对话29岁的Meta首席AI官：顶级AI竞赛的终局，究竟拼什
字幕摘录
| 时间 | 英文 | 中文 |
|---|---|---|
| 0:00 | It was almost exactly a year ago that you started at Meta leading the AI team and I thought maybe | 差不多一年前 你开始在梅塔领导AI团队 我想也许 |
| 0:06 | you could set the scene for us just a little bit. | 你可以给我们安排一下 |
| 0:09 | I'm wondering if you can let us know | 我想知道你能不能告诉我们 |
| 0:12 | where you think Meta is as an AI company today versus where it was a year ago. | 你觉得Meta今天是AI公司 与一年前的情况相反 |
| 0:17 | Specifically, | 具体来说, |
| 0:18 | what's the reputation of Meta as an AI company today versus a year ago when you got there? | 今天Meta作为AI公司的名声是什么? |
| 0:24 | Yeah, I mean, I think it's been a very exciting year for Meta and AI. | 是啊,我的意思是,我认为 这是一个非常令人兴奋的一年 对于梅塔和AI。 |
| 0:29 | I mean, I think when | 我的意思是,我想当 |
| 0:31 | roughly a year ago, a little more than a year ago, Meta released Llama 4 and while that was | 大约一年前, 一年多前,梅塔 释放了Llama 4,而当时, |
| 0:38 | still an exciting release, I think it wasn't quite on the trajectory that Meta needed to | 仍然令人兴奋的发布, 我认为它并不完全 在轨迹上,梅塔需要 |
展开字幕全文（355 条）
| 序号 | 英文 | 中文 |
|---|---|---|
| 1 | It was almost exactly a year ago that you started at Meta leading the AI team and I thought maybe | 差不多一年前 你开始在梅塔领导AI团队 我想也许 |
| 2 | you could set the scene for us just a little bit. | 你可以给我们安排一下 |
| 3 | I'm wondering if you can let us know | 我想知道你能不能告诉我们 |
| 4 | where you think Meta is as an AI company today versus where it was a year ago. | 你觉得Meta今天是AI公司 与一年前的情况相反 |
| 5 | Specifically, | 具体来说, |
| 6 | what's the reputation of Meta as an AI company today versus a year ago when you got there? | 今天Meta作为AI公司的名声是什么? |
| 7 | Yeah, I mean, I think it's been a very exciting year for Meta and AI. | 是啊,我的意思是,我认为 这是一个非常令人兴奋的一年 对于梅塔和AI。 |
| 8 | I mean, I think when | 我的意思是,我想当 |
| 9 | roughly a year ago, a little more than a year ago, Meta released Llama 4 and while that was | 大约一年前, 一年多前,梅塔 释放了Llama 4,而当时, |
| 10 | still an exciting release, I think it wasn't quite on the trajectory that Meta needed to | 仍然令人兴奋的发布, 我认为它并不完全 在轨迹上,梅塔需要 |
| 11 | be able to continue building a lot of the products and experiences that it seek to build | * 能够继续建设它寻求建设的许多产品和经验; |
| 12 | and so we've been hard at work over the past year since I joined and since starting Meta | 自从我加入和开始梅塔公司以来 我们一直在努力工作 |
| 13 | Superintelligence Labs. | 超级智能实验室. |
| 14 | We've been undergoing an entire process of building a new scaling | 我们一直在经历一个全过程 建立一个新的规模 |
| 15 | ladder for our models, developing new sets of infrastructure and new research to be able to | 开发新的基础设施和新研究,以便能够 |
| 16 | power sort of a new series and family of models and back in April, we released the very first | 4月份,我们发布了第一个 |
| 17 | 新的Spark模型和Meta AI的更新 | |
| 18 | The reception to all of that | 对所有这些的接待 |
| 19 | was incredibly positive. | 令人难以置信的正面。 |
| 20 | It was even better than frankly we had expected internally. | 这比我们内部所期望的要好得多。 |
| 21 | We saw | 我们看见了 |
| 22 | incredible gains in usage of Meta AI. | 在使用Meta AI方面令人难以置信的进步. |
| 23 | It was at the top of many of the app stores and we're sort | 它在很多应用商店的顶端 我们排序 |
| 24 | of working on our even larger models today and are excited about what we'll be able to demonstrate | 并且对我们能够展示的东西感到兴奋 |
| 25 | to the world. | 给世界。 |
| 26 | We're on a very exciting and fast trajectory and we're excited to continue showing | 我们正在一个非常刺激和快速的轨道上 我们兴奋地继续显示 |
| 27 | the world what we produce. | 我们创造的世界 |
| 28 | I think that the AI industry has gotten very hot and competitive | 我认为AI产业已经变得非常热门和有竞争力 |
| 29 | obviously in the past year as well and we take that very seriously but we're really excited. | 很明显,在过去一年里也是如此,我们非常认真地对待这一点,但我们真的很兴奋。 |
| 30 | Obviously the goal you're spending in the same way that an open AI, Anthropic, others are spending, | 显然你所花费的目标 与开放的AI,Anthropic,其他人所花费的目标相同, |
| 31 | do you feel like you guys are in that same tier? | 你觉得你们在同一层吗? |
