世界不是由像素组成的：帕西尼的「PIE」、触觉数据工厂与物理AI的最后一厘米

基于音频「机器人也需要"手感"：帕西尼谈触觉、数据与物理AI｜硅谷101年度线下大会」完整转写整理（音频1160秒，约19.3分钟）。本场为英文演讲，开头一段为节目中文旁白。讲者自述为 Dr. Michael Xu（旁白称"Michael徐"）；经公开资料核实，帕西尼感知科技（PaXini，深圳，2021年4月成立）创始人兼CEO为许晋诚博士（Xu Jincheng，早稻田大学菅野实验室工学博士，研究方向即触觉传感与人机交互），且其公开访谈确认公司名刻意取自皮肤机械感受器"帕西尼小体"（Pacinian corpuscle）——与演讲开场自述完全吻合，据此判定为同一人（“Michael"为其英文名属推断，官方页面未见直接确认）。经核实的事实共五组：讲者身份与公司命名来源、五大数采工厂（天津/宿迁/武汉/自贡/赣州，宿迁即演讲中的"East China"工厂）、DexH13灵巧手（即演讲中的H13）、OmniVTLA全模态大模型（转写作"Omni VTA”）、比亚迪与京东为股东；其余数字与主张均为讲者口述口径，未做外部核验。同系列前作：天文学的AI革命篇、半导体AI自进化篇。

先说结论

这是硅谷101 Alignment 2026年度大会系列的第三场，讲者是触觉传感器公司帕西尼（PaXini）创始人许晋诚。他的核心主张一句话就能说完：AI已经能思考，但不理解物理世界——「智能若没有物理理解，是不完整的」（intelligence without the physical understanding is incomplete），而补上这块缺口的关键模态是触觉。讲者的论据来自一个日常场景：机器人可以做出和人类一模一样的擦地动作，但它并不「理解」自己在做什么——人类擦地时靠手感控制喷嘴压在地上的力道，机器人没有这层感知，就是从「知道」（knowing）到「做到」（doing）、再从「做到」到「可靠地做到」（doing reliably）之间两道跨不过去的坎。他的招牌框架是一个缩写PIE——感知（Perception）、智能（Intelligence）、执行（Execution），对应公司的三层业务：自研触觉传感芯片铺满机器人全身、百万小时级含触觉的人采数据工厂加OmniVTLA全模态大模型、经两万小时测试的DexH13灵巧手；在此之上还有第二个PIE——平台（Platform）、集成（Integration）、生态（Ecosystem）。讲者口述的公司盘子很大：累计融资超6亿美元、最新估值超30亿美元、触觉传感器份额全球第一、股东含三星/LG/比亚迪（口述口径；公开报道可确认的股东是比亚迪与京东，融资与估值的公开口径以人民币计价，与口述美元口径存在差异，见文末标注）。对读者的价值在于三条链条：其一，具身智能的瓶颈正在从「大脑」转向「接触瞬间的物理量」——力觉、滑移、纹理；其二，「以人为中心、非遥操作」的数据采集方式决定了数据资产能否跨机器人本体复用，这是触觉数据工厂的真正壁垒；其三，从卖传感器到开源灵巧手、建生态，是「单品硬件」向「物理AI基建」的平台化跃迁。

一、先讲名字：帕西尼小体、鸡蛋与「世界不是由像素组成的」

演讲从公司名讲起，这也是全场最扎实的一段科普。讲者说，人类皮肤里有四类神经，其中负责纹理与快速振动的一类叫帕西尼小体（Pacinian corpuscle）——公司就想做出「像帕西尼小体一样的传感器皮肤」，于是得名（这一命名动机经其公开访谈核实）。他让全场想象一个实验：一只手端满杯的水，另一只手握一枚鸡蛋——你怎么知道鸡蛋没有滑走、没有被捏破？靠的不是眼睛，是触觉信息源源不断传回大脑。节目中文旁白把这一段点题为：机器人大脑越来越聪明，但要走进现实世界，缺的正是「像人类一样感知物体的触感、理解接触时的受力变化」。

