(📷 圖說 👉 這與我們這些年陪著客戶一起實作「先手動再自動」的基本觀念相通，經常那些有用的基礎魔法，就像放在陽台後院那支看來不起眼的掃把，但總能變出窗明几淨。不需要一直想著拔獅子鬃毛這種傳說中的特效藥，反而毫不起眼的知識盤點、流程盤點，就足以帶來色彩繽紛。攝於並木，雖然現在盛開的是櫻花，但我怕忘記，而偏好銀杏。圖片來源：Ernest。)
✳️ Ramp CEO Eric Glyman 解析 Super IC 招募哲學
今天這篇也算是 縱橫古時候與 AI 時代，來看 Keith Rabois 聊聊如何用三個不討喜，換來真心硬團隊 的延伸閱讀，Keith 推薦可以多聽聽 Ramp CEO Eric 的思路，而找到這個 2024 年份的招募實戰分享，講 Super IC（Super Individual Contributors，超級獨立貢獻者）這種被多數公司放棄的路線怎麼長出來。前情提要：Ramp 以超變態高速在成長。
（一邊看一邊覺得心有戚戚焉，更多金句靜待個幾天，我應該會整理逐字稿重點放到部落格，這幾週跟大家一起延伸閱讀、一起延伸討論覺得很開心，希望大家都跟我一樣有所收穫，就算只有一段話，只要能套用到合適的場景發揮作用，都是魔法設計的初衷吧，就這麼相信著，就可以一直寫下去。）
⌬ Spike 大於完美，接受 malformed
- Ramp 找的是 spiky（極端突出）的人，在某個領域真的比誰都深入，然後給他工具放大槓桿。（這是 2024 年的分享，工具可能改變，但是基本魔法的原理不會改變。）
- Eric 說他們找的人有時候偏向 malformed（畸形的）、有些問題，連他自己在 Ramp 某些任務職位可能都做不來。
- 這種路線發展出兩個具體人才集散地：aliens（外星人，少數被傳統大廠放棄的極致人才）跟 black sheep（黑羊，非典型履歷但具備某個極端技能的人才）。
- 不追求零缺點，但可以追求 slope（斜率）。
⌬ Slope 大於 Intercept，新鮮人可以比資深值錢
- Eric 畫了個圖。
- 資深候選人今天的起點 intercept（截距） 高，但曲線可能平；
- 某個 18 歲的大一實習生起點低，但 slope 陡。
- Ramp 會去押那條斜率。
- 他舉 Calvin Lee 為例，18 歲高中輟學後進 MIT，在他們還是公寓辦公室時被邀來做四週冬季實習，後來兩年半從 MIT 畢業。
（當年大學畢業幸運進到台積公司製程整合部門，23 歲接手歐洲客戶，然後接手美國特殊製程客戶，我不確定自己是不是那條被押注（還是壓住）的斜率。但我還記得 26 歲當時準備從台積畢業前幾週，與提拔我的長官們分享為什麼需要離開時，我也是在白板上畫了一條斜率、一條二次函式拋物線，然後帶著台積的養分衝進傳統產業挑戰拆解複雜問題，也開始了與歐洲的一些因緣巧合，也許默默之中都有安排吧。）
⌬ Strong Yes 才算 Yes，拍桌才叫負責
- 參與面試的六位評審如果都評分「還不錯、可以考慮」，在 Ramp 會被判成 no。
- Ramp 的四格評分是 strong no、no、yes、strong yes。
- 一般的 yes 等同拒絕。
- 得有至少一位評審願意把自己的信譽押注下去，公開說「我會讓他成功」。
- 他補充，一個來自對的人的 strong yes，可以蓋過其他人的 no，因為那個人正在具備這個 craft（工藝、手藝）的實務經驗，他知道自己在選什麼。
⌬ Lineage 比名氣重要：你是在黃金年代，還是搭便車
- Eric 提到早期頂尖業務 Max，不是從當紅公司來，是從 Namely 這家 NPS 負 40、產品難用到讓人討厭的公司來。
- 能在難賣的產品上拿第一，才是真本事。
- 他也用 90 年代芝加哥公牛對比 2010 年代芝加哥公牛當類比，同一家公司，在不同時代會有不同經歷。
- 招人時不要只看公司招牌，要問那個人是不是在這家公司或部門單位經歷從 0 到 1 的黃金年代，他實際打造了什麼、經歷了什麼。
⌬ 先手動再自動，SpaceX 的五步順序
- SpaceX 的五步演算法最後一步才是 automate，前面要先 question、delete、simplify、accelerate。
- 中文版：質疑需求、刪除流程、簡化優化、加速週期、最後才是自動化。
- Ramp 的招募、產品、流程都照這個順序，automate 不是捷徑，是最後一哩。
- 很多團隊急著自動化一個糟糕的流程，Eric 的提醒是，那個流程也許根本一開始就該刪掉。
📷 圖說 👉 這與我們這些年陪著客戶一起實作「先手動再自動」的基本觀念相通，經常那些有用的基礎魔法，就像放在陽台後院那支看來不起眼的掃把，但總能變出窗明几淨。不需要一直想著拔獅子鬃毛這種傳說中的特效藥，反而毫不起眼的知識盤點、流程盤點，就足以帶來色彩繽紛。攝於並木，雖然現在盛開的是櫻花，但我怕忘記，而偏好銀杏。
2026 已經過了將近三分之一，一直在調整與平衡自己與團隊與客戶的知識系統以及流程系統，一來滿足自己所需要的誇張大量閱讀，二來今年陸續展開多個實驗題目，其中包含練習大量書寫，特別是鼓起勇氣練習公開的書寫。默默也寫了約莫 8 的二次方篇的日更筆記與圖說。動力來源與原因也許留待下一趟年度回顧再來分享，但在過程中陸續收到各種有趣且神奇、甚至是有點挑戰性的回饋，而繼續前進。默默感謝每天路過輕輕按個「👍」或「❤️」的妳和你。只要記住，我們其實比我們內心的疑慮都更強大，然後每天改變一小步，例如，今天改成按個「❤️」（？
#ramp #slope
✳️ 延伸閱讀
- 縱橫古時候與 AI 時代，來看 Keith Rabois 聊聊如何用三個不討喜，換來真心硬團隊
- 拆解 Ramp 如何搶走 AmEx 客戶：從「賣錢」到「賣時間」的暗物質護城河
- 董事會不是用來解釋的：Capital One 共同創辦人的 2x2 決策框架
✳️ 知識圖譜
(更多關於知識圖譜…)
✳️ 逐字稿與筆記
引言與公司哲學
- all right great to see everybody and thank you so much for joining joining us today uh it is a 2040 uh in ramp terms uh we’re here to talk about uh super I’s uh and Technical hiring I’ll probably maybe almost think about that as as hiring for Craft um uh like with all advice uh try it on see if it fits if it does great if not no worries um but hopefully we we will share a few lessons that have been helpful to us uh in building the company um that we are today um with that um I also want to
好的，很高興見到大家，非常感謝大家今天加入我們。以 Ramp 的說法，現在是 2040。我們今天要聊 Super ICs（超級獨立貢獻者）和技術人才招募，我可能會把這件事想成是在招募工藝 (Craft)。所有建議都一樣，試試看合不合適，合適就太好了，不合適也沒關係。但希望我們能分享一些對我們打造今天這家公司有幫助的經驗。 - give uh an upfront notice that a lot of this will rhyme uh with some of the lessons uh that you heard about an hour ago um because we count the days we can also uh know what percentage of the time the company was with or without Keith uh the answer is 98% of the days uh Keith has been involved um uh he shaped the company deeply I think it’s been a big part of why we’ been able to grow the way we have and so we hope this is very useful uh for you um uh like all things uh you know a 30 minute
先聲明一下，這裡很多內容會跟大家一小時前聽到的一些課題押韻。因為我們記錄日子，所以也能知道公司有 Keith 或沒 Keith 的天數比例。答案是 98% 的日子 Keith 都有參與。他深深地形塑了這家公司，我認為這是我們能這樣成長的一大原因，所以希望這對你們非常有用。跟所有事情一樣，30 分鐘的 - talk is almost like a trailer but the original the movie uh the source the book it’s always better uh these are some of actually the source materials that have taught us uh most of what we know um from the score takes care of itself uh to high output management um how you’re thinking through being a manager uh John Wooden on leadership is incredible uh insanely simple describes a lot of the unusual management practices that shape Apple uh so there are great essays uh on Delan doio and of
演講幾乎只是個預告片，原作電影、原始素材、書籍總是更好。這些其實就是教我們大部分我們所知的原始素材。從 The Score Takes Care of Itself 到 High Output Management，關於如何思考當一個管理者；John Wooden on Leadership 非常精彩；Insanely Simple 描述了很多形塑 Apple 的非典型管理實務。還有 Delan doio〔音譯〕上的精彩文章，還有 - course David senra podcast Founders uh is just extraordinary um I really recommend these sources if you want to go deeper uh on some of the lessons things you’re hearing today after internalizing all these sources there is one consistent theme that keeps showing up whether it’s on hiring on operating a company on building great product on pursuing things we could craft and that is the theme of Simplicity uh at ramp Simplicity is really at the heart of everything that we do uh that lends to a
當然還有 David Senra 的 podcast《Founders》，真的非常棒。如果你想深入今天聽到的一些課題，我非常推薦這些來源。內化這些素材之後，有一個一致的主題會不斷出現，無論是招募、經營公司、打造好產品、追求我們能琢磨的事物，那就是「簡單 (Simplicity)」的主題。在 Ramp，簡單真的是我們所做一切的核心。這帶出了 - Simplicity of focus um something that we talk about a lot inside of the company is our singular focus and our consistent Mission uh the rule at ramp is everything we do every service we offer every product that we put out into the market has to meet one test which is does this save time or money for customers if it does we can ship it if it doesn’t we don’t and so having simple rules for your company will clean a lot of things up uh we also try to have an emphasis on Simplicity in the
「專注的簡單」。公司內部常談的一件事就是我們單一的專注和一致的使命。Ramp 的規則是：我們做的每件事、提供的每個服務、推向市場的每個產品，都必須通過一個測試：這能為客戶省時間或省錢嗎？如果可以，我們就可以出貨；如果不行，就不做。為公司設下簡單規則可以釐清很多事情。我們也試著在產品上強調簡單。 - products that we put out uh we design our products to make very complex things very easy uh what we’re trying to do is streamline uh an abstract way a lot of the complexity of running a high performing company what we do mechanically is we integrate and rebundle often what our today disparate pieces of software into simple experiences and so we’re not trying to make easier to do with fewer step expense reports we’re trying to have expenses that do themselves books that close themselves money
我們設計產品，把非常複雜的事情變得非常容易。我們試著簡化並抽象化經營一家高績效公司的複雜性。實務上，我們整合並重新打包今日分散的軟體，變成簡單的體驗。我們不是要讓費用報銷的步驟變少，我們是要讓費用自己報銷，帳自己結算，錢 - that Finds Its way to you know higher rate of return uh for your business I I I think this is a little bit of where the world is going technology being sold as software we need to learn how to use but actually a technology that does work for you selling work um but hopefully it’s give some context as well as about abuse where some of this advice is coming from and how we’re trying to apply it um and I think the closest analogy for just really really day in day out we’re trying to do uh here
自己找到更高的回報率給你的事業。我覺得這有點是這個世界要去的地方：科技以「需要學怎麼用的軟體」的形式被銷售，但實際上應該是「替你工作的科技」在賣工作本身。希望這也提供了一些脈絡，說明這些建議是從哪裡來的，我們怎麼試著應用。我覺得最能形容我們這裡每天都在做的事情的類比，就是 - at ramp uh is this these are the Raptor 1 two and uh three uh rockets at SpaceX uh they follow a simple fstep algorithm anybody can do it uh step one question every requirement step two delete any part of the process that you can three simplify and optimize for accelerate cycle times go a little faster and five automate notice that five the automation step is last you do not automate first you need to delete delete delete get down to simple requirements before you do any of this stuff and so
SpaceX 的 Raptor 1、2、3 火箭。它們遵循一個簡單的五步演算法，任何人都做得到。第一步：質疑每一個需求。第二步：刪除你能刪除的流程。第三步：簡化並最佳化。第四步：加速週期時間，跑快一點。第五步：自動化。注意第五步的「自動化」放在最後。你不會先自動化，你必須先刪刪刪，把需求降到最簡單，才做後面這些事。
打造高績效團隊
- we we try to do that not just for our products but uh a little more app to what we’re here to talk about today Simplicity and team and so these are some of the rules that we follow to create high performing teams here at ramp um welcome to the art of Hing at ramp um and this came at a good time for us uh the company is getting larger um I think in the coming weeks we will um both exciting but also a little scary for me pass over a thousand people uh here at the company um uh and I worry as we
我們不只對產品這麼做，今天更適合談的就是「團隊的簡單」。這些是我們在 Ramp 遵循的規則，用來打造高績效團隊。歡迎來到 Ramp 的招募藝術。這對我們來說時機很好，公司正在變大，我覺得接下來幾週我們會突破一千人。這讓我興奮但也有點害怕。我擔心的是隨著我們 - get bigger do you become a victim of process do you become more mediocre uh does the blade of the knife get more dull uh and does the next marginal person you hire actually make the company a little bit less effective um uh I think if you’re thoughtful about things uh we hope to be a a little bit different um we were trying to fight that with every uh fiber of of of my being of our being uh and so these are some of the uh things that we do uh at the core I think one of the the the main and
