Building an Autonomous Enterprise for Real-World Services with Netic Founder Melisa Tokmak
频道: No Priors: AI, Machine Learning, Tech, & Startups
视频: https://www.youtube.com/watch?v=wWbX3NL6_Uo
原文语言: en
统计: 共 70 轮 · Elad Gil 29 · Melisa Tokmak 40
[0:05] Elad Gil
Today in our prior we're joined by Melissa Takmack. Melissa is the founder and CEO of Netic, a company that builds AI for different real world services like HVAC, pet care, a variety of other things like that, roofers. Prior to Netic, Melissa was a director of engineering and worked on various aspects of go to market for scale.AI and uh also has experiences for meta. Welcome to No Prize, Melissa. Melissa, thanks for joining us today. No prior.
今天这期 No Priors 请到了 Melisa Tokmak。Melisa 是 Netic 的创始人兼 CEO，这家公司为各种真实世界的服务行业做 AI——HVAC（暖通空调）、宠物照护，还有屋顶维修等等一大堆。创办 Netic 之前，Melisa 在 Scale AI 做工程总监，也负责过 GTM（go-to-market，市场进入）的多个环节，此前还在 Meta 工作过。欢迎来到 No Priors，Melisa，谢谢你今天来。
[0:30] Melisa Tokmak
Thank you for having me.
谢谢你们邀请我。
[0:31] Elad Gil
Yeah, maybe we can start off by talking a little bit about your business and what you're building cuz I think that um you're doing something really interesting in the real world and you're kind of mirroring AI in the real world. So could you tell us more about your company and Netic and you know what you're focused on?
我们不如先聊聊你的生意、你在造的东西。我觉得你们在真实世界里做的事情特别有意思，某种意义上是把 AI 映射进了物理世界。能不能多讲讲你的公司 Netic，以及你们现在聚焦在什么上？
[0:45] Melisa Tokmak
Yeah, Ned builds AI to run millions of real world businesses that keep the world running. That means basically we work with large enterprises in essential services. We started in essential services.
好。Netic 做的是用 AI 去运营那些支撑世界运转的、数以百万计的真实世界生意。具体说，就是我们服务 essential services（必需服务）领域的大型企业——我们是从必需服务这个口子起步的。
[1:00] Elad Gil
What would be an example of essential service?
必需服务，能举个例子吗？
[1:02] Melisa Tokmak
Imagine a billion dollar revenue home services companies in HVAC, plumbing, electric or consumer wellness companies like where you can become a member and really go do a lot of sports or different activities, hospitality, automotive, uh pet services across the board. So every single thing that you need to run your life or that you want to do uh are a good example and a lot of these businesses are actually quite large businesses and they interact with millions of end users themselves usually that are consumers or businesses themselves. So netic exists between the company and its customers. So every single thing to understand the customer need or want and match that with how can we even help that customer with the operational rules of the business and even deploy the services or the labor all happens on netic.
你可以想象一家年收入十亿美元的家庭服务公司，做 HVAC、水暖、电力的；或者消费者健康类公司——就是你可以办个会员、去做各种运动和活动的那种；还有酒店接待、汽车服务、宠物服务，各行各业都有。凡是你过日子离不开的、或者你想去做的事，都是很好的例子。而且这里面很多其实是相当大体量的公司，它们自己就要跟数以百万计的终端用户打交道，这些用户可能是消费者，也可能是企业。而 Netic 就存在于这家公司和它的客户之间。从搞清楚客户的需求或诉求，到把这个需求跟公司的运营规则匹配上、看我们究竟能怎么帮这位客户，再到真正把服务或人力 dispatch（派工）出去——全都跑在 Netic 上。
[2:02] Elad Gil
So basically um say a customer calls an HVAC provider what what happens or what is netic doing for them?
那具体点说，比如一位客户打电话给一家 HVAC 服务商，会发生什么？Netic 在里面替他们做什么？
[2:09] Melisa Tokmak
Yeah. So imagine that you are in the middle of nowhere. It's -20° and your heat stops working. So you first find right a provider and there are many providers. You want to pick the most trustworthy one and the one that you can get to because you're in a minus 20° weather. Maybe you have a kid or an elderly or something is not working. Uh from there once you find a business and usually you can find these from aggregators or by search engines or in now LLMs and you can reach out to the company from any medium you want. If they're using netic, you can call them or text them or find like go to their website to channel 9uler. It's netic agents that talk to the customer. What kind of home do you live in? Do we have any of your records?
好。你想象一下：你人在一个前不着村后不着店的地方，室外零下 20 度，暖气罢工了。你第一步是找服务商，服务商有一大堆，你想挑最靠谱、而且真能赶到你这儿的那家——毕竟是零下 20 度的天气，家里可能还有小孩或者老人，东西又坏了。找到一家公司之后（一般是通过聚合平台、搜索引擎，现在也可能是通过 LLM 找到的），你可以用任何你顺手的方式联系他们。如果这家公司用了 Netic，你打电话、发短信，或者去他们官网上聊天约时间——跟你对话的都是 Netic 的 agent。你住的是什么类型的房子？我们这儿有没有你的历史记录？
[2:51] Elad Gil
I see. So, this could be like a voice call and Netic is actually providing the voice.
我明白了。所以这可能是一通语音电话，而那个声音其实是 Netic 提供的。
[2:54] Melisa Tokmak
Completely.
完全是。
[2:54] Elad Gil
And [clears throat] then the reasoning around, hey, this person needs some help and we should deploy somebody out to fix their H.
然后背后还有一层推理：这个人需要帮忙，我们得派个人上门去修他的 HVAC。
[3:00] Melisa Tokmak
Exactly. And it's actually a lot of the businesses we work with, this is why we started with essential services. The operational needs are very complex. It's not as simple as oh Eli's heat broke and now Melissa goes is it that actually what kind of even units do you have? What kind of needs do you have? Can we come to you? Is it something that we need to come to you today or tomorrow? What is your lifetime value as a customer? So should we be deploying the best person who can only work on let's say boilers or new age um systems today or later? So it's actually quite complex to first of understand what is the need from the customer can this company service it if so who when and to ensure that it is all optimized to in create customer delight to generate more revenue for the company right
没错。而且我们服务的很多公司——这也正是我们从必需服务切入的原因——运营上的诉求非常复杂。不是「哦，Elad 家暖气坏了，那 Melisa 上门」这么简单。得看：你家装的到底是什么型号的设备？你的具体需求是什么？我们能不能上门？是今天必须到，还是明天到也行？你作为客户的 lifetime value（生命周期价值）是多少？那么今天到底该不该把那个只会修比如说锅炉、或者只会修新式系统的最强技师派给你，还是排到晚点？所以这事其实相当复杂：先要搞清楚客户的需求是什么，这家公司能不能接、如果能接由谁来接、什么时候接，还要保证整个过程是最优的，既让客户满意，又能给公司多创造收入。
[3:52] Elad Gil
that makes sense what what what is a sort of like incumbent version of you know what you're doing or what are what are other players in the market or how should people think about you know what this replaces.