| 32 | 是现在打开的AI,Anthropic,Meta | |
| 33 | do you feel like there's still a bit of a gap? | 你觉得还有些差距吗? |
| 34 | Because I'd say certainly a year ago that was the | 因为我会说,肯定 一年前,那是 |
| 35 | perception. | 感觉。 |
| 36 | The new Spark model that we released is not at the tier of the leading frontier models | 我们发布的新Spark模型没有处于领先前沿模型的层次 |
| 37 | but we believe it's a very exciting data point on the trajectory and we expect the upcoming | 但是我们认为这是一个非常令人兴奋的数据点 在轨迹上,我们期待即将来临 |
| 38 | models we release to be quite competitive with the leading models in the world. | 与世界领先的模型竞争 |
| 39 | You called | 你打电话来 |
| 40 | new Spark an appetizer, an appetizer model. | 新的Spark开胃菜,开胃菜模型. |
| 41 | To build that metaphor out, when does the entree | 要建立这个比喻,什么时候要求 |
| 42 | model get here? | 模特儿来了? |
| 43 | Will it be at that tier that you guys ultimately want to be at? | 你们到底想在那个级别吗? |
| 44 | Yeah, we are in process of cooking. | 是的,我们正在做饭。 |
| 45 | The whole conversation is just going to be this back and | 整个对话只是要 回来, |
| 46 | forth. | 离开。 |
| 47 | Yeah, no, I mean we're cooking it. | 是啊,不,我的意思是,我们正在做它。 |
| 48 | We're excited to show it to the world once it's ready. | 我们很兴奋,一旦它准备好就把它展示给世界. |
| 49 | We're seeing very exciting and promising results in the process of training it right now. | 我们现在在培训过程中看到非常令人兴奋和有希望的结果。 |
| 50 | We're | 我们是 |
| 51 | quite excited about it and I think overall we built the entire research effort around | 相当兴奋,我想总的来说,我们 构建了整个研究工作围绕 |
| 52 | predictable scaling. | 可预见规模。 |
| 53 | The entire belief of our overall research effort was that in many ways the | 我们整个研究工作的全部信念是,在许多方面, |
| 54 | central belief behind the current modern AI boom is that as you scale these models you will see | 目前现代AI热潮背后的核心信念是,当你缩放这些模型时,你会看到 |
| 55 | get incredible results and get predictable levels of increased capability. | 获得令人难以置信的结果,并获得可预测的能力提高水平。 |
| 56 | New Spark was an early data point on that scaling ladder for us. | "新火花"是一个早期的数据点 对于我们来说,这个缩放梯子。 |
| 57 | The next models we release will be | 我们公布的下一个模型是 |
| 58 | an even greater point on the scaling curves and we are really excited to show the world what we'll | 我们非常兴奋地向世界展示我们将会做什么 |
| 59 | be able to produce. | 能够生产。 |
| 60 | Obviously you went from basically rebuilding the team, rebuilding the | 显然你从基本重建团队,重建 |
| 61 | lab to this new Spark model in a very short amount of time. | 在很短的时间内,实验室将使用这个新的Spark模型。 |
| 62 | What is the biggest barrier | 什么是最大的障碍 |
| 63 | from getting to the appetizer to the next level? | 从开胃菜到下一关? |
| 64 | I assume it's not resources. | 我想这不是资源 |
| 65 | You guys are spending | 你们要花钱 |
| 66 | a ton of money on this effort. | 一大笔钱 |
| 67 | Is it just simply time? | 只是时间问题吗? |
| 68 | Is it talent? | 是天赋吗? |
| 69 | What's going to bring | 将会带来什么 |
| 70 | your models to that frontier? | 你的模特儿去那个边境? |
| 71 | Yeah, we've talked about some of this in some of our public | 是啊,我们已经谈论了一些 我们的一些公众 |
| 72 | blog posts and whatnot, but it's about continuing to scale the data, the compute going into the | 博客贴文和什么的, 但它是关于继续 放大数据,计算进入 |
| 73 | models, as well as continuing to scale with research, so continue to drive advances in | 以及继续以研究方式扩大规模,因此继续推动 |
| 74 | underlying research breakthroughs to continue driving forward the progress of the models, | 继续推进模型进展的根本研究突破, |
| 75 | and building infrastructure to support all of this. | 并建设基础设施来支持所有这一切。 |
| 76 | This is in many ways a year where | 这在很多方面是一年 |
| 77 | all of the labs are dramatically scaling up their models and we think of a much faster trajectory | 所有实验室都在大幅提升模型 我们想到一个更快的轨道 |
| 78 | to do so because obviously we've been doing all this work over the course of the past year. | 这样做,显然是因为 我们在过去一年中一直在做所有这些工作。 |
| 79 | Yeah, we need to build the infrastructure. | 是的,我们需要建设基础设施。 |
| 80 | We need to scale the data, scale the compute, | 我们需要放大数据 放大计算 |
| 81 | train these large models, and show them to the world. | 训练这些大模型, 并展示给世界。 |