變大，會不會變成流程的受害者？會不會變得更平庸？刀刃會不會變鈍？你下一個邊際聘僱的人，實際上會不會讓公司效能下降一點？我想如果你對這些事情仔細思考，我們希望能做得不一樣。我們用我全身的每一根纖維，用我們所有人的意志來對抗這件事。這些就是我們做的事。核心上，我覺得其中一個主要的 - core ideas and beliefs that we have is is if you find deeply talented very spiky people people who are very good at one thing and you think hard about how to put them in in roles where they can express that and give them tools technology that augment and enhance that ability you can create leverage and so a lot of what we’re trying to do is do that once do that often and replicate it in lots of aspects of our company and so when we talk a lot about uh velocity counting the days we’re trying
核心想法和信念是：如果你找到極具天賦、非常 spiky（極端突出）的人，那種在某件事上真的很強的人，認真思考怎麼把他們放在可以展現這種能力的位置上，給他們工具、科技來擴增和增強那種能力，你就能創造槓桿。我們很多時候就是在做這件事：做一次，常常做，然後在公司很多方面複製。所以當我們常常談論速度、記錄天數時，我們是試著 - to really think about what are the unique skills and how do we augment them consistently over time is a bit of the the framework that’s going on uh in the back of our minds um and to put a little bit more meat uh on the bone um if you were to kind of assess uh where we spending the money how are we actually hiring um one of the first things that’ll jump out um is about half of the funds that we spend actually go towards engineering product design data science uh technical functions that are
認真思考什麼是獨特技能，以及我們如何持續擴增它們。這是我們腦海裡一直在跑的框架。具體一點來說，如果你評估我們把錢花在哪、我們怎麼招人，第一件會跳出來的事情是：我們大約一半的資金實際上是投入在工程、產品設計、數據科學這些技術功能上，這些功能 - augmenting skills uh in a lot of companies uh these teams are just building user facing or infrastructure products um a heavy portion of these teams are actually building tools for salespeople for marketing for design for customer support for talent you name it they are there to enhance uh the capabilities and the strengths of other teams next we have a a very high uh and in some cases very unusual bar for talent there are real Olympians um uh whove worked at ramp some have medled um there
是在擴增技能。在很多公司裡，這些團隊只是在做面向使用者的產品或基礎設施產品。我們很大一部分的團隊其實是在為業務、行銷、設計、客戶支援、人才招募等部門打造工具，都是為了增強其他團隊的能力和優勢。接著，我們對人才有很高的，某些情況下非常不尋常的標準。在 Ramp 工作過的有真正的奧運選手，有些拿過獎牌。 - are math Olympians probably real Olympians in in some sense but um uh you know II IMO they’re they’re here uh there are people really really really really really good at video games um some just like it but sometimes it’s an early signal that there’s something unusual going on in terms of someone’s ability to focus and persevere we’ll talk about that in a little bit um but we try to find um extremes um and look for that and find signal uh and then last um I think Keith emphasized this a lot
有數學奧林匹亞選手，某種意義上也算是真正的奧運選手。我們這裡有 IMO（國際數學奧林匹亞）選手。還有些人電玩真的非常非常強。有些人只是喜歡，但有時候這是一個早期訊號，代表這個人在專注力和毅力上有些不尋常的東西。我們待會會聊。我們試著找到極端，尋找那種訊號。最後，Keith 很強調這一點 - but I think it’s really worth harping uh on um we look uh we hire for slope over inter intercept um the goal is not zero defects uh often we’re finding uh you know maybe very weird in some cases malformed I have problems other people have issues there are roles in which uh I would fail and probably get fired at in ramp if I if I were here uh that is okay um you want to find people who are particularly uh spiky and on a steep slope and we’ll talk about what that means um we also unusually have
但我覺得值得反覆強調：我們是「斜率 (slope) 大於截距 (intercept)」招人。目標不是零缺點。我們常常找的人有些很奇怪，有些情況下偏 malformed（畸形的），我有我的問題，別人有別人的問題。有些職位我自己做會失敗，如果我是在 Ramp 應徵可能會被開除。那沒關係。你要找的是特別 spiky、斜率陡峭的人。我們待會會談這是什麼意思。我們還有一個不尋常的 - a culture of seeking out uh exf Founders um Keith said this that I’m not so well it’s never fun when someone says you know I think I’m ready to go and uh leave and start my own company um uh the day that I actually worry about is uh when that stops happening uh it means we’re no longer finding determined people who want to make things happen who are going to wake up restless and try to move things forward uh and so we actively try to be a place where uh we attract people who want to make a
文化是主動去找前創辦人。Keith 講過這個：當有人說「我覺得我準備好了，要離開去創業」，那從來都不是好玩的事。但我真正擔心的日子，是這件事不再發生的那天，那代表我們再也找不到決心要做事、每天醒來不安於現狀、想要推動事情前進的人。所以我們主動讓 Ramp 成為一個吸引想要在宇宙留下痕跡、想要創造 - dent in the universe want to create something uh worth uh creating uh give them resources build teams around them and hopefully we convince them of why the ramp opportunity is is exciting to stay for lots of years um and sometimes when they’re ready to go uh we want to do our very best to help support and see them succeed as well um uh so this is probably a sum up of of uh some of the characteristics um and done right you can measure the outputs uh uh and the efficiency of people um and so I
值得創造的東西的人的地方。給他們資源、為他們建立團隊，希望我們能說服他們 Ramp 的機會令人興奮，讓他們留很多年。有時候當他們準備好要走了，我們也會全力支持，希望看到他們成功。這大概就是一些特質的總結。做對的話，你可以量化出人員的產出和效率。 - I’ll give you one example um something that we measure uh very consistently uh is the productivity of sales development reps this is one example um sales development rep is sort of the the starter role in sales uh the quota for an average sales development rep at ramp is four times that of our next closest competitor uh some of that is we try to find extreme like very talented spiky individuals folks who show promise in what they do uh we also have Engineers growth Minds folks and analytics
我給你一個例子。我們持續測量的是 SDR（Sales Development Reps，業務開發代表）的生產力。SDR 大概是業務的入門職位。Ramp 一個平均 SDR 的配額是我們最接近的競爭對手的四倍。一部分原因是我們試著找極端的、非常有才華、spiky 的人，展現出潛力的人。我們也有工程師、成長心態的人、數據分析師 - uh who are trying to augment uh their capabilities automate parts of the job so they can focus on uh where humans are uniquely great and the rest if we can have automation uh machines uh doing other aspects of of their role uh so this isn’t just abstract Concepts done correctly um you should be able to measure yourself and hopefully show higher efficiency uh on certain types of outputs that um may be essential to how you grow and succeed as a company all right um let’s get into the good stuff
試著擴增他們的能力，自動化部分工作，讓他們可以專注在人類獨特擅長的地方，剩下的如果能讓自動化或機器處理就交給自動化或機器。這不是抽象概念，做對的話，你應該可以量化自己，並且在某些類型的產出上展現更高的效率。這些產出對於你作為一家公司的成長和成功可能至關重要。好，進入正題。
工藝的重要性
- and actually some of the Practical rules um there’s going to be four parts to this the first is going to be before you hire um in an emphasis on super ic’s uh super individual contributors um Crafts People um who sometimes manage other craft people we’ll talk about that unpack that um uh next um when you know you want to hire people how do you think about various Talent pools where to look uh and maybe find some unique Alpha uh we’ll talk a bit about selection and assessment uh and then last
實務規則會分四部分。第一部分是「招人之前」，重點在 Super ICs（Super Individual Contributors，超級獨立貢獻者），就是那些有時候會管理其他工匠的工匠。我們會來拆解。接下來，當你知道要招人時，怎麼思考不同的人才池、去哪裡找、也許能找到獨特的 alpha。我們會聊選才和評估。然後最後 - what are some of the things that as uh things are starting to work and grow um may change uh over time and so uh look I I I think anyone um you know know in their career wants to work for someone that they can learn things from that is just like a truth if you’re early in your career you want to work for other great people who will make you better so you should look for people who are great at their craft but often people uh forget this um I think a lot of times uh companies when they get
是當事情開始運作、成長，隨時間推移可能會有什麼改變。任何人在職涯裡都想為一個能學到東西的人工作，這就是事實。如果你在職涯早期，你想為能讓你變得更好的其他優秀的人工作。所以你應該找在自己工藝上很強的人，但很多人會忘記這點。很多時候，公司拿到 - round of funding they get the advice of okay you should hire uh this they’ve seen the movie before they’ve done it they’ve been the VP of this go bring them in uh and they’re going to go and know how to do it maybe they know some things but uh they don’t want to get their hands dirty um there’s some things that are unusual about ramp we very very rarely uh hire managers um directly to go into the company we prefer finding people who have done the work you know sold the deals um written the
一輪募資後，得到的建議是「你應該招這個人，他們以前看過這部電影、做過這件事、曾經是這個職能的 VP，把他們找來，他們會知道怎麼做」。他們可能知道一些事情，但他們不想捲起袖子動手。Ramp 有一些不尋常的地方：我們很少很少直接空降管理層。我們偏好找親手做過那些工作的人，真的賣過交易、寫過 - copy themselves um uh designed an interface uh not just once but continue to do that day in day out I still write copy um you know I still look at marketing collateral um I still will get involved in how things are designed uh think that’s really helpful it gives you a lot of uh taste and the ability to actually understand uh who are the people that you’re going to need what makes them great uh and when they run into problems you’ll be able to better train them and I think that uh not only
文案、設計過介面的人，不只做過一次，而是日復一日持續在做。我到今天都還在寫文案，還在看行銷素材，還會參與設計決策。我覺得這真的很有幫助，它會給你很多品味 (taste)，以及實際理解你會需要什麼樣的人、什麼讓他們變得厲害的能力。當他們遇到問題時，你才能更好地訓練他們。我認為這不只 - allows you to hire better people um but also uh allows people who are working uh with you or for you to grow quickly and want to stay working with you um um and so I wanted to start actually with just like a wonderful uh documentary um um some folks may have have seen it if not like you should watch it it’s 70 something minutes um uh it’s it’s about a guy named Jiro Ono um I think he turns 99 later this month um he is still um uh most nights serving Sushi um in what was the first three