有道理。那你们对标的「传统做法」大概是什么样？市面上还有哪些玩家？大家该怎么理解你们替代掉的到底是什么？
[4:02] Melisa Tokmak
So let's let's go back how how do people do this today? Obviously it's we are a two-year-old company. So before before us too, how were they doing that? So it's primarily with people. These companies are extremely large and if anything actually they can't grow because they have to keep investing in people to think of any kind of growth and for them they're eBida businesses. Many times they're actually owned by private equity. They need to care about their margins and what they make to be able to invest back in the business. The most important labor in the business is actually the people who are going to deliver the job, right? You want to focus every single thing to make your end customer happy because a lot of these businesses are also commodities. If you do not answer me, I'm going to go to the next one, next search result on Google, next result on like now charg. So it's very important to be able to invest your money into your labor, the bluecollar labor. So [snorts] until now it was with hundreds and hundreds of people in these companies trying to figure out how to support customers at all times. The teams that support these workflows in the companies are unreliable as well. Let's look at a day in a life with this company that makes a billion dollar revenue. You might show up. The business usually start at 4:00 or 5 am. You don't really have anybody in the company. Then even maybe your technicians start at 6:00 a.m. But before then, no one is in the company.
我们往回退一步——今天大家是怎么做这件事的？毕竟我们只是一家两年的公司，在我们之前他们靠什么？主要靠人。这些公司体量极大，而且说白了它们长不动，因为想谈任何增长，就得不停往人身上投钱。它们是看 EBITDA（息税折旧摊销前利润）的生意，很多其实是私募股权持有的，必须在意利润率、在意赚到多少才能再投回业务里。而这门生意里最重要的人力，其实是真正上门交付的那批人。你希望把每一分资源都花在让终端客户满意上——因为这类生意很多本身就是同质化的商品型服务。你不接我电话，我就去找下一家，谷歌搜索结果的下一条，或者现在 ChatGPT 给的下一条。所以把钱投到你的蓝领人力上非常重要。可到目前为止，这些公司都是靠成百上千号人在那儿琢磨怎么全天候支持客户。而且公司里支撑这些 workflow 的团队本身也很不稳定。我们看一家年收入十亿美元的公司，一天是怎么过的：你来上班，这类生意一般凌晨 4 点、5 点就开门了，可这个点公司里基本没人；就算技师是早上 6 点开工，在那之前公司里一个人都没有。
[5:28] Melisa Tokmak
You might come in as a manager, but there are at least three people that quit that way or that day or five didn't show up. So all these customers are piling in starting at 6:00 a.m. because now it's a heat wave in the country. Actually, this past week, we have hit record waves uh last week in the country. What do you do as a business? So many of these businesses are cyclical businesses. It's very important in winter, summer, holiday season, season people start working out, right? So you have to make sure that you're capturing a lot of your value in those months to be able to sustain yourself and your employees
你作为管理者来上班，结果那天至少有三个人辞职，还有五个人没到岗。而客户从早上 6 点开始就不停涌进来，因为全国正赶上热浪——实际上就在上周，全国刚创了热浪纪录。这时候你这门生意怎么办？而且很多这类生意是周期性的：冬天、夏天、节假日、大家开始健身的那个季节，都是关键节点。你必须保证在这些月份里把该拿的价值都拿到手，才养得起自己和员工。
[6:04] Elad Gil
because you're providing like agents taking overflow. So somebody's giving you a lot of phone calls, they can pick up the phone calls for you.
因为你们提供的像是「接溢出」的 agent——有人电话被打爆了，你们可以替他把这些电话接起来。
[6:09] Melisa Tokmak
Yeah. Over that is many customers how they started. But today actually over 70% of our customers are AI first. We call it netic first. They go N1. Uh and over 70% actually all of their customers first interaction with the company is with NetIC agents.
是的，很多客户一开始确实是这么用起来的。但今天，我们超过 70% 的客户是 AI 优先的——我们叫它 Netic 优先（netic first），他们直接上 N1。而且超过 70% 的情况下，他们所有客户与这家公司的第一次交互，都是跟 Netic 的 agent 完成的。
[6:26] Elad Gil
So there's a few different approaches you can do for markets like this. And so um obviously I'm excited to be an investor in ETIC. I've also invested in AI roll-ups where people will uh you know a company like Long Lake will go and buy a bunch of businesses and then optimize them with AI. And in some cases, these businesses overlap in the profile with your customer set. And you mentioned many of them are owned by private equity already. How did you make the decision to build this service versus just doing a rollup or buying and running the assets or the trade-offs in terms of those approaches?
面向这类市场有好几种打法。首先声明，我很高兴是 Netic 的投资人；同时我也投过 AI roll-up（并购整合）——比如 Long Lake 这样的公司，去收一堆生意，然后用 AI 把它们优化一遍。有些情况下，这些被收的生意画像跟你的客户群是重合的，而且你也提到，其中很多已经被私募股权持有了。你当初是怎么决定去做这样一个服务，而不是干脆自己做 roll-up、把这些资产买下来自己经营？这两条路各自的取舍是什么？
[6:54] Melisa Tokmak
I think three reasons. One, what I want to build in the world, skill set and scale. Um you know before building this I was at scale for about four years and I built a lot of their real world businesses from government to large enterprises.
我觉得有三个原因。第一，我想在这个世界上造出什么；第二，我自己的技能组合；第三，规模。创办这家公司之前，我在 Scale 待了大概四年，把他们很多面向真实世界的业务从零做了起来，从政府客户到大型企业。
[7:10] Elad Gil
And scale by the way uh for uh folks who are watching or listening to this uh was uh started off as a data labeling company. It ended up eventually um having some sort of licensing or agreement with Meta where Meta paid I think it was $28 billion.
顺便给在看或在听的朋友补一句，Scale 最早是一家数据标注公司，后来跟 Meta 达成了某种授权或者合作协议，Meta 付了我记得是 280 亿美元。
[7:26] Elad Gil
Yeahish a little 30 around there. So real money. Um and so you were a very early and like important person I think at scale um in terms of the the variety of the different businesses you worked on and then you decided to leave before that happened actually to to start this business.
差不多吧，接近 300 亿。反正是真金白银。所以你在 Scale 算是非常早期、而且我觉得相当关键的一个人，因为你操盘过的业务线跨度特别大。然后你其实是在那笔交易发生之前，就决定离开出来创业的。
[7:41] Melisa Tokmak
Yeah, for me there were two things really important. It's scale that showed me a lot more about the real world businesses. So a lot of my business units to give you a sense I built at scale the government business unit and large enterprises in logistics, manufacturing, financial services, healthcare across the board. And I think the tangible impact you can make in the world is incredible. It is the fulfillment you actually get. You know, in tech, we keep talking about this fulfillment that doesn't usually come to many people. But I I I love doing that. My background, I grew up in a very small town in Turkey. I grew up with nothing and really came here only for college. When I got a full scholarship to Stanford, I didn't even own a computer before. And all of my town is actually in these type of industries. So to me, all of us come here and we talk about tech, we talk about creating impact, but majority of the companies that only really serve other startups or tech companies. So to me that was very important. What are we building for the real world? And I combined that with a technical pursuit which was a selfish pursuit is that AI is good at assisting consumers or being a co-pilot. But the next most unsolved problem is that how are we using AI in mission critical workflows? What does that look like to have a fully autonomously executing actually systems? So netic came out of really combining these two for me and at the end I loved being at scale. I was a very good culture fit I think and I loved my team. Some of my best friends actually are from there.