| 82 | Yeah. | 对 |
| 83 | I want to talk about model | 我想谈谈模特儿的事 |
| 84 | strategy a little bit. | 有点策略 |
| 85 | Before you got to meta, everything was open source. | 在你得到元之前,一切都是开源的. |
| 86 | That was definitely | 这绝对是 |
| 87 | the overarching strategy. | 总体战略。 |
| 88 | New Spark model is not open source. | 新的 Spark 模式不是开源 。 |
| 89 | I believe I heard you on a | 我相信,我听到你在一个 |
| 90 | prior interview basically say that as you guys were testing it, it didn't feel safe to open | 之前的采访基本上说,当你们测试的时候,打开它感觉不安全 |
| 91 | source. | 来源. |
| 92 | Can you go deeper on sort of what you mean by that and how you made that decision? | 你能更深入地了解你的意思 以及你是如何做决定的吗? |
| 93 | Yeah, so one of the things that we did as part of meta super intelligence labs is we updated | 是啊,所以我们做的一件事 作为元超智能实验室的一部分 是更新 |
| 94 | our what we call our advanced AI scaling framework, which is really our view of what are the risks | 我们称之为先进的AI缩放框架, 这实际上是我们对风险的看法 |
| 95 | that we see in developing these very powerful models and how do we want to handle those risks | 我们在开发这些非常强大的模型时看到,我们想如何处理这些风险。 |
| 96 | as we see them in early testing. | 正如我们在早期测试中看到的 |
| 97 | We published a lot of what we saw in the process of training | 我们在训练过程中看到很多东西 |
| 98 | Spark in our preparedness report. | 我们的备战报告火花 |
| 99 | Some of the things that we saw is that it actually triggered | 我们看到的一些事情是 它实际上触发了 |
| 100 | some high risk areas in the course of early training, particularly around bio risk, | 早期培训过程中的一些高风险领域,特别是生物风险领域, |
| 101 | but also a number of the risks were elevated. | 但也增加了一些风险。 |
| 102 | This is something I think the entire industry | 我觉得这是整个行业的事 |
| 103 | has seen as the models have improved pretty dramatically over the past year. | 这些模式在过去一年中得到了相当显著的改进。 |
| 104 | We certainly | 当然 |
| 105 | aren't the only ones to see a host of these risks show up as we scaled up the models and as we | 并不是唯一一个看到这些风险 出现在我们扩大模型和我们 |
| 106 | kept pushing the frontier of research, but we saw these risks trigger and we realized I think | 不断推动研究的前沿, 但我们看到了这些风险触发,我们意识到,我想 |
| 107 | fundamentally those when we launched a model like new Spark in a product, we have a lot of ways to | 基本上,当我们推出一个模型 像一个新的火花 在产品,我们有很多方法 |
| 108 | mitigate some of these risks and ensure that we're able to launch it in a safe and responsible way. | 并且确保我们能够以安全而负责的方式发射 |
| 109 | It's much harder to do that when you open source a model and people can use that model in all | 当你打开一个模型的源头 并且人们可以使用这个模型时,做这个就困难得多了 |
| 110 | sorts of contexts that we may not have full understanding of. | 我们可能没有充分理解的某种背景。 |
| 111 | So we're in the process right | ┮и《笆琌癸 |
| 112 | now of developing models that we believe are fit and safe to be open source while still maintaining | 现在开发的模型 我们认为是合适的和安全的 开源的同时仍然保持 |
| 113 | as much of the performance capabilities as possible. | 尽可能多的发挥能力。 |
| 114 | So you will still do open source. | 所以你还是会做开源。 |
| 115 | It sounds like llama, though, is not the brand or the pillar that you're going to do. | 这听起来像拉玛,虽然, 不是品牌或支柱 你要去做。 |
| 116 | Is everything | 全部都是这样 |
| 117 | open source going to be new Spark or adjacent? | 开源将是新的 Spark 还是相邻的 ? |
| 118 | You know, we have exciting debates about branding | 你知道,我们有令人兴奋的辩论 关于品牌 |
| 119 | internally and nothing to share right now, but yeah. | 内部和没有 分享现在,但是啊。 |
| 120 | Okay. | 摆 |
| 121 | The big models that are coming that | 未来的大模特儿 |
| 122 | you've hinted at, give us a general sense if you can. | 你已经暗示了, 给我们一个一般的感觉,如果你可以。 |
| 123 | Obviously, you know, each model, each company | 显然,你知道,每个模型,每个公司 |
| 124 | is perhaps known for certain things. | 也许以某些事物而闻名。 |
| 125 | Do you feel as though you're moving in a direction where | 你觉得你正在朝着一个方向前进吗? |
| 126 | Meta's models are going to be known for, you know, best in class at X versus Y? | 梅塔的模特儿们将会以在X对Y的班级成绩最好而闻名? |
| 127 | Like what are you | 喜欢什么你 |