讓你招到更好的人，也讓跟你一起工作、為你工作的人能快速成長，想要繼續和你一起工作。我想從一部很棒的紀錄片開始，有些人可能看過，沒看過的話你應該去看，大概 70 幾分鐘，講一個叫 Jiro Ono〔小野二郎〕的人。我想他這個月底就要 99 歲了。他每晚都還在捏壽司，在那家曾經是世界第一家三星 - Michelin star uh sushi restaurant in the world um and it’s actually in the basement of a subway station um in Tokyo he’s a pretty unusual Guy and um I won’t spoil the movie but there’s an amazing scene where the narrator essentially asks him you know Jiro what is your secret how do you make such um great sushi and his answer is is kind of deceivingly simple he basically says in order to make delicious food you need to eat delicious food sounds kind of obvious and he says the quality of
的米其林壽司餐廳工作。餐廳其實位於東京一個地鐵站的地下室。他是個相當特別的人。我不想暴雷，但有一個很棒的場景，旁白問他：「Jiro，你的秘訣是什麼？你怎麼做出這麼棒的壽司？」他的答案看似簡單：「要做出美味的食物，你需要吃過美味的食物。」聽起來很理所當然。他說
培養品味與鑑賞力
- ingredients is important but one must develop a pallet capable of discerning good and without good taste you can’t make good food and if your sense of taste is lower than that of the customer how is it that you will impress them then the guy goes off on some tangent about like the guy he buys rice from homi and how he knows more about rice than anything but it’s actually I think a really profound um uh moment that teaches you a lot about what it uh means to have a care of of craft and I think
食材的品質很重要，但你必須培養一個能分辨好壞的味覺 (pallet)。沒有好品味，你就做不出好食物。如果你的味覺比客戶的味覺還差，你怎麼能打動他們？然後他講到他向 Homi〔音譯〕這個人買米，那人對米的了解比任何事都多。我覺得這其實是一個很深刻的時刻，教會你什麼叫「對工藝有所在意」。我認為 - at at almost 99 he’s probably eaten more fish than probably anyone any of us have ever met you know he’s still doing it um and I think the lesson is you don’t start by by making sushi you start by tasting it you need to continue to hone your craft sales is a craft design is a craft engineering is a craft great hiring is a craft um before you try to go hire other people you need to actually do it you need to actually taste it um you need you to have your own perspective on it um and so I think
快 99 歲了，他大概比我們認識的任何人都吃過更多魚，他還在做。課題是：你不是從做壽司開始的，你是從吃壽司開始的。你必須持續磨練你的工藝。業務是一種工藝，設計是一種工藝，工程是一種工藝，優秀的招募也是一種工藝。在你試著去招別人之前，你得先自己做過、自己嚐過，你得有自己的觀點。所以我想 - actually before you even hire you should spend a lot of time upfront really thinking through uh these types of um these types of thoughts um and I think if you hire people who are great at their Craft um I think and hope they are going to create uh guilds um uh people who are dedicated um to a certain way of thinking and excellence in something that they do um name the fields you name it there is there is a a highest level of standard that you can seek and you know you can keep going pushing
在你招人之前，你應該花很多時間先認真思考這些想法。我覺得，如果你招了在工藝上很強的人，我認為並希望他們會建立「行會 (guilds)」，也就是一群致力於某種思考方式和某件事卓越程度的人。任何領域，都有你可以追求的最高標準，你可以一直 - the limits of that uh more than just a team um and I think uh doing that will change people’s even relation with with their job I think when people have great tastes um uh you can discern the difference between good ingredients great ingredients you understand when someone is a good designer um what does that mean uh mechanically in what sense in what context um um I think discernment is very important and um I will agree with with with Keith points uh earlier sometimes you know Great Taste
推那個極限。不只是一個團隊。這會改變人們與他們工作的關係。當人有很棒的品味時，你可以分辨好食材和絕佳食材的差別，你理解一個好的設計師是什麼意思，在什麼意義上、在什麼情境下。我覺得鑑賞力 (discernment) 非常重要。我同意 Keith 稍早的觀點，有時候個人 - individually doesn’t necessarily means you have good taste in assessing others but it’s a good start um because in in in in really trying to hire someone to run uh and build your company you’re asking them to do a lot of complex set of tasks um but I think without this element you need more but I think it’s a foundational element uh of hiring it’s something you want to look for um I wanted to talk a little bit about uh person problem uh matching um uh hopefully we’re thinking and you know
有好品味不一定代表你有評估別人的好品味，但這是一個很好的起點。因為當你真的要招一個人來經營和建造你的公司時，你是在要求他們做一堆複雜的任務。沒有這個元素，你還需要更多，但我認為這是招募的基礎元素，是你想要找的東西。我想談談「人與問題匹配 (person-problem matching)」。希望我們是在思考 - before we let somewhat higher um uh hopefully they work for 3 four months in your company um they have sufficient uh they’ve demonstrated their ability to work as a crafts person they’ve shown the taste they know a little bit of what they’re looking for um they may say I want to go um we have this critical need um I need to open a you know a JD or or um a job job requisition and I think one of the the critical components before you even open roles is thinking about what uni unique Spike uh do
在讓某人升職之前，希望他們已經在公司工作 3、4 個月，他們展現了作為工匠的工作能力，他們展現了品味，他們知道自己在找什麼。他們可能會說「我要開一個 JD（Job Description，職位描述）或職位請購」。在開職位之前，一個關鍵的元素就是思考你需要哪種獨特的 spike - you need in a certain style of engineering and design and I’ll give you one example from from ramp so in the early days there were effectively two archetypes of Engineers uh that worked uh at ramp when you think about ramp we issue credit cards these your payments it’s sort of like not really funny if your credit card you swipe it and doesn’t work um like it has to work all the time there’s there’s certain parts um of our stack which which uh need to have very low latency very fast response
在某種風格的工程和設計裡。舉一個 Ramp 的例子，早期 Ramp 有效地有兩種工程師原型。Ramp 是發信用卡的，這是你的錢。如果你刷卡不能用，那可一點都不好笑。它必須一直能用。我們堆疊的某些部分必須有非常低的延遲、非常快的回應 - time very low tolerance for error uh and you want to find perfectionist for these types of tasks uh and um you know when we were you know even thinking about where would we go out and look for um we wanted to think about what is the job they are going to do um and can we find people who obsess and have the specific qualities uh not just the care of craft but even the way they work the way they think for those types of tasks there’s other elements of our product where the other archetype were
時間、對錯誤極低的容忍度。這些任務你要找完美主義者。當我們想要去哪裡找人時，我們會想他們要做什麼工作，能不能找到執著的人，具備特定的特質，不只是對工藝的在意，連他們工作的方式、思考的方式都要能對得上這種任務。我們產品的其他部分，另一種原型是 - people who could just like learn really quickly and ship things daily like hourly um we get feedback from customer within 10 minutes it’s fixed um they could be sloppy sometimes and I think that’s okay um in certain context um there are some places where you know if let’s say you turn in a receipt and you know sometimes it takes uh half a second other times because the cues are too long it takes 10 seconds to the match that’s not a huge deal as long as you’re detecting it and you’re
能非常快速學習，每天甚至每小時出貨的人。我們 10 分鐘內從客戶那裡得到回饋，就修好了。他們有時候可以馬虎一點，在某些情境下那是可以的。比如你交一張收據，有時候半秒搞定，有時候因為佇列太長要 10 秒才比對到，那不是什麼大問題，只要你有偵測到、有在 - responding quickly and so there’s going to be different analogies for this in lots of different contexts but you want to be thinking about uh person problem matching just are they they good um at the art of what they’re doing but what is the company need what’s the context of the services that they’re going to be building um and um I think if you were to distill some there may be other rules that you’ll get into for certain jobs but um before you hire anyone um you know you’ll want to check
快速回應。不同情境下會有不同的類比，但你要思考「人與問題匹配」，他們不只在自己做的事情上優秀，還要看公司的需求是什麼、他們要建造的服務的情境是什麼。我想如果你要提煉一些規則，某些職位可能會有其他規則，但在你招任何人之前，你要檢查 - have they done the work do they know the craft or they graded it is there a clear need maybe a person problem match uh does the person have good judgment and discernment if you check all these boxes only then maybe they’re ready to hire uh and you can proceed on to maybe the subject of the talk how do we hire um I think this is a big one um I think this is especially true um for you know preed seed series a series B uh it is probably still true uh even at series D and above um uh people
他們做過那個工作嗎？他們懂那個工藝嗎？他們有沒有評等？有沒有明確的需求？有沒有人與問題的匹配？這個人有沒有好的判斷和鑑賞力？所有項目都打勾了，他們才準備好被招募。然後你就可以進入演講的主題：我們怎麼招人。這點很重要，尤其對 pre-seed、A 輪、B 輪來說。對 D 輪以上可能也還是如此。人們
從《魔球》中學習
- forget it sometimes when you have lots of money um um but in the abstract um I think any company at its core especially technology companies um they’re not in the technology business they are in the people business the team you build is the is is the company that you build um and uh in the world’s most competitive market which is the market for people um uh and you have very few dollars um how can you find people who are going to create a lot more value than than what it costs uh maybe to pay
有時候有很多錢的時候就會忘記。但抽象來說，任何公司，尤其是科技公司，核心上不是科技業，而是人的行業。你建造的團隊就是你建造的公司。在世界上最競爭的市場，也就是人才市場，而你錢很少，你要怎麼找到能創造遠高於支付他們 - them uh at the time um and we learned a lot uh uh from this book or if you prefer the movie that’s fine too Moneyball um it is based on a true story uh it’s it’s quite good basically the heart of it um is Billy Bean becomes the the general manager uh for the Oakland Athletics uh they have a really tiny budget um they’re losing and uh he probably was going to get fired unless he could change things again because they’re managing things for for a cost and uh the status quo the way people hired