对，对我来说有两件事特别重要。一是 Scale 让我见识了大量真实世界的生意——给你个概念，我在 Scale 亲手做起来的业务单元包括政府业务，以及物流、制造、金融服务、医疗等各个领域的大企业业务。我觉得你能在这个世界上产生的那种看得见摸得着的影响力，实在了不起，那才是你真正拿到的成就感。在科技行业我们总在谈成就感，但很多人其实并没有真的得到。而我是真心喜欢做这个。说说我的背景：我在土耳其一个很小的镇上长大，一无所有，来美国完全是为了上大学——我拿到斯坦福全额奖学金的时候，自己连一台电脑都没有过。而我们镇上的人基本都在干这类行业。所以对我来说，大家都跑到这边来谈科技、谈创造影响，可绝大多数公司真正服务的，其实只是其他创业公司或者科技公司。所以「我们到底在为真实世界造什么」这个问题对我特别重要。我把它和一个技术上的追求结合起来了——那算是个挺自私的追求：AI 现在很擅长辅助消费者、当 co-pilot，但下一个最没被解决的问题是，我们怎么把 AI 用到 mission critical（关键任务）的 workflow 里？一套真正能完全 autonomous（自主）执行的系统，究竟长什么样？Netic 就是这两件事在我身上结合出来的东西。最后说一句，我在 Scale 待得很开心，我觉得自己跟那儿的文化非常合拍，也很爱我的团队，我最好的几个朋友就是在那儿认识的。
[9:14] Melisa Tokmak
Obviously Alex was amazing person to work with directly as a founder and but to me that it wasn't missing the product right what is the product because it is an operationally heavy business to build the infrastructure for AI even though people don't want to talk about it and it was rewarding but missing that product edge for me so I knew I wanted to build a product that can scale and compound and two personal skill set I think in a lot of those rollups the main important thing is the M&A itself. I'm not an M&A person. I'm a builder. I am an engineer. I'm a product person. It just doesn't fit my skill set and I don't want to build a business where clearly what I can provide is not the most important thing. And I think third is the scale of the product too. What I have seen there's a lot of successful roll-ups and tech enabled rollups, but at the end all the products you're building are for the company you just bought. it can't really be actually applied to any other company. So you are committing to buying these a few companies in whatever industries that you are interested in and serving those with your products versus what I'm interested in is how every real world business can run on netic right if we didn't have to limit them if they could focus on what they are good at in that business which is the labor which is the differentiation quality of the service how could netic run the rest
Alex 作为创始人，直接跟他共事的体验也非常棒。但对我来说，那里缺的是产品那一块——产品到底是什么？因为给 AI 造基础设施本身是一门运营很重的生意，虽然大家不太愿意谈这一点。它很有回报感，可对我来说少了那份产品的锋芒。所以我很清楚，我想做一个能规模化、能复利叠加的产品。第二是个人技能组合：在很多 roll-up 里，最关键的能力其实是 M&A（并购）本身。我不是做 M&A 的人，我是个 builder，是工程师，是产品人。这跟我的技能根本不匹配，而我不想去做一门「我最能提供的东西显然不是最重要的东西」的生意。第三是产品能做多大。我见过很多很成功的 roll-up，包括技术驱动型的 roll-up，但归根结底，你造的所有产品都只服务于你刚买下的那家公司，很难真的迁移到别的公司身上。所以你等于是在承诺：在你感兴趣的那几个行业里买下几家公司，然后用你的产品服务它们。而我感兴趣的是——每一家真实世界的公司都能跑在 Netic 上，会是什么样？如果我们不必给它们设限，如果它们可以只专注自己真正擅长的那部分——也就是人力、也就是服务质量上的差异化——剩下的全部交给 Netic 来跑，会怎么样？
[10:38] Elad Gil
how do so when people talk about um AI for the real world they mean two or three things right so obviously there's is what you're doing at Etek and serving these businesses that you know go and implement different services for people in their homes or their pets or you know different aspects of their life. There's people who talk about robotics they're self-driving like do you think all these things converge over time? Do you think the timeline of that is very far in the future and doesn't kind of matter in terms of these other types of things?
那么，大家谈「面向真实世界的 AI」时，一般指的是两三类东西：一类显然是你在 Netic 做的，服务这些上门给人家里、给宠物、给生活各个方面提供服务的公司；另一类是有人在谈的 robotics（机器人）、自动驾驶。你觉得这些东西最终会收敛到一起吗？还是说那个时间点太遥远，遥远到对眼下这些事其实没什么影响？
[11:02] Melisa Tokmak
Yeah, it's almost you're looking at the same book maybe but different chapters. We're talking about much more now chapters and for the next few decades. I do believe that there's a chapter in the future. It is all about robotics. But in some of the industries that I'm working with that is quite far in the future. And I always say you know if you spend a few days with these business which is very important for us like you know I have to tell you about how I learned about these uh industries and every single engineer in our company do have to go visit customers on site to build the right product back home in San Francisco. And it if you look around, if we look outside of a window, look at every single building. If robotics is going to do what we're doing with these companies today, every single of those buildings need to be either 3D printed
是的，感觉像是在看同一本书，只不过是不同的章节。我们现在谈的是当下这一章，以及未来几十年的那几章。我确实相信未来会有一章完全属于机器人。但在我服务的一些行业里，那还相当遥远。我常说，你得跟这些生意实地待上几天——这对我们非常重要，我得跟你讲讲我自己是怎么摸清这些行业的：我们公司每一个工程师都必须去客户现场实地拜访，然后才回旧金山做出对的产品。你环顾一下四周，往窗外看，看看每一栋楼。如果要让机器人做到我们今天替这些公司做的事，那这些楼要么得是 3D 打印出来的，
[11:53] Melisa Tokmak
or completely like standardized. I don't see that happening. And that's just because you just think robotics will be able to navigate the different variation or variety in terms of
要么就得完全标准化。我看不到这种事会发生。——这仅仅是因为你觉得机器人没法应付现实里这么大的差异和多样性吗？
[12:03] Melisa Tokmak
I think if you look at robotics capabilities today it is quite far in terms of dexterity on how to be able to handle different types of um screws even what type of homes are how am I going up how am I going into the tiny areas to be able to fix something many times it's not clear you have to open up the whole walls to even see what do you need to fix how and on On top of that, there is in these industries, there's a lot of the human element too because when a lot of people are dealing with these industries, it is worst day of their lives, right? It's either their home is flooded or they want to take the day and reduce stress and go play some tennis at Bay Club, right? But whatever it is that they're actually going through something and there's an human element, I do think there's quite a bit of a while for robotics to be closer in our chapter.