| 128 | hoping to accomplish with what you guys come out with next? | 希望你们能完成接下来的任务? |
| 129 | So already in new Spark, some areas | 因此,已经在新的 Spark, 一些区域 |
| 130 | where we were really impressed by the capabilities, even though it was, again, like a much smaller | 在那里,我们真的印象深刻 能力,即使它,再次, 像一个小得多 |
| 131 | model than ultimately we intend to train, were around multimodality capabilities, so its ability | 而不是我们最终打算训练的模型, 围绕着多式联运能力, 所以它的能力 |
| 132 | to handle images, video, audio, and that's obviously very important and critical for Meta's | 处理图像,视频,音频, 这显然非常重要和关键 对于Meta's |
| 133 | business. | 业务。 |
| 134 | Also its capabilities in health were really impressive and that was very exciting for | 它的卫生能力也令人印象深刻 令人兴奋 |
| 135 | us. | 我们 |
| 136 | You know, health is an area that we view as really critical as we scale these models out to | 你知道,健康是一个领域 我们认为真正关键 因为我们推广这些模型 |
| 137 | billions and millions of people all around the world. | 百亿亿世界众生. |
| 138 | And then also a lot of the early results | 还有很多早期的结果 |
| 139 | we saw in the ability of the model to create, you know, vibe code and create little games or | 我们从模型的能力中看到了 创造,你知道, 气氛代码和创建小游戏或 |
| 140 | artifacts or whatnot were very powerful. | 文物或什么的 是非常强大的。 |
| 141 | So we are doubling down on some of these and continuing | 所以,我们正在加倍 在一些这些和继续 |
| 142 | to invest into the agentic capabilities of the model. | 投资于模型的代理能力。 |
| 143 | So we're really excited for the upcoming | 所以,我们真的很兴奋 对于即将来临 |
| 144 | models to be released to be very, very capable agents paired with a lot of these other strengths | 释放的模型 是非常,非常有能力的剂 与很多其他的优势配对 |
| 145 | around multimodality, around health and many others. | 环绕着多式联运,环绕着健康和许多其他。 |
| 146 | And ultimately what we're really excited | 最终我们感到兴奋 |
| 147 | to build for the world are the best personal agents for everybody around the world as much as | 为世界而建设是全世界每个人最好的个人代理,就像 |
| 148 | possible. | 也许吧 |
| 149 | I want to get to agents in just a second because you just made some news on agents | 我想马上找探员 因为你刚给探员做了些新闻 |
| 150 | actually yesterday. | 其实是昨天 |
| 151 | You have other stuff in the works that we can talk about. | 作品中还有其它东西我们可以谈. |
| 152 | But before we pivot | 但是在我们开始前 |
| 153 | off safety real quick, I want to ask a question just about China. | 我要问一个中国的问题 |
| 154 | You've been, you know, you've | 你已经,你知道,你已经 |
| 155 | talked publicly about the risks and the threats of AI coming out of China. | 公开谈论AI出中国的风险和威胁. |
| 156 | You guys have also | 你们也有 |
| 157 | trained on some Chinese open source models. | 对一些中国开源模型进行了培训。 |
| 158 | I'm just wondering, can you give us a sense of like | 我只是在想 你能不能给我们一点感觉 |
| 159 | how you view China right now in the, you know, through the lens of AI? | 你现在如何看待中国 在,你知道, 通过AI的镜头? |
| 160 | Is it a threat in the | 这是威胁吗? |
| 161 | way that we've heard historically? | 我们从史上听说的? |
| 162 | Do you feel like that's changed? | 你觉得这改变了吗? |
| 163 | Yeah, I think that, I think | 是的,我觉得,我想 |
| 164 | it is incredibly important for the United States to lead on technology and the economic benefits | 对美国来说,在技术和经济效益方面起带头作用极为重要。 |
| 165 | that can be created from AI. | 从AI可以创建。 |
| 166 | I think this is very, very critical. | 我认为这是非常非常关键的。 |
| 167 | If you look at the history of | 如果你看历史 |
| 168 | civilization, you know, technological advances are very important for countries or civilizations to | 技术进步对于各国或文明 |
| 169 | adapt to and be able to adopt. | 适应并能够收养。 |
| 170 | And that really defines, you know, really truly the course of | 这真正地定义了,你知道, 真正的过程 |
| 171 | history over long arcs. | 长弧的历史. |
| 172 | So I think it's very important that the United States is able to lead | 因此,我认为非常重要的是 美国能够领导 |
| 173 | on AI. | 关于大赦国际的报告。 |
| 174 | And that's a huge part of our focus at Meta as well is ensuring that we are able to contribute | 这也是我们在梅塔的一大重点 也是确保我们能贡献 |
| 175 | to the United States leading. | 到美国领先。 |
| 176 | Where do you think we are right now as a country? | 你觉得我们现在作为一个国家在哪里? |