成本的價值的人？我們從這本書學到很多，當然你偏好電影版也沒問題，就是《魔球 (Moneyball)》。這是基於真實故事。核心是比利．比恩 (Billy Beane) 成為奧克蘭運動家隊的總經理。他們預算很小，一直輸，他可能會被開除，除非他能改變事情，因為他們在成本管理中做事。現狀是當時人們招募的方式 - uh back then was they were looking for kind of Sluggers and well-rounded uh players who were ready in in any situations and that’s what Scouts look for uh he took a different approach uh he wanted to look for very specific stats that stacked together could lead to the outcomes that he wanted which was a lot of runs um and he was looking just for that and he found um you know people who could hit but couldn’t catch it was a very strange team but being able effectively to assemble this they
是找打擊者和全能型球員，在任何情況都準備好，這是球探在找的。他採取了不同的方法，他找非常具體的數據，這些數據疊在一起可以帶來他想要的結果，也就是很多分數。他就專注在那個上面。他找到了會打擊但不會接球的人。這是一個很奇怪的隊伍，但能有效地組合起來，他們 - were able to win a lot more uh than what you might expect uh and have their dollars go a lot further so they’d find people a very specific necessary spikes um and would go after them uh pretty hard um so I’ll give you a few tactics for what this can mean in practice uh first uh I recommend hiring aliens uh with extraordinary uh abilities um this is a photo um of my co-founder uh Kareem um Kareem is extraordinary he was born uh and raised in Beirut Lebanon um there are a lot of things that
贏得比預期多得多，讓他們的錢走得更遠。他們找有非常具體、必要 spike 的人，然後用力追下去。我給你幾個戰術，這在實務上是什麼意思。第一，我推薦招募「外星人 (aliens)」，那種有非凡能力的人。這是我的共同創辦人 Kareem 的照片。Kareem 很不凡，他在黎巴嫩貝魯特出生長大。有很多事情 - make him a very special individual um but I can tell you from going to college with him was uh he was one of these smartest people that could never get internships like companies just did not uh they weren’t built and well formed uh to be able to hire them so maybe the first uh Alien that we uh hired and had to figure out together was how can we actually work together uh and incorporate a company and navigate through that uh and you know today at ramp I I think that we’ve hired uh something
讓他成為非常特別的個人。但我可以從跟他一起讀大學的經驗告訴你，他是那種最聰明卻從來拿不到實習的人，因為公司就是沒有建好、沒有成形到能招他們。他可能是我們招到的第一個「外星人」，我們必須一起搞清楚要怎麼真的一起工作、成立公司、找到出路。今天在 Ramp，我想我們已經招了大約 - on the order of 20 to 30 um uh aliens with extraordinary abilities sometimes you go through an H1B sometimes you go through an O1 um but one of the unusual ways we are willing to compete is we will spend the time to actually find uh great people sponsor visas uh and that means that this excellent talent pool where there’s not a lot of competition uh unfortunately um we can find great people U perhaps this this will work for you uh spotting early Talent um this is Calvin Lee um sorry this is
20 到 30 位有非凡能力的「外星人」。有時候走 H1B，有時候走 O1。我們願意競爭的一個不尋常的方式是，我們會花時間真的去找厲害的人，贊助簽證。這意味著這個非常棒的人才池，很可惜，沒什麼競爭，我們可以在那裡找到厲害的人。也許這對你也會有用。發掘早期人才。這是 Calvin Lee。 - so grainy uh and this was in our original office um apartment um uh we we we met Calvin when he was 18 years old he was a high school dropout uh who went to MIT and he wanted a a winter one four-week internship um uh as a college freshman and I can tell you that is not a competitive market at all no one is doing this and there were certain things that were very special uh about Calvin um you didn’t need to be a genius you just needed to look at his resume uh in order to see this um
抱歉照片這麼模糊，這是在我們原本的辦公室公寓。我們遇到 Calvin 時他 18 歲，高中輟學後進了 MIT。他想要一個冬季四週的實習，當時還是大一新生。我可以告訴你，那一點都不是競爭的市場，沒人在做這個。Calvin 有些非常特別的地方，你不需要是天才，你只需要看他的履歷就能看出這些 - extraordinary capabilities uh in mathematics extraordinary achievements uh and I think had we waited a summer um we probably never would have been able to spend time with him um but he came over uh to parabus and we kept up with him um you he later interned at Google brain Facebook AI research um uh and uh he was set to go to to Citadel uh and we said how about working out of our living room um uh instead um and luckily he knew a couple things about us um and after some back and forth um uh
非凡的數學能力、非凡的成就。我覺得如果我們多等一個夏天，我們可能永遠都沒機會跟他相處。他後來來了 Parabus〔Ramp 前身〕，我們持續保持聯繫。他後來在 Google Brain、Facebook AI Research 實習，原本準備去 Citadel，我們說「不如來我們客廳工作吧？」。幸運的是他對我們有些了解，來回溝通之後 - agreed uh to come over um uh and he is an extraordinary person I think he finished from MIT in two and a half uh years um with with a 5 and and what I would say is if we waited to when most companies were trying to hire um of the junior internship or um you know it’s that it’s it’s priced in there’s enough context um about um but if start getting to know you can form deeper relationships and so i’ say for startups uh I know um you know zows there’s other folks in the audience who are a lot
他同意加入。他是個非凡的人，他用兩年半就從 MIT 畢業，拿 5.0 GPA。我想說的是，如果我們等到大多數公司都在招暑期實習或大三實習時，那時候已經被定價了，脈絡夠多了。但如果你提早認識人，就能建立更深的關係。對於新創公司，我知道在場有一些更年輕的
招募「外星人」與早期人才
- younger spend time uh with these types of people I think it goes a very very long way uh and you can find some unique advantages and we try to do this to and so we’ll we’ll offer uh winter internships fall spring um meeting people early and investing in genuine relationships is another way um uh that you can um uh compete in mispriced Talent pools uh Black Sheep um let’s talk about this one it’s a little bit personal for me my my uh my older brother um has has Asperger’s he has learning
朋友，花時間跟這種人相處會走得非常非常遠，你可以找到獨特的優勢。我們就試著這樣做，我們會提供冬季、秋季、春季實習。提早認識人、投資在真誠的關係裡，是另一個在「被錯誤定價的人才池」裡競爭的方式。接下來聊「黑羊 (Black Sheep)」。這個對我來說有點私人。我哥哥有亞斯伯格症 (Asperger’s)，有學習 - disabilities and I remember as a kid um you know your older brother is your hero uh in a lot of ways and uh you know as as you would get older and kind of go through standard School uh things got really hard for him um because he did not fit in the traditional uh way uh that people people learn um and so I couldn’t quite understand as a kid for someone who had the most unbelievable memory um could could tell you you know within an instant if you ask him about what was going on and you your
障礙。我記得小時候，哥哥在很多方面都是你的英雄。隨著你長大，走過標準的學校教育，對他來說變得很困難，因為他不適合大家學習的傳統方式。小時候我不太能理解，一個擁有最不可思議記憶的人，你問他任何事情，他馬上就能告訴你 - 421 BC in this part of the world could go right into it he had extreme spikes but um frankly really struggled um to go through through uh through high school and I I I I always just remembered thinking that in in certain contexts he could really Thrive there are things that he could do really well and I I think uh ha happily in he’s found that parents have car cared about that and we’ve been fortunate to be able to work through that and what I’d say is there are many people with stories uh
西元前 421 年在這個地區發生了什麼事。他有極端的 spike，但坦白說高中過得非常辛苦。我一直記得自己在想，在某些情境下他可以真正發揮，有些事他真的做得很好。很幸運他找到關心這個的父母，我們也很幸運能一起走過這段路。我想說有很多人有類似 - like him um uh there are um there’s an extraordinary uh engineer um here who um you know has autism um he went to to Community College um it’s not a very obvious resume but there were certain things uh about him um he was very well known in the Minecraft Community um he built extraordinary things if you actually kind of assessed and again not a lot of people are looking in the Minecraft pool um uh of who is has created kind of successful private servers but there were signals that there was
他的故事。我們這裡有一位非凡的工程師有自閉症。他讀社區大學，履歷看起來不明顯。但他有些特別的地方：他在 Minecraft 社群裡非常有名，蓋了很了不起的東西。如果你去評估，沒很多人在看「誰在 Minecraft 裡打造過成功私人伺服器」這個池子。但裡面有 - an extraordinary kind of talent which was very relevant uh when you’re building very complex um databases you need extreme attention to detail where his skill sets was very extraordinary um you’re going to find other versions of this contrarians people with um you know things that make them less obvious um uh but you want to look for that um because you can find really extraordinary people who can create a unique culture uh for you uh hi taste um or as I was putting this together people with
一種非凡的人才訊號，當你要建造非常複雜的資料庫時，你需要極端的細節注意力，他的能力組合非常不凡。你會找到其他版本的反共識者、有些讓他們不那麼明顯的特質的人。你要找那種人，因為你可以找到能為你創造獨特文化的非凡人才。高品味 (high taste)，或者當我在整理這個簡報時我想到「頭髮多 (a lot of hair)」的人。
高品味的價值
- a lot of hair um was what I wondered so I I some people have seen this interview but I thought it was actually just like excellent um um Anderson Cooper was interviewing Rick Rubin um uh wrote a great book called the creative act um and he says do you know how to work a soundboard he’s a producer he says no I have no technical ability I know nothing about music uh and Anderson asked him you must know something well and this is what I loved I know what I like and what I don’t like and I’m