你看今天的机器人能力，在灵巧性上还差得远——怎么应对各种不同类型的螺丝、房子是什么结构、我怎么爬上去、怎么钻进那些狭小空间去修东西。很多时候问题都不明确，你得先把整面墙拆开，才能看清到底要修什么、怎么修。除此之外，这些行业还有很重的「人」的成分——因为很多人打交道的时候，正处在他们人生中最糟糕的一天，对吧？要么家里被水淹了，要么他们只是想少点压力、抽一天去 Bay Club 打打网球。不管是哪种，他们确实正在经历某件事，这里面有人的因素在。所以我确实觉得，机器人要走到我们这个环节，还得等相当长一段时间。
[12:56] Elad Gil
And then I've kind of heard you talk about um the big labs, so open AI, anthropic, uh maybe Google, Meta, etc. Um how do you think they'll process industry or how do you think what you're doing is different from what they can do?
我听你聊过大模型实验室——OpenAI、Anthropic，可能还有 Google、Meta 这些。你觉得它们会怎么切入这个行业？或者说，你们在做的事和它们能做的事，区别在哪？
[13:08] Elad Gil
Different way of asking that maybe can labs do this? Mhm.
换个问法可能就是：这件事实验室自己能不能做？嗯。
[13:12] Melisa Tokmak
Um I think that's a fair question because there's a lot of startups in today's world that do that build things functionally and visually very similar to the labs key products, right? It's either a chatbot or a coding agent. Um I do chuckle at the question a little bit though because I think 10 years ago that same exact question was can Google do this and then now it became can labs do this? Sure, some of the things the core competencies they can do it but some of the other things they're not investing in it. For Ned's case, I don't see them as a competitive risk. It's actually I think the two especially the leading labs, there are amazing businesses. They're also great partners to companies like us. So I really respect that. But in terms of looking at what we provide to these industries and companies, I actually think two things are very important. one in focus in what you're building and I will say you know it would be a funny question to these enterprises right open AI builds amazing products really fast but also it kills them really fast so I don't think enterprises or at least enterprises in these industries looking for that really fast or in anthropics case you know Silicon Valley converged in the idea that you know they pulled ahead in coding agents um or like claude because they had focus but you see exactly the opposite in the enterprise case there's about like 20 products like what is really happening um and I don't really see that and a meta focused question maybe about the labs and specifically researchers they really care about solving the most generalizable way of the problem right so in this case maybe looking at the problem we're solving the answer would be well when we get the a AGI we'll ask how to solve it for essential services and I think that is both operationally and intellectually a bit lazy thinking And the finally is for to solve these type of extremely difficult problems
我觉得这个问题问得合理，因为今天有大量创业公司，做出来的东西在功能上、视觉上都跟实验室的核心产品高度相似，对吧？不是聊天机器人，就是 coding agent。不过这个问题我听着还是想笑，因为十年前一模一样的问题是「Google 会不会自己做？」，现在变成了「实验室会不会自己做？」。当然，有些属于它们核心能力范围的事，它们做得了；但另一些事，它们根本没在投入。就 Netic 这个场景，我不把它们看成竞争风险。实际上我觉得——尤其是头部这两家——它们是非常了不起的生意，也是我们这类公司很好的合作伙伴，这点我是真心敬重的。但如果看我们给这些行业和企业提供的东西，我觉得有两点特别关键。第一是「专注」，你做的东西够不够专注。我会说，这个问题拿去问那些企业客户其实挺好玩的——OpenAI 造产品又快又好，但它砍产品也一样快。所以我不认为企业、至少这些行业里的企业，会想要那种「快」。或者拿 Anthropic 举例，硅谷有个共识是：它们在 coding agent 上跑到了前面，Claude 之所以领先是因为够专注。但在企业级这块你看到的恰恰相反，大概有二十个产品同时铺开，到底在发生什么？我不太看得懂。另一个更元一层的问题，是关于实验室、尤其是研究员——他们真正在意的是用最泛化的方式解决问题，对吧？所以放到我们在解决的这个问题上，他们的答案可能是「等我们拿到 AGI 了，再去问它怎么解决必需服务这件事」。我觉得这在运营上和思维上都有点偷懒。最后一点是，要解决这种极端困难的问题——
[15:08] Melisa Tokmak
with millions in the country that have completely different worries, different accents. How do they want to engage different contexts and also even engage them again to make them multi-time customers. There's quite a bit of lost mile that you really have to do that doesn't only come from models that has to come from your harnesses and orchestration, the software and the product that you have to build on top. So if anything actually companies like us and Netic have to be good at all three layers.
——面对全国上百万各不相同的人，他们的顾虑不同、口音不同、想要怎样被服务的方式不同、所处情境也不同，你还得让他们愿意再来一次、变成回头客。这里有一大堆「最后一公里」的活儿是你必须亲自做的，而这些不只来自模型，它得来自你的 harness 和 orchestration，来自你在上面搭的软件和产品。所以要真说的话，像我们 Netic 这样的公司，反而必须三层都做得好。
[15:34] Elad Gil
Yeah, that makes sense. One thing I've noticed um which is more sort of a side comment on what you're saying is that one big shift I see from four years ago and now you know four years ago I was funding things like Harvey and Perplexity and uh you know a little bit later decagon and a bridge and a lot of the sort of vertical applications obviously netic um and I feel like a lot of founders now are almost a little bit too worried about the labs they're doing and so they're not entering new verticals or sort of staying away from things that they think is in the road map of the lab when I think in traditional times people would have kind of fought it out a bit more and so I think that's a little bit surprising.
对，这说得通。我注意到一件事——算是对你说的话的一点旁注——我看到四年前和现在有个很大的变化。四年前我投的是 Harvey、Perplexity 这类，稍晚一点是 Decagon、Abridge，还有很多垂直应用，当然也包括 Netic。我感觉现在很多创始人对实验室的动作有点过度担心了，结果他们不敢进新的垂直领域，或者刻意躲开自己觉得会出现在实验室 roadmap 上的方向。放在以前，我觉得大家会更愿意正面拼一拼的。所以这点让我有点意外。
[16:06] Melisa Tokmak
Yeah, I do think that's because there's a lot of building going on currently that's focused on how can I exit immediately.
是的，我觉得原因在于，现在有很多创业是奔着「我怎么才能尽快退出」去做的。
[16:13] Elad Gil
Oh, interesting.