| 177 | Is the U.S. leading? | 美国领先吗? |
| 178 | I think right now the U.S. is leading, yes. | 我想现在美国正在领先 |
| 179 | And I think it's, you know, this is one of these | 我认为这是,你知道, 这是其中之一 |
| 180 | situations where you, you know, it's important for us always to track progress from many other | 在你,你知道, 这对我们很重要 总是跟踪从很多其他的进展 |
| 181 | countries, but especially China, you know, be very thoughtful and understand exactly | 特别是中国,你知道的, 深思熟虑和理解准确 |
| 182 | what's happening within each country and what are the reasons those things are happening. | 每一个国家发生的事情 以及那些事情发生的原因 |
| 183 | But | 不过 |
| 184 | I think right now we're ahead. | 我想现在我们领先了 |
| 185 | And what, sorry, I said I was going to get the agents, but one more. | 什么,对不起,我说我要得到 特工,但还有一个。 |
| 186 | What could put that at risk? | 怎么会有风险? |
| 187 | What could put that lead at risk? | 是什么让铅处于危险之中? |
| 188 | What's the most threatening | 什么是最危险的 |
| 189 | thing to stop that? | 能阻止这一切吗? |
| 190 | That's a good question. | 问曰. |
| 191 | I mean, ultimately I think we are in a phase where | 我的意思是,最终 我认为我们正处于一个阶段 |
| 192 | the research advancements industry-wide that we're seeing from continuing to scale these | 研究进步了整个工业 我们从继续扩大这些 |
| 193 | models, apply more compute, apply more data to these models, are just incredibly exciting. | 模型,应用更多的计算,应用更多的数据 这些模型,只是令人难以置信的兴奋。 |
| 194 | In some ways the progress and pace of research today is nearly miraculous. | 在某些方面,今天的研究进展和速度几乎是奇迹。 |
| 195 | And so I think that | 所以我觉得 |
| 196 | it's important that we're able to continue this pace of progress, that we're able to continue, | 重要的是我们能够继续这个进步的速度, 我们能够继续, |
| 197 | you know, being able to continue scaling these models. | 能够继续推广这些模型 |
| 198 | Agents now, finally. | 特工们终于来了 |
| 199 | You guys just yesterday, I believe it was announced, a business agent, | 你们昨天,我想是宣布的,一个经纪人, |
| 200 | so advertisers can use this to interact with customers, I presume eventually help even | 广告商可以用这个与顾客互动,我猜最终会帮助甚至 |
| 201 | develop ad campaigns, things like that. | 发展广告宣传,类似的东西。 |
| 202 | But you're also developing a consumer agent. | 但你也在开发一个消费者代理。 |
| 203 | Talk me through your vision for how ultimately agents will reach all of us. | 让我透透你的远见吧 探险家最终会怎样接近我们所有人 |
| 204 | Like, I guess I'm | 喜欢,我想我 |
| 205 | wondering, is an agent going to be similar to like my email address where I have one core agent and | 好奇的是,一个代理会 类似我的电子邮件地址 在那里我有一个核心代理和 |
| 206 | maybe a secondary agent? | 也许是副特工? |
| 207 | Or is it going to be like the apps on my phone where I have one agent for | 还是会像手机上的应用程序 我有一个经纪人 |
| 208 | every single task in my life? | 我一生的每一个任务? |
| 209 | Like, what do you envision we're going to be using as a society? | 喜欢,你设想什么 我们将作为一个社会? |
| 210 | Yeah, I, you know, we ultimately think it'll probably land somewhere in between those. | 是啊,我,你知道,我们最终认为 它可能会降落 在中间的某个地方。 |
| 211 | I think, | 我觉得 |
| 212 | you know, we really believe that people are probably going to have one, maybe two, maybe a | 你知道,我们真的相信 人们可能会有一个,也许两个,也许一个 |
| 213 | small handful of agents that they rely on. | 他们依赖的少量特工 |
| 214 | And maybe they have a personal agent that's focused on | 也许他们有个人特工 专注于 |
| 215 | things like their health and maintaining their personal relationships and, you know, helping them | 比如他们的健康和保持他们的个人关系 以及帮助他们 |
| 216 | be a better parent and be better with their friends and family. | 做一个更好的父母 并更好的与朋友和家人在一起。 |
| 217 | And then, you know, perhaps | 然后,你知道,也许 |
| 218 | they use that same agent in their work lives, especially if they're working in a small business | 他们在工作中使用同样的代理, 特别是如果他们在一家小企业工作 |
| 219 | or they're an entrepreneur or, you know, working within a smaller organization. | 或者他们是一个企业家,或者,你知道,在一个较小的组织内工作。 |
| 220 | And then, you know, | 然后,你知道吗, |
| 221 | maybe there's worlds where if you work within a larger enterprise or a larger company, then | 也许有世界 如果你在一个更大的企业或一个更大的公司工作, 那么 |