有些人看過這個訪談，但我覺得這個真的很棒。Anderson Cooper 訪問 Rick Rubin，他寫了一本很棒的書叫《The Creative Act》。Anderson 問他：「你會用混音台嗎？」他是個製作人。他說「不會，我沒有技術能力，我對音樂一無所知。」Anderson 問：「那你一定知道些什麼吧？」我愛他的回答：「我知道我喜歡什麼、不喜歡什麼，而且我 - decisive about what I like and what I don’t like what are you being paid for is the confidence in my taste and my ability to express what I feel has proven helpful uh for many artists um you know I think as a as a small startup um or even a company scaling you need to be fairly opinionated in the products that you make what you put out in the world you need to be distinct in order to stand out uh in a very busy World um uh and uh a lot of people um are good generalists they can handle lots of
對我喜歡或不喜歡的東西很果斷。我被付錢的原因是對我品味的信心，以及表達我感受的能力，這些對很多藝人都證明有用。」我覺得作為一家小新創、甚至是擴張中的公司，你對你做的產品、推向世界的東西必須有相當的主見。你必須獨特才能在非常繁忙的世界中脫穎而出。很多人是好的通才，他們能處理很多 - situations it’s much rarer to find people uh with particular opinions people who’ve created a certain Craft um and I think in early um uh hires you want to be looking for people who have very specific um sometimes stubborn um but um certain views about the world um to make it more real Diego who also has a lot of hair um is our now VP of design uh we hired we couldn’t hire him originally but we could convince him to be an adviser um for a couple times you a couple hours a week um you’d think
情況。找到有特定意見、有創造某種工藝的人稀有得多。我覺得早期招募時，你要找有非常具體、有時候固執、但對世界有明確觀點的人。用具體的例子，Diego 頭髮也很多，現在是我們的設計 VP。我們原本沒辦法招他，但可以說服他當顧問，一週幾個小時。你會 - about it give feedback um to our um Junior designer at the time um about what was getting built um week after week after week slowly he was waking up and actually kind of falling in love uh with the work he was doing in his spare time uh and um not only could be fine some was taste but eventually we were able to uh I don’t know if seduce is the right word but convince him to to come over um to ramp and so I would say early on you want to find people with a very distinct and particular point
對我們當時的資淺設計師給回饋，關於正在打造的東西。週復一週，他慢慢醒過來，甚至愛上了他在閒暇時間做的工作。不只是能感受品味，最終我們說服（也許「勾引」這個詞不對）他加入 Ramp。所以我會說，早期你要找有非常獨特、特定觀點 - of view um looking for that and there’s ways you can find uh vintages and lineages um I think is an important thing sometimes you say okay great we’re ready to scale we need to hire great salespeople um or we need amazing design um and people say okay awesome let’s like sales who has great sales let’s go to uh Salesforce um or let’s go to uh Rippling and I I sort of think about this it’s easier to explain almost in basketball terms like if you wanted to hire and recruit and someone who’s
的人。尋找「年份 (vintages) 和血統 (lineages)」有方法，我覺得這很重要。有時候你說「好，我們準備好擴張，需要招好的業務」，或者「我們需要很棒的設計」。人們說「好，那業務去哪裡找？Salesforce 或 Rippling 吧。」我覺得用籃球的比喻比較好解釋。如果你要招一個
理解血統與情境
- amazing at basketball um recruiting from the ’90s Chicago Bulls in the 1990s like man that like you couldn’t do better like that was an amazing team um recruiting from the Chicago Bulls in the 2010s like you know like might be fine um uh but you know it’s a little a little bit tougher it’s not like quite the same caliber of team and yet people say stuff like this all the time they’re like oh yeah like uh this company they know sales this company you got to hire a marketing person from them
超強的籃球選手，從 90 年代的芝加哥公牛 (Chicago Bulls) 招募，那你簡直不能做得更好了，那是個了不起的隊伍。從 2010 年代的芝加哥公牛招募，可能還好，但有點困難，已經不是同一個水準的隊伍。但人們一直講這種話：「這家公司很懂業務」、「你要從這家公司招行銷的人」 - right um and so you want to think not just about who is great at the craft of what they’re doing but what was the period to even what was getting done uh often I find people who are excellent at building a function versus operating a function once it’s at scale are very very different and so matter the function like I I do recommend looking at companies and studying and being a student of what made other companies excellent at what they do um but you want to dig a little bit deeper uh into uh
你要思考的不只是誰在自己的工藝上很強，還要看那個時期、那時候做了什麼。我常發現「打造一個功能」的優秀人選和「運作已規模化的功能」的優秀人選差別非常非常大。不管什麼功能，我推薦你去研究那些讓其他公司在他們做的事情上卓越的原因，當它們的學生。但你要再深入挖一點 - was this the person who created this extraordinary function um who earn that reputation was this the right time period go one layer deeper than people often do um often a lot of people can can make mistakes on this um I will also argue there is um a hard negative there is a counter uh uh to this uh which is um you know I I’ll pick on one of our favorite neighbors uh across the street from our office is figma uh I think figma is no doubt an extraordinary company um and they might even have
是不是這個人創造了那個非凡的功能、賺到那個名聲、是不是對的時間段？比大多數人多深入一層。很多人會在這裡犯錯。我也要說，有個反向的論點。我們辦公室對街的鄰居 Figma 是我們最喜歡的公司之一。我覺得 Figma 毫無疑問是一家非凡的公司，他們甚至可能有 - world class salespeople but I have no ability to figure that out I don’t know right the product is so good um I I actually think you could be the best in the world salesperson or you could be totally fine and a lot of times uh people will go and try to find sales Talent uh maybe at the companies with extraordinary um you know product engineering whatever cultures um one of the best uh early salespeople that we ever hired uh was Max Max is still with a company um he went back from being an AE
世界級的業務，但我沒辦法判斷，我不知道。產品太好了，你可能是世界上最好的業務，也可能只是還行。很多時候人們會去有卓越產品、工程、文化的公司找業務人才。我們最早期招到最棒的業務之一是 Max。Max 還在公司，他從 AE 退回 - to being an SDR um and started the function here at ramp and something that really caught my attention Was Not only was he the top of his leaderboard but he was doing it at Namely um and Namely was a really interesting company for me because they had scaled from something like zero to I don’t know 70 80 million in Revenue in a very compressed period of time three four years something fast and the most surprising thing was they had an NPS of-40 and they like wow how did how did they do that um
SDR，在 Ramp 開啟了這個功能。真正吸引我注意的是，他不只是業績榜第一，而且他是在 Namely 這家公司做到的。Namely 對我來說是一家很有趣的公司，因為他們在很短的時間（三、四年）從零成長到 7,000 到 8,000 萬美元營收。最驚訝的是他們的 NPS 是負 40。我想「哇，他們是怎麼做到的？ - how did they sell something that people actually hated that’s got to be a good salesperson um and so there are certain functions where actually um selecting for good lineages will lead you totally astray um uh and so if you can find someone who can sell a product that um you know doesn’t doesn’t work great um and bring them to a place that actually has extraordinary products that works great just do that that’s that’s really good all right um so we’ve thought about different pools um we’re
他們怎麼賣出一個人們實際上討厭的東西？那一定是個好業務。」所以某些功能上，選擇「好的血統」反而會讓你完全走偏。如果你能找到能賣一個「不怎麼好用」的產品的人，把他帶到真的有卓越產品的地方，就這麼做，那真的很棒。好，我們想過不同的人才池，我們
撰寫簡短的職位描述
- selecting um from some places that we’re advantaged we know some rules we know what to look for uh and we’re starting to get people into the building how do we think about uh assessing uh Talent um so let’s talk a bit more about spikes um uh and so first um I think it’s worth stating uh one of the basics which is if you look at most job descriptions for most jobs in America there’s like a lot of requirements um you know three of seven years of experience um has done this 10 other things it’s
在我們有優勢的地方選擇，我們知道一些規則、知道要找什麼，也開始把人招進來。我們怎麼想評估人才？再多聊一些 spike。首先，值得說明一個基本原則：如果你看美國大多數工作的職位描述，有很多需求，三到七年經驗、做過這個還有其他十件事。這 - very like yesand and Jazzy people can add things and it sounds fun when you’re doing the job um uh description um but actually what you’re doing is with every new requirement you’re narrowing and narrowing and narrowing and narrowing and dividing and dividing whereas you could have thought about in the the world of people that could have met the things that actually mattered to your business could have been uh tens hundreds thousands uh tens of thousands and maybe you’ve narrowed it down with
很「是的，而且 (yes-and)」、很 jazzy，人們可以加東西，寫 JD 的時候聽起來很好玩。但實際上你在做的是，每加一個需求，你就在收窄、收窄、收窄、分割、分割。而原本能符合你業務真正重要事情的人，可能有幾十、幾百、幾千、幾萬人，你也許用 - this long checklist uh to maybe like five um which probably raises your price and So you you’re not you’re not getting in an advantage in that um I think as a startup and early on uh to the extent you can remember the lesson of Simplicity there should be one thing that you’re looking for maybe two U but probably one uh if you are looking just for one particular thing you can look a lot more broadly you’re going to be better able to assess spikes and you will more quickly be able to discern is
這個長長的檢查清單把它收窄到 5 個，這可能會拉高你的價格，所以你並沒有在那裡獲得優勢。我認為作為一家新創，在早期，記住「簡單」的課題，你應該只找一件事，也許兩件，但大概是一件。如果你只找一件特定的事，你可以看得更廣，你會更能評估 spike，更快能分辨 - someone truly worldclass and you can compare at one vector that really matters for your business and so uh practically what I would take away from this is uh you should write uh unreasonably short job descriptions and know exactly what you’re looking for uh and make it easy to assess someone’s performance on one metric first again this is advice really for ic’s you’ll turn into managers later um but I think it’s a good way uh of hiring um let’s talk about slope uh over intercept um so this is
這個人是不是真的世界級，你可以在對你業務真正重要的一個向量上比較。實務上我的建議是：寫不合理地短的職位描述，清楚知道你在找什麼，讓評估一個人在某個指標上的表現變得容易。這個建議主要是針對 IC，之後才會變成主管。但我覺得這是個好的招募方式。我們來聊「斜率大於截距 (slope over intercept)」。這是
斜率大於截距