哦，有意思。
[16:14] Melisa Tokmak
Uh instead of I think being a founder I actually think was a more honorable thing before like you knew you're dedicating your life to it. Like I I'll give you an example before I built I actually really looked for a job. Is there something I want to build? I looked at the fourth thing I want to build and scale. Is there anything else I'm interested in? And afterwards, is there someone that I really want to help further the mission? I think being a founder is a very difficult thing that you have to dedicate your whole whole life for decades. It is not something to be taken lightly, but people are worried about this too much because I think they're looking at it on like, oh, will I be able to exit the short time frame exits? And then the second thing is, maybe this goes into hiring philosophy a little bit. People who are at Netic who deeply care about what they are building themselves customer obsessed for the long term have the agency to start thing urgency to solve problems but at the same time to really rigor and patience to carry it forward right I see especially Gen Z in these days when I'm you know looking at hiring obsessed with this permanent underclass mentality if I don't make my money in the next 18 months or if I don't learn everything in the world in the next six months months [clears throat] then I am forever poor or this is a AG agi pill view of you know in 18 months uh AI will subsume a lot of people skill set and so your value decreases yeah
而不是——我觉得以前当创始人是件更值得敬重的事，你清楚自己是在把一生投进去。举个我自己的例子，在创业之前我是真的去找过工作的。我先问：有没有什么是我想做出来的东西？我看的第一件事是我想构建并规模化的东西是什么。还有没有别的我感兴趣的？之后再问：有没有哪个人是我真心愿意去帮他把使命推进下去的？我觉得当创始人是件非常难的事，你得把整个人生投进去，一投几十年，这不是能轻率决定的事。但现在大家担心得太多了，因为他们的视角是「我能不能在短周期里退出」。第二点，这可能有点扯到招聘理念了——在 Netic 的人，是那种发自内心在乎自己所做之事、长期以客户为中心、有 agency 去把事情启动、有紧迫感去解决问题，同时又有足够的严谨和耐心把它一路推下去的人。我在招人时特别注意到，现在的 Gen Z 有一种「永久底层」的执念：如果我不能在接下来 18 个月里挣到钱，或者不能在接下来六个月里学完世上所有东西，那我就永远穷了。或者是一种被 AGI 洗脑的世界观——18 个月后 AI 会吞掉很多人的技能，你的价值就贬值了。是的。
[17:44] Melisa Tokmak
which I think is actually a very dangerous mindset in fact building really good things take a very long time I have learned some of the most important lessons in my life from you know committing to things and keeping up with it with all the problems and with especially at scale for example over really those four years and now with building a company because you have to see not just what's the first version of a thing
我觉得这其实是非常危险的心态。事实上，真正好的东西都要花很长时间才能做出来。我人生中最重要的一些教训，恰恰来自于对某件事的长期投入，以及带着所有问题坚持下去——尤其是在规模化的状态下，比如我在那四年里的经历，现在是做公司。因为你要看到的不只是一件东西的第一个版本。
[18:09] Melisa Tokmak
how are you thinking about what does it do in the world what do you have to improve how do you keep up with other research how do you keep up with people how do you need to improve that right I think um especially we look for that we really avoid shiny object seekers and there's a recently I was reading uh this you know quote attributed to Martin Luther actually it says you know the Christian shoe maker, you know, doesn't honor God by putting little crosses on the shoes. It does so by building the best shoe,
你得想：它在真实世界里到底起了什么作用？你需要改进什么？你怎么跟上其他研究进展？怎么跟上人的变化？你需要在哪些地方提升？我们招人时特别看这一点，我们非常刻意地避开那种追新鲜玩意儿的人。最近我读到一句据说出自马丁·路德的话，大意是：一个基督徒鞋匠，不是靠在鞋上刻小十字架来荣耀上帝，而是靠做出最好的鞋。
[18:40] Melisa Tokmak
right? The best shoes. Uh because God cares about craftsmanship. And I think I really resonate with that when you think about as a founder, what are you building? Why are you doing this? Right? It really is about building the best product that are solving the problems of the people you care about. If anything, you actually achieve more with less, with more focus and with a small group of people that are really that came together to take down the whole world. Right.
对，做出最好的鞋。因为上帝在意的是手艺。我对这句话真的很有共鸣——当你作为创始人去想，你到底在做什么？你为什么要做这件事？归根到底，就是做出最好的产品，去解决你在乎的那群人的问题。真要说的话，你反而能用更少的资源、更强的专注、一小群真正凑到一起想要征服世界的人，做成更多的事。
[19:09] Elad Gil
You mentioned um selecting for people with agency. How do you screen for agency? Like what do you look for? Is it an interview question? Is it experiences? Is it something else?
你刚提到挑有 agency 的人。你是怎么筛 agency 的？具体看什么？是靠一道面试题？还是看经历？还是别的什么？
[19:18] Melisa Tokmak
Um I think agency what I look for is like not agency. Now, if you have agency, you care about agency, you have shown agency continuously in your life. So, I will actually dig in. How how was it? Like, if you're a new grad, you don't have to have a job. Did you do something in college? Did you do something to be able to get into college? Do you have a project do you really care about? And not just that, like I did this for a weekend, but did you keep up with it? Right? So, what are really those examples? And I think I do ask many times. It's just one question. I'd like to know about what people do, what has been the hardest thing they have ever done in life and they can take it anywhere but it's not just one question but we'll really dig in to understand the why
我看的其实不是「agency 这个词本身」。如果你有 agency、你在意 agency，那你这辈子应该是持续地展现过 agency 的。所以我会往下深挖：当时是怎么回事？比如你是应届生，你不一定有工作经历，那你在大学里做过什么？你为了考进这所大学做过什么？你有没有一个自己真心在乎的项目？而且不只是「我做了一个周末」，你有没有坚持下去？那些具体例子到底是什么？我很多时候会问——就一道题——我想知道人们都做过什么，问他们这辈子做过的最难的事是什么。他们可以往任何方向答，但这不是一问一答就完了，我们会往里挖，去理解背后的「为什么」。
[20:03] Melisa Tokmak
or you pull that apart from someone's life story right but the the most important thing I care about not to have one agency example is just however long you have been alive and conscious
或者说，你是从一个人的人生故事里把它拆解出来。但我最在意的，不是找到某一个 agency 的例子，而是：在你活着、有意识的这么多年里——
[20:17] Melisa Tokmak
have you been showing agency in things that you did in life and Did you keep up with them? Did you stick through?
——你在自己做的事情上有没有持续展现出 agency？你有没有一直坚持下来？有没有咬牙撑到底？
[20:24] Elad Gil
So, you care about the follow through. You started something, you did something unique that most people don't do
所以你在意的是后续的执行力。你启动了某件事，做了一件大多数人不会做的、独特的事——
[20:28] Elad Gil
and then you kept going on it. It's funny. I used to ask the same question of what's the hardest thing you've ever done.
——然后你还一直做了下去。有意思的是，我以前也问同一个问题：你做过的最难的事是什么。
[20:33] Elad Gil
And I got such bad answers that I stopped asking it.
结果我拿到的答案太糟了，后来我就不问了。
[20:36] Melisa Tokmak
Well, I mean, you know, I don't know why you asked it, but you know, if you get bad answers, you know, maybe that person is not really a good
呃，我不知道你当时为什么问这个，但如果你拿到的都是糟糕的答案，那也许说明那个人本身就不太行——
[20:42] Elad Gil
Well, it's just consistently. I just mean like the hiring pool that I was at least looking after that.