| 222 | these become bifurcated and separated. | 它们相互交织和分离 |
| 223 | Not too much unlike email, let's say, or plenty of other, | 和电子邮件没什么大不了的 比如说 还有其他的 |
| 224 | you know, key technologies that we use from a day-to-day basis. | 我们日常使用的关键技术 |
| 225 | So, you know, we think that | 所以,你知道,我们认为 |
| 226 | ultimately, it should be, you know, agents will be something that become deeply personal | 最终,它应该是,你知道, 特工会成为 深深的私人 |
| 227 | and should be things where over time you find yourself being able to rely on them more and more | 应该是你发现自己越来越依赖他们的东西 |
| 228 | and more for more and more of your personal life, more and more of your work life. | 以及你越来越多的个人生活, 你越来越多的工作生活。 |
| 229 | And that'll be a | 这将是一个 |
| 230 | process that, you know, all society goes through together. | 过程,你知道,所有的社会 一起经历。 |
| 231 | Do you feel that meta in particular, | 你是否特别觉得 |
| 232 | and again, a lot of this happened before you got there, but long history of privacy-related issues, | 再说一遍,很多事情发生在你到那里之前 但是与隐私有关的问题的悠久历史 |
| 233 | are people going to be willing to, you know, trust a meta agent with the personal tasks of | 人们会愿意,你知道, 相信一个元代理人 个人的任务 |
| 234 | their life that you're describing? | 你描述他们的生活? |
| 235 | Yeah, I think that this is like one of the most important | 是的,我觉得这就像一个最重要的 |
| 236 | societal question for agents writ large. | 对特工们来说 社会问题很严重 |
| 237 | I mean, I think that there's incredible amount of innovation | 我的意思是,我觉得有 令人难以置信的创新量 |
| 238 | and technologies we built out on things like agentic safety, you know, ensuring that these | 和科技,我们建立的东西 像代理安全,你知道,确保这些 |
| 239 | agents are respectful to your privacy, ensuring that they're respectful of your boundaries, | 特工们尊重你的隐私 确保他们尊重你的边界 |
| 240 | and continuing to design products in a way that are able to support that. | 并继续以能够支持的方式设计产品。 |
| 241 | So this is definitely | 这绝对是 |
| 242 | something that we're taking very, very seriously and we're being quite thoughtful about. | 我们非常,非常严肃地对待这件事, 并且我们非常深思熟虑。 |
| 243 | And | 还有 |
| 244 | ultimately, we're excited to, you know, show the world what we've built. | 最终,我们兴奋地,你知道, 向世界展示我们所建造的东西。 |
| 245 | But yeah, we think that | 但是,是的,我们认为 |
| 246 | this is not even just a meta problem. | 这不仅仅是一个元问题。 |
| 247 | This is an industry-wide problem as we build more and more | 这是一个全行业的问题 随着我们建设越来越多的 |
| 248 | powerful agents. | 强大的特工 |
| 249 | I think it is a redefinition of, you know, humans' relationship with technology in | 人类与科技的关系 |
| 250 | work through together. | 一起通过。 |
| 251 | How soon will we see a meta consumer agent? | 我们多久才能见到一个元消费者代理? |
| 252 | We are actively cooking it, cooking the entree. | 我们正积极烹饪它, 烹调主菜。 |
| 253 | But no, this is like, I think this is one of the | 但是,不,这是喜欢, 我认为这是其中之一 |
| 254 | things that was very exciting for us internally about the MuSpark launch and MetaEye is that, | 我们内部非常兴奋 关于MuSpark的发射和MetaEye是, |
| 255 | you know, we, even when those launched, we were cooking things internally, both the larger models | 你知道,我们,即使那些启动, 我们正在做东西的内部, 两个更大的模型 |
| 256 | as well as some of these products that you referred to that we're, if anything, more excited about | 以及你提到的这些产品 我们更兴奋的是 |
| 257 | than what we came out with in April. | 比我们四月出来时所见所闻 |
| 258 | Okay. | 摆 |
| 259 | And how are you using agents right now? | 你现在如何利用特工? |
| 260 | Your boss, | 你的上司,你现在... |
| 261 | Mark Zuckerberg, it's been reported he has like a Zuck bot, essentially, or various versions of | 马克・扎克伯格,据报道 他像一个扎克机器人, 本质上,或各种版本 |
| 262 | agents that he's tasking some of his CEO duties to. | 代理,他的任务 他的一些首席执行官的职责。 |
| 263 | Is there an Alex bot that's doing part of your | 是否有一个亚历克斯机器人 做你的一部分 |
| 264 | job right now while we're on stage? | 现在在舞台上工作? |
| 265 | Well, I definitely use agents to support and help me | 嗯,我肯定使用代理 支持和帮助我 |
| 266 | in my work a lot. | 我经常工作 |
| 267 | I think that in many ways, you know, being a leader within a company is really | 我认为,在许多方面,你知道, 领导一个公司,真的 |