- a way to visualize I think we’ve been talking about this in different iterations um but if someone has a lot of years of experience maybe they’re more Akin in the point of time is today uh to this blue uh line um you know they can create more value um uh and it’s more known and it’s priced in versus like the 18-year-old freshman intern is like here uh and maybe that intern is on a really steep slope but um what you want to be doing is hiring uh people that are closer to the red uh line um you
一個視覺化的方式。我們已經用不同的迭代談過這件事。如果有人有很多年經驗，在今天這個時間點他們比較接近這條藍線，他們能創造更多價值，而且更為人知、被定價了。而 18 歲的大一實習生在這裡，也許這位實習生有非常陡的斜率。你要做的是招更接近紅線的人。你
設計有效的面試
- want to find people early in your career you want to look for particular spikes were they created a video game do they win some competition is there something unusual when you go deep you actually can find something very unusual and go and try to find that so look for slope over intercept every time um when you’re interviewing uh people uh things should be rangy um uh disregard some of the attacks but um a lot of times what happens let’s say you meet someone you’re looking for someone in um
要在職涯早期找人，尋找特定的 spike：他們做過一款電玩嗎？他們贏過某個比賽嗎？有什麼不尋常的地方？當你深挖，你可以找到非常不尋常的東西，試著去找。每次都找「斜率大於截距」。當你面試人時，事情應該要有範圍 (rangy)。假設你遇到某人，你在找 - engineering and you want it you set up an interview with like the six other people at the company uh what happens at a lot of companies is there’s six different interviews and everyone kind of asks variations of the same thing tell me about your background um tell about a project um that you’re particularly proud of um you know uh you can go deep into something Technical and and if you kind of look at the actual time spent maybe half of the questions were actually the same thing and you keep
一個工程師，你安排他跟公司其他六個人面試。很多公司的情況是，六場不同的面試，每個人都問同一件事的變體：「跟我說說你的背景」、「說一個你特別驕傲的專案」，你可以深入某個技術主題。如果你看實際花的時間，大概一半的問題其實是同一件事，你一直 - on getting measurement of the same note and maybe you all like the note but you know why’ you have six interviews you might as well have had like two um what three I when you think about interviews mechanically you actually want to construct them um to be eliciting very different signals and very different responses um so to give you an example like you know in certain functions uh you may not know this up front but if let’s say um you’re interviewing for something at ramp uh one of those
量測同一個音符，也許大家都喜歡這個音符，但那你何必要六場面試？你不如兩場就好，或三場。機械性地思考面試，你其實想要設計它來引出非常不同的訊號和反應。舉個例子，在某些功能上你可能不會事先知道，但比如你在 Ramp 面試某個職位，其中一場 - interviews may actually be designed to debate you on something find something you say and just disagree with you on it um and the goal of that is actually to pick up one particular signal which is if you have a bunch of determined stubborn black sheepish very individual people you are almost certainly get to get into a debate how is someone going to handle that and some of the pressure um you want someone who’s looking particularly what people want out of life and out of their career U maybe
面試可能是刻意設計來跟你辯論某件事。找到你說過的某句話，就直接反對你。目標是撈取一個特定的訊號：如果你有一群決心要做事、固執、黑羊式的、非常個人化的人，幾乎一定會遇到辯論。這個人會怎麼處理？怎麼應對壓力？你要的是能觀察別人從人生和職涯中想要什麼的人。也許 - you want you want to actually see work product and work on things with people uh the point is you can design different types of interviews um but if there’s certain things not just in their craft but also in the context of working with others if there’s specific signals you are trying uh to pick up be very intentional uh in the design of your interview process uh strong yes or or bust I think this is really important um I think one of the most common outcomes for interviews um is you have six
你想看實際的作品，跟他們一起做事。重點是你可以設計不同類型的面試。如果有些東西不只在他們的工藝上，也在和別人合作的情境中，有特定的訊號要撈，那就要在面試設計上非常刻意。「強烈同意 (strong yes) 或拉倒」。這很重要。面試最常見的結果之一，是你有六個 - maybe they’re they’re really good um and you ask what do you think um so we have uh basically four grades you can give people um uh strong no no yes or strong yes uh and often with people who are kind of good it’s like I like them yes and you get to the end and it’s six yeses and at ramp that is a no um uh you want someone who is uh going to pound the table and say I’m going to go and make this person successful particularly if they’re the hiring manager and I even think if you have a strong
還不錯的面試官，你問他們怎麼想。我們有四個等級可以給人：strong no、no、yes、strong yes。通常對於還不錯的人，大家會說「我喜歡他，yes」，最後六個都 yes。在 Ramp 這就是 no。你要的是願意拍桌說「我會讓這個人成功」的人，特別是招募主管。我甚至認為，如果你有一個從 - yes from the right person that can overwhelm uh even knows in other conditions because it’s someone saying uh I am running this function I’m leading this craft uh and I’m telling you and I will put myself on the line to make this person successful um uh I think again in looking for convictions and certain spikes you want to be looking for very strong uh yeses otherwise keep waiting passing you want people with deep taste spikes capabilities uh in last um you know what would I what would I say
對的人來的 strong yes，可以蓋過其他條件下的 no。因為有人正在說「我在經營這個功能、我在領導這個工藝，我告訴你，我會把自己押上去讓這個人成功」。尋找信念和特定 spike 時，你要找非常 strong 的 yes，否則繼續等、繼續跳過。你要找有深度品味、spike、能力的人。最後，你知道我會說 - um you might even be able to disregard almost everything we just talked about um which is to observe the following um let’s say that you have like an unreasonable set of interviews and you spend like eight hours uh with a candidate U that’s like a lot of time you do the dinner you do all the meetings you have the Retro and you say okay great let’s hire them uh within uh two business days or maybe one at certain companies um you will have more information about what it is actually like to work
什麼？你甚至可以忽略我們剛才講的幾乎所有東西，來觀察以下：假設你有一組不合理的面試，你花八小時跟候選人相處。那是很多時間。你們一起吃晚餐、開各種會、做 Retro，你說「好，我們招他吧」。某些公司裡，他加入後兩個工作天、甚至一個工作天內，你對跟這個人實際工作的感受 - with that person uh than in your entire interview process uh and most people aren’t that good at interviews yet uh and so probably my top piece of advice which probably relates to what PayPal did in the early days what Facebook did what Airbnb what a lot of uh companies with a lot of uh without a lot of resour ources do uh which is hire based on referrals um you have some asymmetric information you or someone you’ve worked with or know or get advice from has worked with observed built things
就會比整個面試過程更多。大多數人還沒那麼會面試。所以我大概最重要的建議，這跟 PayPal 早期、Facebook、Airbnb、很多沒有多少資源的公司做的事相關，就是基於推薦 (referrals) 招人。你有不對稱資訊：你、或跟你一起工作過的人、或你認識的人、或給你建議的人，曾經跟他們一起工作、觀察過、一起打造過
核心抽樣與創辦人審核
- with someone for many years uh and you’re going to get a lot more signal maybe you never seen them do the job but you can know a lot of things about people um and so um you can build a great interview process but you’re better off just uh you know uh trying to use longterm information about people if you can okay uh last things I I’ll fly through what changes as you scale um and you want to look out for and this is again probably more relevant as you start to get up there in terms of number
某個人很多年，你會得到更多訊號。你可能從來沒看過他們做這個工作，但你對他們這個人知道很多。所以你可以建立很棒的面試流程，但如果可能的話，利用對人的長期資訊會更好。好，最後我快速帶過「隨著擴張會有什麼改變、你要留意什麼」。這比較相關於當你開始 - of people uh the first thing that you want to do consistently um is this exercise of core sampling um uh this often comes um uh feels where you’re studying the Earth you want to kind of drill down down down down um maybe do a first interview um and you get a certain degree of signal but if you keep asking the same question the same you go deeper on it you start seeing the depth that someone can speak at you want to do similar things with your orgs and so sometimes this shows up inide of
變大時。第一件要持續做的事是「核心抽樣 (core sampling)」的練習。這感覺很像研究地球，你要一直往下鑽。也許你做一場初步面試，得到某程度的訊號，但你一直問同一個問題，往更深處挖，你會開始看到某人能講到什麼深度。你想對組織做類似的事情。這有時會在 - organizations as you don’t do one-on ones just with someone who reports to you but their skip level their skip level you go down deep into the source to actually see um are things behaving as uh you expect um you want to be going deep into your organization to ensure that the assumptions that you’re making in scaling with other people over time are playing out in the way that you actually anticipated uh you you cannot not do this uh you there’s something that we learned from Facebook that um
組織裡展現為：你不只跟直接下屬做一對一，還跟他們的下屬、下屬的下屬做，深入到源頭看事情是不是如你所預期。你要深入你的組織，確保你在和其他人一起擴張時所做的假設，實際上真的如你預期那樣發生。你不能不做這件事。我們從 Facebook 學到一件事 - probably saved the company from going off a bit of a cliff at one point um it got to the point where um we were growing quickly enough we were adding like a 100 200 people uh a year uh and I was no longer meeting everybody um and um I was looking at one of the the recommendations an offer was about to go out um where there was all kinds of red flags all types of issues and someone was going to go and we made all sorts of mistakes someone who joined like a month go was hiring someone really