不，是一直如此。我意思是，至少在我当时接触的那个招聘池里是这样。
[20:46] Melisa Tokmak
I do get really good answers. I think they're very different answers, right? Some will really open up and I I I'm not looking for a work work answer like it can be anything in your life that truly the hardest thing and sometimes you learn a lot about people's personal lives that something they had to go through they didn't choose but it was all about how they reacted what they controlled how did they take something they can control in their hands and actually change the future right and sometimes it's about uh you know it also doesn't have to be this gr grandio thing I'll give you an example of someone we just hired actually he's starting uh next week and you know he gave this answer this was the first time someone gave an answer like this he said I live a very simple life I deeply care about my work and I have a crazy regimen about how I think about my health and really outlined a little bit what they meant right like outside of that I don't really have a lot but the hardest thing for me has been I kept up with this for x amount of years every single day like waking up this hour doing these things every other hour at night um showing up for my co co-workers and the hardest thing has been doing that for over 15 years and without getting bored without really thinking about anything else and committing myself to what I care about and I thought that was also a very creative answer right like I'm not really looking for um you know maybe when I gave that answer like my answer is a lot about building a company or about how I came from a very tiny town when my parents didn't have anything to hear but I'm not looking for that everybody's story is different. How did you make your own story? Right. I thought that was a great answer and we're very happy to have him.
我拿到的答案是真的很好。我觉得答案会非常不一样，对吧？有些人会敞开心扉。我要的不是一个「工作上的」答案，可以是你人生里任何真正最难的事。有时候你会因此了解到对方个人生活里的很多东西——某些他们并没选择、却必须经历的事，但重点全在于他们怎么反应、他们能掌控什么、他们怎么把能掌控的部分抓在自己手里、真的去改变未来。而且有时候，它也不必是什么惊天动地的大事。给你举个例子，我们刚招了一个人，下周入职。他给的答案是我第一次听到有人这么答：他说我过的是非常简单的生活，我对我的工作极其上心，我对健康有一套近乎疯狂的作息规律，然后他稍微展开讲了讲具体是什么。他说除此之外我没什么别的，但对我来说最难的事，是我把这套东西坚持了 X 年、每一天都如此——几点起床、几点做什么、晚上每隔几小时做什么，还有为同事负责地出现在那儿。最难的事就是这样坚持了十五年多，没有厌倦，没有真的去想别的，就是把自己交给我在乎的事。我觉得这也是个非常有创造性的答案，对吧？我不是在找那种标准答案。比如换我来答，我的答案会跟做公司有关，或者跟我来自一个很小的镇子、父母来这边时一无所有有关——但我要的不是这个。每个人的故事都不一样，你是怎么写出自己那个故事的？我觉得那是个很棒的答案，我们很高兴他能加入。
[22:25] Elad Gil
That's amazing. How do you think about where you want to be with the company in 5 years or what's your longer term vision for
太棒了。那你怎么看五年后公司要走到哪一步？或者说你更长期的愿景是什么——
[22:31] Elad Gil
what you folks have accomplished, what you're doing? What's your sort of northstar
——关于你们已经做成的、正在做的这些事。你们的北极星是什么？
[22:35] Melisa Tokmak
in terms of our vision? We are building an autonomous enterprise,
说到愿景，我们要造的是一个 autonomous enterprise（自主运转的企业）。
[22:40] Melisa Tokmak
right? We want every single thing in these companies to be handled autonomously with netic except the actual services and the labor itself that human component so that these companies can fully focus on that to deliver and create customer delight. Um that is number one thing with the company. Every single product every single thing that we build on top of our intelligence layer that compounds and connected to the other products is for that purpose. Mhm.
对吧？我们希望这些公司里的每一件事，都由 Netic 自主处理掉，唯独留下真正的服务交付和劳动本身——那个属于人的部分，好让这些公司能百分之百专注在那上面，去交付、去创造让客户惊喜的体验。这是公司最重要的一件事。我们在自己的智能层之上做的每一个产品、每一样东西，都会复利叠加、彼此打通，全都是为了这个目的。嗯。
[23:10] Melisa Tokmak
Um second is you know building a company is very hard and I'm also doing that for the vision but also to work with remarkable people. I think for me as the company grows something I deeply care about how do we keep working with only remarkable people and there there's no one answer. I think every single company that I know that I respect um that even thought about these things for the long term like imagine Palante or SpaceX or notion who really care about craftsmanship and really delivering for their audience for many many years have struggled with this. So for me um deeply thinking about that and achieving that is very important um in the company. So you're serving a lot of industries that are perceived as very slow to adopt technology or not at the cutting edge or you know things that take a long time to you know use new things and you're selling AI solutions and agents to them. Has that been challenging? Has that been straightforward? Like is that perception correct? Like how do you think about your customer base in terms of just like the rate at which they'll adopt new things or try new things?
第二点是，创业本身非常难。我做这件事，一方面是为了那个愿景，另一方面也是为了能和真正出色的人共事。对我来说，随着公司变大，我特别在意的一件事就是：怎么才能始终只和最出色的人一起工作。这件事没有标准答案。我认识并且尊敬的每一家公司——那些真正长期认真想过这个问题的公司，比如 Palantir、SpaceX 或者 Notion，这些真正在意手艺、真正想为自己的用户长期交付价值的公司——在这件事上都挣扎过。所以对我来说，把这件事想透、并且真的做到，在公司里非常重要。
（Elad Gil）你服务的很多行业，外界的印象是技术采纳很慢、不在前沿、要花很久才肯用新东西，而你在向他们卖 AI 方案和 agent。这个过程难吗？还是其实挺顺？这种印象到底对不对？在接受和尝试新事物的速度上，你怎么看你的客户群体？
[24:15] Melisa Tokmak
I think it's a big misconception to think about these industries as old school. Um actually some of the most techforward business focused people owners founders I have met have been in these industries. So first of all we we work with large enterprises so they're extremely value focused and they have to be tech forward right for example to give you a sense you know like just one of the businesses we closed it can be like a half a million contract um and it took from end to end 14 days
把这些行业当成「老派」，我觉得是个很大的误解。实际上，我见过的最懂技术、最有生意头脑的老板和创始人，有一批恰恰就在这些行业里。首先，我们服务的是大型企业，他们对价值极其敏感，也必须走在技术前面。给你个概念：我们签下来的其中一单，合同额可能是五十万美元，而从头到尾只用了 14 天。
[24:48] Melisa Tokmak
and it's not because you know there is you know they I'm not a magic potion and the company is not or just AI is not they're very thoughtful about what they have to do and checking if the value is there right even in their buying uh behavior. Uh that being said, I think the way that I think about these industries is that they are extremely primal and tech forward at the same time. So if you think about a large roofing company, right, is actually they have to have door knockers. So door knockers mean act, you know, going neighborhood to neighborhood thinking about different you know roofs and trying to talk to you about why you should think about a new way or like having solar on your roof, right? But at the same time today in Netc it's a compilation of product from hand handling everything inbound outbound analytics and even delivery of their labor. So we connect to satellite data to be able to think about in different neighborhoods how the hurricanes affect different roofs. How should you think about different materials? should that be autonomously fed into the context of our agents so that not only you're better equipped to handle that conversation with the customer but also you can spot who to go after. So these are not
