge for their interfaces. 
  5. Open-weight models are often estimated to be 4-8 months behind the frontier closed-weight models. How much are people willing to pay for a model that's just a few months ahead?
  6. This is not an either/or choice either. A couple of months ago, a developer told the FT that he spends $50 a day using Moonshot’s Kimi model for roughly 80% of his work, reserving Anthropic’s Claude for more complex tasks because using Claude alone would cost him $900 a day. 
  7. And the open-weight competition is real globally. It was reported that US technology companies have been adopting Chinese models due to cost concerns. I can see why: if these models can do most tasks sufficiently, if not equally, well, but for less than 10% of Claude's token price, they are viable alternatives in many cases. 
  8. If many developers end up using cheap open-weight models for most tasks and reserving Claude / GPT only for the hardest ones, that will have an impact on OpenAI and Anthropic's financial positions and valuations. 
  9. There have been a lot of discussions about AI safety recently. Any analysis that does not address China and/or open-weight models that can be run locally is at best incomplete. 
  10. The following statements are not mutually exclusive: (i) competition is good because it increases total welfare; and (ii) much of that welfare may show up as consumer surplus instead of profits or anything measured in GDP.
(proofread by GPT-5.5. All errors are human and mine.)

Generated by ChatGPT-5.5, with the lyrics of Eason Chan's 人工智能 ("Artificial Intelligence", 2001) as the prompt.