可能救了公司免於掉下懸崖。當時我們成長夠快，一年增加一兩百人，我已經不是每個人都見過。我正在看一個推薦，一個 offer 快要發出去了，到處都是紅旗、各種問題，這個人就要加入了，我們犯了各種錯誤。一個月前才加入的人，正在非常快速 - quickly all kinds of stuff I learned this story from Facebook where mark for a long time did interview everyone but at some point it changed um and effectively they they they the story is um they created effectively a Zuck review um I think up through 2,000 people prior to the approval of any packet the hiring managers people who want to go and bring someone in would bring it Mark would review it they talk about it live and sign it off at the beginning I think it was pretty bumpy um uh you
地招人，各種事情。我從 Facebook 學到這個故事，Mark 很長一段時間每個人都面試，但某個點改變了。基本上他們做了 Zuck review。我想一直到 2,000 人之前，任何聘僱方案 (packet) 核准之前，想招人的招募主管會把它帶給 Mark，Mark 會審閱，他們會現場討論，然後簽核。一開始應該滿顛簸， - know people who are new to the company were were wondering but effectively you’re you’re training you’re you’re you’re you’re you’re continuing to show what it is you want people to know uh about the company and you’re able to kind of quality assurance and scale yourself and we’ve we’ve introduced this at ramp when we were a few hundred people we still do this today where before we hire anyone I do Kareem does Colin does uh must all sign off on a person uh and that’s really effective um so
新加入公司的人會很疑惑。但基本上你在訓練，你持續展現你希望大家知道的關於公司的事情，你可以做某種品質保證並且擴張你自己。我們在 Ramp 幾百人時也引入了這個，今天還在做。在我們招任何人之前，我、Kareem、Colin 都必須簽核。這非常有效。 - that as you’re scaling one not only uh does it mean for someone hiring the Founders personally are saying yes I want you to have this person I’m excited about this but it’s also effective for closing too uh they can go to the candidate and say the founders want you here they looked at and excited when you meet them on the first day um uh you’ll know something about them as well uh and so if you can interview people for as long as you can if you can’t do that there are other ways to continue
擴張時，一方面這代表招募者說「創辦人親自說 yes，我要你有這個人，我對此感到興奮」，另一方面對收尾 (closing) 也有效。他們可以跟候選人說「創辦人們想要你來，他們看過你，當你第一天來的時候見到他們，你會對他們有所了解。」所以如果你可以盡量久地親自面試人，那就面試。如果不能，也有其他方式可以繼續 - to scale that um uh and last um as you’re now starting to build whole departments and functions uh there’s different algorithms that you might follow for individual one um for go to market um you know it’s often payback period um Associated is someone uh cost this they’re able to return this you can go and measure uh over time three months six months out are the bets that you’re making actually playing out if not you can um sort of meter how you hire uh GNA I’m General administrative that’s
擴張。最後，當你開始建整個部門和功能，你可能會對每個部門遵循不同的演算法。對於 go-to-market，通常是回收週期：這個人的成本是多少，他們能回收多少？你可以隨時間測量，三個月、六個月後，你下的賭注實際上有沒有發酵。如果沒有，你可以某種程度地計量你怎麼招人。G&A 就是「General & Administrative（一般與行政）」也就是 - me too um we need to go down um as a percentage of expenses uh as the company grows and so you want to grow uh sublinearly uh in terms of of how the company an R&D uh you can assess things often the output of those organizations takes places over many years uh versus um like a single quarter um and so you often want to be looking at input input metrics productivity and throughput metrics in order to assess um will the marginal higher um continue to slow things down or actually allow us to
我這類，隨著公司成長，佔支出的比例要下降。公司的 G&A 在這個維度上你要次線性 (sublinearly) 成長。R&D 則常常是以多年為單位評估產出，而不是單季。所以你常要看輸入指標、生產力、吞吐量指標，來評估邊際新增的人會繼續拖慢事情、還是讓我們
Q&A：具體招募情境
- speed up and take new bets that we couldn’t before um follow some simple algorithms and you’ll be able to um hire at scale um last things I’ll say this I hope this was a useful crash course into some of how we think about hiring at ramp um you know um to the extent you know anyone um you know we’re also hiring too this is my Shameless plug we have 60 open rules uh come talk to us and and if you’re you know here listening to this you might be the high slow person we’re looking for um my social
加速、下一些之前做不了的賭注。遵循一些簡單的演算法，你就能規模化地招人。最後我想說，希望這是一個關於 Ramp 怎麼想招募事情的有用速成課。如果你認識誰，我們也在招人，這是我厚臉皮的推銷，我們有 60 個職缺空著，來找我們聊聊。如果你正在聽，你可能就是我們在找的高斜率 (high slope) 人選。我的社群
早期招募 Recruiter
- media team wanted to put this one on there I I like this a lot um we work above a Home Depot which is pretty weird and interesting it has its own perks but um um look um I hope this was use SCH um uh it has been a ton of fun uh building this company over the past few years uh and with that uh wanted to open things up for for [Applause] Q&A hey uh Victor here from Ava U my question is how soon did you hire a recruiter on the team and what was the decision behind that yeah um this is
媒體團隊想把這張照片放上來，我很喜歡。我們在一家 Home Depot 樓上辦公，滿奇怪也有趣，有它自己的好處。希望這有用。過去幾年打造這家公司真的超有趣。接下來開放 Q&A。嗨，我是 Victor，來自 Ava。我的問題是：你們多快就招了第一個 recruiter？背後的決定是什麼？這是 - probably one of the more dangerous pieces of advice like I’ll I’ll give um very early uh in the context of of ramp I think that we were um four to five months old uh at the time that we hired our first recruiter I think it was our ninth or 10th hire uh at the company and so it was like seven Engineers one designer me and then this Talent person and I think the specific reason that we did that was we we knew exactly like what were the the specific experiments we were trying to run how did De
我會給出比較危險的建議之一。非常早。在 Ramp 的脈絡下，我們大約四到五個月大的時候招了第一個 recruiter，我想那是公司的第 9 或第 10 位員工，大概是七個工程師、一個設計師、我，然後就是這個 talent 人。我們這麼做的具體原因是，我們清楚知道我們要跑什麼具體的實驗、怎麼 - risk it um and we were investing ahead pretty quickly and building the capacity in order to scale uh aspects of of uh kind of ramps engineering team and um what we really were looking um wasn’t for like a head of people or HR a lot of people who want to be the first often want to do that it was very specific we want to find someone who is really excellent uh at recruiting and hunting um I think it comes from um kind of the notion of really excellent people often aren’t looking um uh sometimes
降低風險，我們很快就提前投資、打造產能，來擴張 Ramp 工程團隊的各個面向。我們真正在找的不是 head of people 或 HR。很多想當第一個的人會想做那個。我們非常具體，我們想找一個真的在招募和獵人頭上很強的人。這來自於一個觀念：真的很優秀的人常常不在找工作，有時候 - they are but uh they’re they’re building great products they’re they’re um companies are trying to keep them around uh and so the intention was to have a lot more at bats see a lot of people in certain cases where I didn’t know the function well it was to try to develop taste uh and discernment to meet lots of people over time see what good looks like um but was meant to really expand the pool that we were seeing so that hopefully we could discern a little better but it it was very early um I
他們在找，但他們在打造很棒的產品，公司想留住他們。意圖是有更多打擊機會，看到很多人。某些情況下我自己不太懂那個功能，這是為了培養品味和鑑賞力，隨時間見很多人，看看什麼叫好。但這真的是為了擴大我們看到的人才池，希望我們能分辨得好一點。不過這真的很早期。我
發展與延伸員工能力
- think if it’s your first go around uh at companies I usually would recommend probably closer in the 20s to 30s going a little bit slower is probably more more correct but um that’s at least how we thought about things yeah I think um very Roberto from um yeah other like the talent density at ramp is that actually a couple of friends they work at they’ve been here for a while um to he like principles on yeah I mean so first I uh there are a lot of companies that um they have an effective
覺得如果是你第一次創業，我通常建議接近 20 到 30 人再招 recruiter，慢一點可能比較對。但至少這是我們當時的想法。Roberto 說：Ramp 的人才密度，他有幾個朋友在這裡工作了一段時間。關於這方面的原則？好，首先，有很多公司有一種有效的 - function and they hire people above um that may be if you fully Embrace that in an effective way of building a company I don’t know how to do it um I I think I’m probably pretty ineffective at it a lot of what we try to do is find people who are really talented uh and they just keep stretching and stretching and stretching um either giveing them more responsibilities or saying I’m going to give you you know I want you to hire uh hire this um uh person lead this function um and you we see she
做法是，他們在既有職能上面再招人進來。如果你完全擁抱那個，那是打造公司的有效方式，但我不知道怎麼做，我覺得我可能非常沒效率地做這件事。我們試著做的很多事情是找真的有才華的人，讓他們一直延伸、延伸、延伸。要嘛給他們更多責任，要嘛說「我要你去招這個人、領導這個功能」。 - mentions the you know Jacob uh and Jacob by name I’ll even talk about one I mean um Jacob um was uh you know prior to coming to ramp um you know a startup founder had built something uh great wanted to see think at the next set of scale and actually came into our operations team it was supposed to be business operations he ended up working in Risk operations and underwriting really quickly um and we could tell whatever we gave him he was fast he was he was clear and people were constantly
她提到 Jacob，我甚至可以具體講一下。Jacob 在來 Ramp 之前是新創創辦人，打造過很棒的東西，想看看下一個規模是什麼樣子，實際上加入了我們的營運團隊。原本應該是 business operations，他很快就轉去做風險營運和核保。我們看得出來不管給他什麼，他都很快、很清晰，大家不斷 - going over to to Jacob to learn things from him um and I think one of the most important parts at that juncture uh in our company’s history I think Jacob was probably about nine months a year in roll uh was how do we scale um our people uh and recruiting uh function uh and so what we did is we actually say hey I know you’ve never spent time on recruiting but this is a really tough problem I’m going to throw you in we’re g to give some resources can you go and do that um and I I think that