这并不是因为我有什么灵丹妙药，也不是公司或者 AI 本身有魔法——是他们对自己该做什么想得非常清楚，会认真核验价值到底在不在，连采购行为本身都很审慎。话虽如此，我看这些行业的方式是：它们同时既非常「原始」，又非常前沿。比如一家大型屋顶施工公司，他们必须有 door knocker（挨家敲门的销售），也就是一个社区一个社区地跑，看不同人家的屋顶，跟你聊为什么该换个新做法、或者在屋顶上装太阳能。但与此同时，今天在 Netic 上，产品是一整套组合：从处理全部 inbound、outbound，到分析，甚至到他们施工人力的调度交付。我们接入卫星数据，就能判断不同社区里飓风对不同屋顶的影响、该考虑什么材料；这些能不能 autonomously 喂进我们 agent 的 context 里——这样不仅你跟客户对话时更有底气，还能判断该主动去找谁。所以这些绝不是……
[26:02] Elad Gil
is this is this technology they were using before or is this uh you know and then you adopted it for the netic platform or
这个——这套技术是他们以前就在用的，还是说是你们做出来之后，接进 Netic 平台的？
[26:08] Melisa Tokmak
yeah I think so they were always interested in how do we serve our customers better but they had to figure out how to do that right do we get this data somewhere else or who looks at it do we look at it manually so now imagine with one platform okay not only I am actually assisting with the inbound that's coming from the ads we have to do but at the same time I'm building context about the neighborhoods that I have to serve and reaching out to bring net new revenue. So that is now they can do it very easily in one platform with netic but they were already thinking about these like additional data and sifting through that with different tools or humans etc and then send door knockers before and that's just one industry right and then we talked a little bit about rollups and being private equity owned now imagine being a private equity and you own maybe 20 30 of these these businesses in different industries so now if those are on net what you can do is well if you're going to go after someone for a wellness offering right and I have a dog that I love and if I'm going to be a member in a wellness offering I would actually very much like to pick the one that has pet care right so if you know that about me and as the owner of the company how would you think about talking to me about that that is all about context and really understanding and doing that easily for your business
是这样。我觉得他们一直都很想把客户服务做得更好，只是得自己琢磨怎么实现：数据从哪儿来？谁来看？是不是得靠人工看？现在想象一下，在一个平台里：我不仅在帮你处理投放广告带来的 inbound，同时还在积累我要服务的那些社区的 context，并且主动触达、带来净新增收入。这些事现在用 Netic 一个平台就能轻松做到；但他们本来就已经在琢磨这些额外数据了，只是得靠各种工具或者人力去筛，然后再派 door knocker 出去。而这还只是一个行业。
我们刚才也聊到了 rollup（整合并购）和被 private equity 持有。现在想象你是一家 PE，手里有二三十家分布在不同行业的公司。如果这些公司都在 Netic 上，你能做的是：比如你要推一个健康/健身会员，而我养了一条特别爱的狗，那我挑会员的时候，会非常倾向于选一个带宠物照护的。所以如果你知道我这一点，作为老板，你会怎么跟我聊这件事？这全都是关于 context——真正理解客户，而且让业务方能轻松做到这件事。
[27:25] Elad Gil
how how well do you or how do you view private equity firms as responding to this wave of AI I think um there there are varying opinions in terms So everything from people are really leaning in and they're trying new things and in some cases they're having to deploy companies. They're working with folks like um Brainco or other companies. In other cases they're uh slower to to move on it. What's your what's your perception of like how private equity is interacting with AI?
你怎么看 private equity 机构对这波 AI 的反应？我感觉分歧挺大的：有些人非常投入、什么都愿意试，有些情况下他们甚至得亲自推动被投公司落地，会找 Brainco 这类公司合作；另一些则动作慢得多。你的感受是，PE 现在是怎么跟 AI 打交道的？
[27:49] Melisa Tokmak
I think already the private equity playbook has changed. Before it used to be right we're going to find a gem and that has an amazing multiple and we are going to go change the team and create the value and then sell it again. But those gems don't really exist anymore like undiscovered. So I think generally the playbook has changed to how do we create tangible value with these businesses that we're working. I think you are right that some are leaning in maybe a bit too much like without um you know trying everything as if it's software and you know it that any product I use in AI it's not like you know the first week that shouldn't be how you look at the platform. It should be the beginning of the relationship. If anything actually you want to make sure those results exist throughout the year and keep getting better and not keep getting worse, right? Um so I think that's important to change the shift on testing the AI like AI and separating from vaporware to true ROI and how do you do that? I think a lot of them at times can be more software oriented you know deterministical but why can I not try that like in a week and tell but many I think do understand with this new role that private equity companies hire it's like different operating partners in AI or they'll have a few engineers that might be AI forward and they'll educate it as well and in our job you know I always see in the company too something is not hard is that like what can we do better so in a lot of these conversations. The conversation is all about the value they're going to get and what they can see tangibly now. It's not demos. We'll pull up and show show a live deployment, right? Because like if it's working and I have it, why can't you just show you what this customer is making that you know we have, you ask me about my northstar recently, we have made so far, I think over $600 million for our customers that have been really
我觉得 private equity 的打法其实已经变了。以前是：找到一块被低估的璞玉、倍数很漂亮，然后换掉团队、把价值做出来，再卖掉。但现在这种「还没被发现的璞玉」基本不存在了。所以整体打法变成了：我们怎么在手里这些生意上，真正创造出看得见的价值。
你说得对，有些机构确实用力过猛——把 AI 当成传统软件那样，什么都试一遍。可我自己用任何 AI 产品，都不会只看第一周，那不该是你评估一个平台的方式，那只是关系的开始。反过来说，你真正要确认的是：这些效果能不能贯穿一整年、而且越来越好，而不是越来越差。
所以我觉得关键是改变测试 AI 的方式，把「PPT 产品」和真实 ROI 分开。怎么做到？很多机构的思路有时候还是偏传统软件、偏确定性——「为什么我不能一周内试完就下结论？」但我觉得也有很多机构确实想明白了：PE 现在会招一种新角色，比如专门做 AI 的 operating partner，或者养几个懂 AI 的工程师，同时也做内部教育。而在我们这边的工作里——我在公司内部也一直这样——有件事其实并不难，就是不断问：我们还能把什么做得更好。所以这些对话里，谈的全是他们能拿到什么价值、现在能看到什么实打实的东西。不是 demo，我们会直接拉出一个真实的线上部署给你看。因为既然它已经在跑、我手上就有，为什么不直接让你看看这个客户挣到了多少？你刚才问我北极星指标——到目前为止，我们已经为客户创造了超过 6 亿美元……
[29:44] Melisa Tokmak
generated from AI handled interactions. Show that. how does it work? How is the real world customers interacting with this technology? So I think once you do that they also get educated and they guide their companies but still it doesn't it's is not taking away the notion of you have to have a relationship direct to the company because at the end uh private equity is more of a guidance for the companies they own. There's not many times rel types of relationships where you can just shove it down the throats of the companies, but you can be a good counselor and you can be a good advisor to the companies on what to check for.
