目前世界上最成功的社交媒體平臺最初只是一個點評女生“漂亮與否”的初級網站。由哈佛大學(Harvard University)學生馬克•扎克伯格(Mark Zuckerberg)創建的Facemash網站，向用戶隨機展示兩張女生照片，讓人們評價誰更有吸引力，照片都是他從哈佛校園網中搜集來的。該網站最終被封，但扎克伯格現在又回到了他在宿舍創業時的老本行。Facebook開始推出一項“約會”服務。The world’s most successful social-media platform began as a rudimentary “hot or not” site. Facemash, created by Harvard student Mark Zuckerberg, presented viewers with random pairs of student photographs he’d scraped from the university’s intranet, and asked them to rate who was the more attractive. The site was shut down but Zuckerberg is now returning to his dorm-room roots. Earlier this year, he announced that Facebook would launch a dating service.
這對該公司來說是件好事，但可能會讓用戶感到些許不安。我從2007年開始使用Facebook；它有我過去10年的帖子、狀態更新、聊天記錄和照片可以分析。很顯然，我們不知道Facebook對我們進行了多少監視，但它的確擁有大量關於我們的數據。試想，如果人們用它來尋找約會對象的話，它的算法將可以隨意使用所有那些令人尷尬的私人信息。This makes sense for the company but might leave users a little queasy. I’ve used Facebook since 2007; it has 10 years of my posts, status updates, chats and photos to analyse. It’s becoming clear that we don’t know the extent to which Facebook monitors us, but it certainly holds a huge amount of data on us. Think of all the excruciating private messaging its algorithms will have at its disposal if people use it to find a date.
然而，或許正是出於這個原因，Facebook和其他科技公司可能會成為出色的“紅娘”。當你可以讓無所不知的機器人幫你選擇伴侶時，為什麼還要費心去碰運氣呢？大多數app都要求用戶通過向左或向右滑動屏幕來接受或拒絕可能的約會對象——這正是人類一直尋求用機器來取代的那類重複勞動。然後是枯燥乏味的閒聊。如果你可以用一個聊天機器人做一些基本工作，淘汰那些特別不適合或者令人討厭的人，會怎麼樣呢?Yet it may be for just this reason that Facebook — and other tech companies — could be good matchmakers. Why bother leaving things to chance when you could just let all-knowing robots choose your partner? Most apps ask users to accept or reject potential dates by swiping left or right — repetitive labour of the sort humans have always sought to replace with machines. Then there is the small talk, which is quite boring. What if you could get a chatbot to do the basics, weeding out anyone who is particularly unsuitable or offensive?
2014年推出的應用程序Bernie，配合在線交友應用Tinder使用，替用戶完成枯燥的滑屏篩選和閒聊。其網站承諾“讓你不用每天花幾小時在滑動屏幕上”，並指出，該應用“通過清除不感興趣的約會來節省時間”。聽起來很有趣。但該應用在2017年突然關閉。（當時，該公司創始人賈斯汀•隆(Justin Long)表示，Tinder要求關閉該應用；記者無法聯繫到他和Tinder對此置評。）Bernie, an app launched in 2014, offered to work alongside Tinder and do the boring swiping and chatting for users. Its website promised “freedom from hours of daily swiping”, and suggested it would “save time by eliminating dates that aren’t interested”. Sounds intriguing. But it was suddenly shut down in 2017 (at the time, founder Justin Long said Tinder had asked the app to close; neither he nor Tinder could be reached for comment).
面部識別技術將給科技公司帶來一個迅速發展的新興數據集。程序員可以利用用戶的“漂亮與否”評分來推算人們認為有吸引力的面部特徵。應用程序可以通過這種方式來優化匹配，不過，不難想象，這些信息對廣告商也很有用。One burgeoning data set for the tech companies will be generated by facial recognition. Programmers could use users’ hot-or-not ratings to calculate which facial features people find attractive. Apps could optimise matches this way, but it’s not hard to imagine the information also proving useful for advertisers.
Hily是另一款應用，其聯合創始人揚•普羅寧(Yan Pronin)將其描述為“社交發現app”。它不會為你完成滑屏篩選，但其作用不僅僅侷限於顯示潛在匹配對象的照片。該應用仍處於測試模式，通過過濾找到興趣和關鍵詞選擇相同的匹配人選，並通過掃描信息找到“對話深度”匹配的人選。分析口頭交流會比Tinder根據地理位置遠近來推薦約會對象的做法更進了一步。Another app, Hily, which co-founder Yan Pronin describes as a “social discovery app”, doesn’t swipe for you, but it does more than show pictures of potential matches. The app, which is still in beta mode, filters matches for similar interests and word choices, and scans messages for “depth of dialogue”. Analysing verbal exchange takes things a step further than offering dates based on geographical proximity, as Tinder does.
將算法用於約會app並不是什麼新鮮事——只是最優秀的科技人才還沒有把他們的代碼分享給我們這些普通人而已。2012年，洛杉磯一位名叫克里斯•麥金雷(Chris McKinlay)的博士生意識到，他沒有有效利用約會網站OkCupid。該網站要求用戶回答一系列問題，並給未來約會對象答案的重要性打分。例如：你可能會說你相信性別平等，並給任何未來約會對象也持同樣觀點的重要性打高分。麥金雷用Python（一種代碼），根據女性的回答將她們分類。然後他編寫了多份不同的個人資料——分別側重不同的性格特徵——以提高他在不同類別女性中的匹配分值。他最終遇到了他的未婚妻，並關閉了賬戶。Using algorithms for dating apps isn’t new — it’s just that the best tech brains haven’t made their code available to the rest of us yet. In 2012, a PhD student in Los Angeles called Chris McKinlay realised that he wasn’t using dating website OkCupid efficiently. The site asks its users to answer a series of questions and rate how important their prospective dates’ answers are. For example: you might say you believe in gender equality, and mark it of high importance that any prospective date does too. McKinlay used Python — a type of code — to group women into categories based on their replies. He then wrote different profiles — highlighting slightly different character traits — to boost his match scores with different groups. He ended up meeting his fiancé and closing his account.
Facebook是最新一家進入約會市場的大型科技公司，但其他公司肯定也會效仿。我們最終可能會生活在這樣一個世界裡，Facebook、亞馬遜(Amazon)或谷歌的海量數據庫會自動為我們找到完美的約會對象。但我們真的想讓硅谷的程序員預測我們可能喜歡，或不喜歡誰嗎？我已經開始擔心，音樂流媒體服務Spotify會讓我的世界變得越來越小，該應用一直試圖向我推薦我可能喜歡聽的歌曲。亞馬遜通過推薦書籍也在做同樣的事。如今，約會app將為潛在的伴侶提供這種服務。這麼做的前提假設是人類的口味和選擇是可以預測的——這或許是真的。但在我看來，這聽起來像是機器人的觀點。Facebook is the latest dating market entrant from big tech, but others will surely follow. We could end up living in a world where the perfect date is automatically found for us among the vast data vaults of Facebook, Amazon or Google. But do we really want Silicon Valley coders to predict who we may or may not like? I already worry that the music streaming service Spotify, which keeps trying to suggest what songs I might like to listen to, diminishes my world. Amazon does the same with books. Now dating apps will do it for potential partners. The assumption is that humans have predictable tastes and choices — and maybe that’s true. But it sounds like a robot’s view to me.
閱讀更多 喵大翻譯 的文章