找 Jacob 學東西。公司歷史上那個節點最重要的部分之一，那時 Jacob 大概在這個職位九個月到一年，就是我們怎麼擴張我們的 people 和招募功能。我們做的是：「嘿，我知道你從來沒花時間做過招募，但這是個很難的問題，我要把你丟進去，我們會給一些資源，你去做做看可以嗎？」我覺得 - kept I think the first time I asked him to do it he was like looked at me like I had a hole in my head and and like what do you mean you’re asking me to go do this but um he got into it fell in love with it um uh spending a lot of time um with people in and around the company some like Keith others like outside you know board advisers mentors people who are great at hiring uh and I think if you take really smart people people who are capable and just continue to stretch them uh I think that
我第一次問他時，他看我像我腦袋有洞一樣，「你叫我去做這個是什麼意思？」。但他投入進去，愛上了它，花很多時間跟公司內外的人相處，像 Keith，還有外部的董事顧問、導師、招募高手。我覺得如果你找真的很聰明、有能力的人，持續延伸他們，這 - allows um gives you a shot at keeping people who um like in the case of Jacob could go leave and start his own company uh could go do other interesting things uh and so I I think probably the thing to remember and that I would try to internalize for everyone here for employees is in some sense like everyone is the hero of Their Own Story right they’re um moving toward some goal and I think as a startup founder it’s really easy to think about like what does my company need and and just
給了你一個機會留住人。像 Jacob 這種可以離開去創業、可以做其他有趣事情的人。我覺得大概要記得的事情，我會想對在場所有人內化的是：某種意義上每個人都是自己故事的主角，他們朝著某個目標前進。作為新創創辦人，很容易想「我公司需要什麼」，然後 - optimize only for that um spend a little bit of time thinking about like who are um who are these people um what’s going to be important maybe skills things to pick up opportunities that may help them whether it’s at ramp or in the future uh and when you can find the intersection of both um I think often it leads to um people who want to stay um who are growing constantly uh and if we do our job rights hopefully convince to stay uh as long as possible um so hope that’s helpful yeah Kelly from
只為那個最佳化。花一點時間想想這些人是誰，對他們重要的是什麼，可能想學的技能、對他們有幫助的機會，無論是在 Ramp 或未來。當你能找到兩者的交集，常常會帶來想要留下、持續成長的人。如果我們做對事，希望能說服他們盡可能久地留下。希望這對你有幫助。Kelly 說：
招募高階主管
- thank you of course yeah um as a company grows how do does your rubric change if you’re hiring an executive oh um so that as the company grows um does your rubric change if you are hiring an executive um was the was the question I think it was a was a great one um so I will admit um I am not great at hiring executive and um I I think there’s probably some saying something like you know if you were world class at hiring Executives you have a 50% hit rate um I don’t even know if I’m if I’m
謝謝。隨著公司成長，你招高階主管的評估標準會變嗎？問題是：隨著公司成長，你招高階主管的標準會變嗎？這是個很棒的問題。我承認，我不是很會招高階主管。我覺得有句話說，如果你是世界級的招高階主管高手，命中率是 50%。我不確定我是否 - there I’m probably not um and so uh For Better or For Worse a bit of the ramp way um is we try to find people who are like super I’s um um people who are great at the craft and stretch them um like one example of someone who joined very early was somebody named Nick kov Nick joined um he was the DP of of engineering at a firm um and he joined ramp when we were 15 people and what was unusual about about it um as you said look if done right in three months here’s the team here’s how it’s going
有到那裡，大概沒有。好壞都是，Ramp 的做法是我們試著找 Super ICs 那種在工藝上很強的人，然後延伸他們。很早期加入的一個例子是 Nick kov〔音譯〕。Nick 加入時是某家公司的工程 VP，他加入 Ramp 時我們 15 個人。不尋常的是，我們說「如果做對的話，三個月後這個團隊會是這樣，但 - to look but you’re going to be an IC um actually for the first three months and a lot of people would say no he was really into it um he really liked it and so um I think we probably miss out um at times and we constrain um uh the set of people that we can work with because we are so stubborn uh on actually if you’re going to lead um you know instead of craft people at doing something like you need to earn their respect their admiration show them how it’s done so they’ll want to follow you in
前三個月你實際上是一個 IC」。很多人會說不，他真的很投入，他很喜歡。我覺得我們有時候會錯過人，因為我們很固執地認為，如果你要領導這些工匠，你必須贏得他們的尊重、敬佩，展現給他們看怎麼做，他們才會
全球遠端團隊
- the battle and a lot of people just simply said I’m too I’ve done too much in my career to put up with this and and and opt out um but sometimes you find people who opt in uh into it and so um it is not the only way I think there’s probably great techniques to assess um but I’ll uh just say may a c I actually um and probably not the person you want to ask about that question but it’s a good one yeah yeah hello this is mitat from Florida so I’m originally from turkey and you mentioned about
想在戰場上跟你。很多人只是說「我職涯做太多了，不想忍受這個」然後退出。但有時候你會找到選擇加入的人。這不是唯一的方式，應該有很棒的評估技巧，但我得承認，我可能不是你問這個問題最好的對象，但這是個好問題。嗨，我是來自佛羅里達的 Mitat，原本來自土耳其。你提到 - extraordinary aliens and I went ahead and actually started hiring from uh turkey and uh at like 30% of the cost of the people here like very talented people like from top companies from top schools and what’s your perspective to that like uh building a like a global remote uh Team versus like building uh together yeah I think it can be I actually think that’s a great thing uh that people can do um and I would probably venture to say your ability to assess Talent from Turkey is far above mine
非凡的「外星人」，我實際上開始從土耳其招人，成本是這裡人員的 30%，非常有才華，來自頂尖公司、頂尖學校。你對這個觀點是什麼？打造全球遠端團隊 vs 大家一起工作？我覺得這可以，我實際上認為這是人們可以做的很棒的事。我會說你評估土耳其人才的能力遠遠高於我 - and far above most people in this room uh and that probably generates Alpha for you um and I think that’s a very good thing um you know if the core thing you were trying to do um is find people who are capable of creating a lot more value than maybe it costs and hopefully you can uh pay them more um and they can participate in the upside and all that uh I actually think that’s a win-win for everybody and so so what I’d say is I I think it is underused um uh as a technique um I think um the
和這個房間裡大多數人，這可能為你產生 alpha，這是一件很棒的事。如果你核心想做的事是找到能創造比成本更多價值的人，希望你能付更多錢給他們，他們能參與上行空間，我實際上覺得這對每個人都是雙贏。我會說這個技巧被低估了。
處理高品味個體的衝突
- world like things like Zoom um digitization of work generally has brought uh the world a lot closer um uh and I actually think it’s a very good thing to be doing and um yeah I hope that’s it sounds like it’s working very well for you yeah sure thanks oh you have a might coming thanks uh one thing that I find really interesting about with Ramp is that you are very outwardly Pro Black Sheep aliens like finding people on the fringes who have very high tastes I imagine that leads to more conflict
世界，像 Zoom 這種工作的數位化，把世界拉得更近。我實際上覺得這是很棒的事。聽起來對你很好。謝謝。有人要丟一個問題。謝謝。Ramp 一件讓我非常感興趣的事是，你們公開支持黑羊、外星人，尋找在邊緣但有很高品味的人。我想這會在組織裡帶來更多衝突 - in an org than at most orgs and I’m just curious do you strive for consensus or do you strive for disagree and commit or like how do you deal with a lot of high taste High conviction folks in one room oh my gosh um I love question um look I I I think one of the greats if you listen to Founders podcast David Senra often talks about David Ogilvy uh quite a bit and he has this this very short but I think extraordinary quote which is you know tolerate genius um uh people who are particularly good
相較大多數組織。我很好奇你追求共識，還是「不同意但承諾 (disagree and commit)」？你怎麼處理一屋子的高品味、高信念的人？天哪，我愛這個問題。如果你聽 Founders podcast，David Senra 常常聊到 David Ogilvy，他有一句很短但我覺得很棒的話：「容忍天才 (tolerate genius)」。特別好的人 - are are are are are stubborn they’re a little bit out there hard to deal with um but there’s a reason that you tolerate it um because often it can lead to Unique interesting um extraordinary thoughts and um that is definitely the way that we lean um as an organization um what I would add to it is that if you’re if you’re hiring Black Sheep founder types uh out there kind of people you can be certain um especially if they’re determined uh you’re going to have fights on your hands um um and I I
是固執的、有點脫線、難相處，但你容忍是有原因的，因為常常會帶來獨特、有趣、非凡的想法。這絕對是我們作為組織傾斜的方向。我要補充的是，如果你招黑羊、前創辦人類型的人，你可以確定，特別是當他們有決心時，你手上會有戰鬥。我 - think actually not everyone does this the same way but I’ll tell you like what I often look for and try to emphasize is the goal is not to outsmart uh the other person uh this isn’t a philosophy class um you’re not trying to win a debate the goal is to uplift um uh and get there together faster Done Right hopefully Iron Will sharpen iron um two opposing th thoughts will actually help uh each other uh get to the truth you need to have a fairly low ego environment um to separate your own views
認為不是每個人都用同樣的方式做這件事，但我告訴你我常找並試著強調的是：目標不是在智力上勝過另一個人，這不是哲學課，你不是要贏得辯論。目標是相互提升、一起更快地到達那裡。做對的話，希望「鐵磨鐵」，兩個相反的想法會幫助彼此到達真相。你需要一個相當低自我 (low ego) 的環境，把你自己的觀點 - uh you know um yourself from the position and and thought you’re taking uh and well it’s hard and at times you’re going to need to just like referee certain things um I think it’s actually leads some more empowering cultures um uh it means you know best Done Right best ideas can come from anywhere I don’t care who’s the boss of some Department um uh best ideas should win um and I think often that attracts um deeper thinkers um so we don’t always get it right but uh that’s how we lean yeah
跟你採取的立場、想法分離。這很難，有時候你會需要當裁判。我覺得這實際上會帶來更能讓人發揮的文化，代表最好的想法可以來自任何地方。我不在乎誰是某個部門的老闆，最好的想法應該贏。這常常會吸引更深度的思考者。我們不是每次都做對，但這是我們傾斜的方向。