……都是由 AI 处理的交互带来的。把这个摆出来看：它是怎么运作的？真实世界里的客户是怎么跟这套技术打交道的？我觉得你一旦这么做，他们自己也被教育了，然后再去指导他们投的公司。但这仍然替代不了「必须和公司本身建立直接关系」这件事，因为归根到底，private equity 更多是给自己持有的公司做指导。很多时候那种关系并不是能硬塞给公司的，但你可以当一个好的顾问、好的参谋，告诉他们该看哪些东西。
[30:20] Elad Gil
Yeah, it seems like a lot of um private equity shops as well have kind of shifted from the 80s style, you know, come in do big layoffs, take apart a conglomerate and much more to your point, how do you optimize value or increase the value of something? I will say though it's still they are always the first conversations very focused on cost cutting because I think they don't see a lot of products or platforms like ours. I'm not really there to cut your costs. Sure that is happening in this way but I'm really interested in this is how you're going to make net new revenue. So that is new. You have to actually start that conversation. You have to show them intangible examples because otherwise it still focuses on how do we get to the bottom line and cut some costs, right? But they have to almost expand their horizon on thinking about what else is possible with AI, right? It would be pretty sad if we used AI only for cost cutting.
是啊，感觉很多 private equity 机构也已经从 80 年代那一套——进来大裁员、把集团拆开卖掉——转向了你说的那种：怎么优化价值、怎么把资产本身做得更值钱。
（Melisa Tokmak）不过我得说，第一轮对话往往还是死盯着降本，因为我觉得他们没见过多少像我们这样的产品或平台。我来这儿不是为了帮你砍成本的。当然从某种意义上成本确实也降了，但我真正感兴趣的是：你要怎么做出净新增收入。这是新东西，你必须主动把话题引到这儿，必须给他们看具体的例子，否则谈来谈去还是围绕怎么改善利润、砍掉一点成本。他们几乎必须拓宽自己的视野，去想 AI 还能带来什么别的可能。如果我们用 AI 只是为了降本，那也太可惜了。
[31:13] Elad Gil
Yeah. Yeah. Yeah. Sure. What do you outside of what NetC is doing? What are you most excited about in terms of the future in AI?
对，是这样。那除了 Netic 在做的事之外，AI 的未来里，你最兴奋的是什么？
[31:18] Melisa Tokmak
I am really excited a little bit about my personal story. So one of the you know when I didn't have access to anything uh one of the in my time maybe the technology that helped is that I would find one or two people who graduated from my high school like to read some of my essays or look at my work and give feedback or Facebook messenger them and it was it was just being able to talk when they were in the US in colleges and they went from Turkey. It was pretty crazy. So like I will give the answer of education because I I think it will change so many lives of people who otherwise would not even think about this world exist. So I'm personally really excited about that and no one will be limited anymore that like everything you want to learn and everything you want to do is kind of in your pocket, right? So we made the world a lot more about agency which I'm excited about. Um, and I cannot wait to see that unfold.
这跟我自己的经历有点关系。当年我什么资源都没有，那个时代真正帮到我的「技术」，就是我能找到一两个从我高中毕业出去的人，请他们帮我看看文书、看看我做的东西，给点反馈，或者在 Facebook Messenger 上跟他们聊。他们从土耳其去了美国上大学，那时候能这样直接对话，感觉挺不可思议的。所以我的答案是教育。我觉得它会改变非常多人的人生——那些原本连「还有这样一个世界存在」都不会想到的人。所以我个人特别期待这件事：以后没有人再会被限制住，你想学的一切、想做的一切，基本都在你口袋里。我们让这个世界变得更取决于个人的能动性（agency），这点让我很兴奋，我迫不及待想看到它怎么展开。
[32:21] Elad Gil
Yeah, it's very exciting. I think uh one big lesson for me is always you can just go do stuff like you don't need permission, you don't need to learn stuff, you can just go do things. So,
对，非常令人兴奋。我自己最大的一个体会一直是：你其实可以直接去做——不需要谁批准，也不用先去学一堆东西，直接干就行了。
[32:29] Melisa Tokmak
but you know that makes it also a little bit scary too then when you see people not doing stuff then it's a little bit of a dystopian version of it because in in any era or technology you live with you can always do stuff. it ends up it stays with the choice you have to make. We keep making it easy and easier easier and easier to make that choice but still I guarantee you majority of the world will not be making that choice. So it is something to something to think about a little bit more negative. So um but I am happy that at least we're taking away the resourcing as a question uh whether you have access to something or not. The second thing I am really excited about a lot in America or in generally right now more mainstream topic about AI is very negative. M
但你知道，这也让人有点害怕——当你看到有人就是不去做的时候，就有点反乌托邦的味道了。因为在任何一个时代、任何一种技术下，你其实都可以去做事，最后还是落回你必须自己做的那个选择。我们在不断把这个选择变得越来越容易、越来越容易，但我敢保证，世界上大多数人依然不会做出那个选择。所以这是件值得多想想的事，稍微有点消极。不过我至少高兴的是，我们正在把「有没有资源、能不能拿到」这个问题从等式里拿掉。
第二件让我很兴奋的事是：现在在美国、或者说整体的主流舆论里，关于 AI 的讨论普遍是负面的。
[33:20] Melisa Tokmak
um what I'm exciting about with use cases like I'm working on it's many times the people's worst day in their lives and they really need help and how do we make sure that you get help or in education or in about being able to look at your health records and really understand you know I was recently using the health feature and trying to understand how can I sleep better and there's no answer it's you got to stop being a founder um
而像我在做的这类用例，很多时候面对的是人们这辈子最糟糕的一天，他们是真的需要帮助——我们怎么保证他们能得到帮助。教育也是；还有能看懂自己的健康记录、真正理解它。我最近就在用健康那个功能，想搞明白怎么才能睡得更好，结果没有答案——答案是「你得先别当创始人了」。
[33:45] Melisa Tokmak
but I think there are a lot more positive the ways that we we are going to start talking about AI and how people's first interaction with AI is not only about like why XY Z is going away or jobs are going away but positive impact of AI that's more in terms of the conversation in the world I'm also excited about
但我觉得，我们谈论 AI 的方式会出现很多更正面的角度：人们第一次接触 AI，不该只是「某某东西要消失了」「工作要没了」，而是 AI 带来的正面影响。世界上关于 AI 的讨论能往这个方向走，也是我很期待的。
[34:05] Elad Gil
amazing yeah well thank you so much for joining us today prior it was a pleasure
太棒了。非常感谢你今天来 No Priors 做客，聊得很愉快。
[34:08] Melisa Tokmak
thank you for having me
谢谢你们的邀请。
[34:12]
find us on Twitter at no prior pod subscribe to our YouTube channel if you want to see our faces follow the show on Apple podcasts, Spotify, or wherever [music] you listen. That way, you get a new episode every week. And sign up for emails or find transcripts for every episode at no-briers.com.
在 Twitter 上关注我们 @NoPriorsPod；想看到我们的脸，就订阅我们的 YouTube 频道；在 Apple Podcasts、Spotify 或者你常用的播客平台关注本节目，这样每周都能收到新一期。也欢迎订阅邮件，或者到 no-priors.com 查看每期的文字稿。