| 268 | about how well are you able to, you know, understand everything that's happening at the | 关于你有多好,你知道, 了解一切发生于 |
| 269 | company to the best you can and help, you know, support your team and being able to continue to | 尽你所能 帮助你 支持你的团队 继续 |
| 270 | execute better and better. | 执行更好和更好的。 |
| 271 | So I think there's all sorts of things that agents can do to help you | 所以我觉得有各种各样的事情 特工可以帮你 |
| 272 | there. | 这里 |
| 273 | But I think the uses of agents that are probably most exciting to me are the ones where | 但我认为使用 可能最让我兴奋的特工 |
| 274 | the ways I use them in my personal life. | 我个人生活中用它们的方式 |
| 275 | So, you know, I use an agent to help me be healthier. | 所以,你知道,我用一个代理 来帮助我更健康。 |
| 276 | And I use an agent to help me keep in touch with my friends and like ensure that I maintain those | 我用一个经纪人来帮助我保持联系 和我的朋友,喜欢确保我保持这些 |
| 277 | relationships. | 关系 |
| 278 | And I think these are use cases where they're like very notably different from | 我认为,这些是使用的案例 他们喜欢非常显著的区别于 |
| 279 | the world where I didn't use an agent at all. | 我根本没用过经纪人的世界 |
| 280 | Like these are things that I think are like hard, | 像这些东西 我认为是很难的, |
| 281 | have been hard historically for me to like stay on top of, you know, both my health as well as | 从历史上来说,我很难 喜欢留在上面,你知道,我的健康和 |
| 282 | to help support you do a good job. | 帮助你做一个好工作。 |
| 283 | Those things have been pretty transformation for me. | 对我来说,那些东西是相当的转变。 |
| 284 | Yeah. | 对 |
| 285 | Somehow we only have a couple of minutes left, but I want to ask some sort of bigger | 不知怎的,我们只剩几分钟了 但我想问一个更大的问题 |
| 286 | picture questions about AI and society. | 描绘AI和社会的问题。 |
| 287 | But I actually want to start at Meta, right? | 但我真的想从梅塔开始 对吧? |
| 288 | So | 这么说 |
| 289 | your team is getting an immense investment. | 你的团队得到了巨大的投资。 |
| 290 | You guys are investing hundreds of billions of | 你们正在投资上千亿 |
| 291 | dollars in AI writ large. | 人工智能中的钱 |
| 292 | At the same time, there were layoffs at the end of last month. | 同时,上个月末出现了裁员现象. |
| 293 | So there | 就这样 |
| 294 | offset that. | 冲抵。 |
| 295 | As the person in charge of AI, I'm just wondering how do you deal with that reality, | 作为AI的负责人,我只是想知道你如何处理这个现实, |
| 296 | right? | 对吧? |
| 297 | That you are working on a product that you're excited about, but at the same time, | 你正在研究你兴奋的产品 但与此同时 |
| 298 | you know, the company is saying, hey, this is costing jobs as well. | 你知道,公司是说, 嘿,这是成本 工作也。 |
| 299 | Like how does that make | 就像这样 |
| 300 | you feel? | 你感觉到了吗? |
| 301 | How do you deal with that internally? | 你怎么处理内部的? |
| 302 | I'm just wondering how you, you know, deal with | 我只是想知道你,你知道,如何处理 |
| 303 | that reality. | 那个现实 |
| 304 | It's incredibly difficult to say goodbye to teammates. | 很难跟队友说再见 |
| 305 | And I think it's, you know, | 我认为,你知道, |
| 306 | it's no, I think it's well known that this is a challenging thing to go through as a team and | 不,我认为这是众所周知的 这是一个挑战性的事情 作为一个团队, |
| 307 | something that, you know, is important to acknowledge. | 有些东西,你知道, 重要的是承认。 |
| 308 | You know, we are ultimately really | 你知道,我们最终真的 |
| 309 | excited about the progress that we're making in AI and the products that we're building. | 对AI和产品的进展感到兴奋 |
| 310 | And, | 还有, |
| 311 | you know, we're excited to bring those to the world. | 你知道,我们很兴奋 把这些带到世界。 |
| 312 | But yeah, I think that there's, you know, | 但是,是的,我认为有,你知道, |
| 313 | running a large company is very complex and, you know, we're working through a lot of those | 经营一家大型公司非常复杂 而且,你知道,我们正在通过很多 |
| 314 | issues, but, you know, we don't take any of it lightly, but, you know, at the end of the day, | 问题,但是,你知道, 我们不轻视任何它, 但是,你知道,在一天结束时, |
| 315 | we're excited about what we're building. | 我们为我们建造的东西感到兴奋。 |
| 316 | I think job loss in general is probably like one of the | 我觉得一般的失业 可能像其中之一 |
| 317 | biggest fears with AI. | 与AI最大的恐惧。 |
| 318 | What's your view on that? | 你对此有什么看法? |
| 319 | Is, if AI goes the way that everyone envisions | 是,如果AI走的方式 每个人都设想 |
| 320 | and we reach super intelligence, is there a world in which that can operate and live alongside all | 我们到达超级智能, 是一个可以运作和生活的世界, |
| 321 | of us staying employed? | 我们继续工作? |
| 322 | I mean, I think it's something that we should pay a lot of attention | 我的意思是,我认为这是一些 我们应该多注意 |
| 323 | to, and I think that we should, you know, track closely and try to understand what the impacts are. | 我认为我们应该,你知道, 密切跟踪和尝试了解 影响是什么。 |
| 324 | I think one thing, though, that we never, you know, I think rarely talk about, but is also | 我想有件事,虽然, 我们从来没有,你知道, 我想很少谈论, |
| 325 | happening that's very exciting is that AI is enabling the creation of more businesses than | 令人兴奋的是 AI正在创造更多的企业 比 |
| 326 | ever before in the world. | 从来没有在世界上。 |
| 327 | And, you know, we see this in our data. | 而且,你知道,我们从数据中看到这一点。 |
| 328 | I'm sure many companies see this | 我相信很多公司都看到了这个 |
| 329 | in their data. | 在他们的数据。 |
| 330 | There are more new companies being started today, like, through use of AI tools than | 今天通过使用AI工具创建的新公司比 |
| 331 | ever before, and those numbers are only growing. | 这些数字在不断增长。 |
| 332 | And I expect as AI tools become more and more | 我期望随着AI工具变得越来越多 |
| 333 | powerful, we'll see more entrepreneurs, more small businesses being started. | 强大的,我们会看到更多的企业家, 更多的小企业开始。 |
| 334 | And so I think | 所以我觉得 |
| 335 | there's, you know, the economy is a complex, you know, machine, and one of the things that we see | 有,你知道,经济是一个复杂的,你知道,机器, 以及我们看到的一件事 |
| 336 | within our data today is that there are more small businesses being started, there are more | 在我们今天的数据中, 越来越多的小企业正在开办, 更多的小企业正在开办。 |
| 337 | entrepreneurs, there's more opportunity for entrepreneurs, and so, you know, and we're really | 企业家们,企业家们有更多的机会, 所以,你知道,我们真的 |
| 338 | excited about that. | 很高兴见到你 |
| 339 | We're excited about supporting small businesses throughout the world. | 我们对支持世界各地的小企业感到兴奋。 |
| 340 | Is it the kind of thing that you feel, we saw the Trump executive order just this week | 这是你感觉到的吗,我们看到了特朗普这个星期的行政命令 |
| 341 | where the administration wants to sort of review some of these models before they're released. | 行政部门想在这些模型发布之前 对这些模型进行审查 |
| 342 | I'm curious what you thought of that, first of all, but two, is there another type of regulation | 我好奇你的想法,首先, 但有两个,有另一种类型的规范 |
| 343 | that you think actually might be helpful to prevent, you know, job loss or prevent a world | 可能有助于防止 失业 或防止一个世界 |
| 344 | in which AI is taking priority over the humans? | AI在其中占据比人类优先的地位? |
| 345 | Yeah, I think, well, I mean, first of all, I think | 是啊,我想,嗯,我的意思是, 首先,我认为 |
| 346 | this is a really important and powerful technology, and so I think it's great that | 这是一个真正重要和强大的技术, 所以我认为这是伟大的 |
| 347 | the administration is, you know, deeply considering how we should be thinking about this technology, | 行政部门在深思熟虑 我们该如何思考这个技术, |
| 348 | how we should think about responsible deployment and what all that entails, so I think that's, | 我们应该如何考虑 负责任的部署 以及所有的一切, 所以我认为, |
| 349 | broadly speaking, I think it's been really great that this has been an issue that the administration | 从广义上讲,我觉得这真的很棒 这个问题是政府 |
| 350 | has been really involved on and very thoughtful of, and you don't think it slows innovation to | 而且你并不认为创新会减慢 |
| 351 | have a review process like that. | 有这样的审查程序。 |
| 352 | Well, I think it's always a balance, you know, regulation is | 嗯,我觉得这总是一个平衡, 你知道,监管是 |
| 353 | notoriously difficult to get right in almost all contexts, but I think that, you know, | 在几乎所有情况下都很难正确,但我想,你知道, |
| 354 | what we're seeing from these models is just that they're becoming dramatically more capable, | 我们从这些模型中看到的只是,它们正在急剧增强能力, |
| 355 | and I think it's important that we're thoughtful about how we deploy them ultimately. | 我认为重要的是,我们要思考 我们如何最终部署他们。 |
该视频共有字幕 355 条。解锁更多字幕为会员功能，请移动到 价格