顺着这个实验，讲者给出了整场演讲的两个判断。第一个是对「理解」的拆解：机器人擦地和人类擦地，动作可以完全相同，但人类会用触觉去「感受地面」、去控制清洁机喷嘴的压力——机器人做同样的动作时，是否真的理解自己在做什么，「这是一个疑问」。第二个是对AI现状的批判，也是演讲的标题级金句：「世界不是由像素组成的」（the world is not made of pixels）——现实不只有我们看见的部分，还有我们触摸、感受并改变的部分；下一代AI不能只理解指令、只对提示词做反应，而要与世界交互、让事情真实发生，从「观察」（observation）走向「交互」（interaction）。中间还有个轻松插曲：讲者放视频时和观众掰扯了一会儿「现在该说AI还是SI（super intelligence，超级智能）」——术语在这几年膨胀得太快，他自己也笑着承认说不清。

从这些信息可以提炼出本场第一条机制链：触觉模态缺失 → 机器人只能复现动作而无法闭环受力，任务成功率不可靠 → 具身智能的落地瓶颈从「大脑模型」转移到「接触瞬间的物理量感知」→ 对读者的启发是：评估具身智能公司时，别只看基础模型的参数和演示视频，要看它有没有力觉/触觉闭环，以及真实任务的可靠性数据。

二、PIE第一角：一颗触觉芯片，铺满机器人的「全身皮肤」

PIE的第一角是感知，讲者的说法是「一切从我们自己的传感器芯片架构长出来」。他展示了同事Dr. John的手：机器人手指捏着一根尖锐的针接触人体——因为指尖有触觉感知，机器人能在接触瞬间即时反应，这是人机物理共处最直接的安全论据。以这套芯片架构（讲者称之为fused chip/fused technology）为底，产品线从指尖、手掌延展到机器人皮肤（robot skin）、脚底（robot sole），再加关节内部的力矩传感器和六轴力/力矩传感器（转写原文作"NDFactor"，疑为口误或专名，未能确认，保留原文）——目标是让机器人拥有「全身的物理感知能力」。

商业基本盘他也讲得很直白（均为口述口径）：触觉传感器份额「全球第一」；应用不只机器人，还包括汽车等领域；股东有三星、LG、比亚迪这类大公司，且「不只是股东」——帕西尼同时帮他们造自己的机器人和自动化设备（比亚迪为股东经公开报道核实，京东亦为股东但演讲中列在客户侧；三星、LG的股东身份未在公开报道中找到对应，保留口述）。一个有意思的信号是：讲者说最近大量客户来提「闻所未闻」的传感器应用需求——全身触觉打开的应用空间，连供应商自己都事先没想到。

这一段的机制链是：统一传感器芯片架构 → 一个技术族覆盖指尖/掌/皮肤/脚底/关节/六轴力矩的全品类，量产摊薄成本 → 机器人从「指尖装传感器」升级为「全身神经系统」，并外溢到车载、消费电子等非机器人市场 → 启发：触觉传感器的竞争终局不是单品精度，而是谁能成为「全身接触量」的标准供应商。公开报道可作旁证：帕西尼把多维触觉传感器单价从约十万元人民币级做到最低199元，降幅约500倍——没有统一架构与量产，这个曲线画不出来（此为公开报道口径，非演讲内容）。

三、PIE第二角：数据工厂——「这不是遥操作」，以及百万小时触觉数据

PIE的第二角是智能层，讲者先补了一个认知前提：智能需要的不只是知识（knowledge，比如文本），更是经验（experience）——机器人「知道」杯子是个圆柱体，但未必知道怎么抓它。经验从哪来？从人与物体的真实交互里来。帕西尼的答案是数据工厂：真人佩戴触觉采集设备（演讲中展示了一只采集手套）去抓握、分拣真实物体，把每一次交互的触觉信息连同动作一起记录下来。讲者特别强调了一句关键的话：「这不是遥操作（this is actually not the teleoperation）」——采集的是人的数据，再用这些数据训练「很小的模型」，就能产生非常精确的动作。这条路线的公开注脚很扎实：帕西尼官方口径称传统遥操作采集成本高、机器人本体一迭代旧数据就贬值、且多数采集本体根本没有触觉传感器，数据「看着多、密度不够」；其「以人为中心」的人因采集设备不依赖特定机器人本体，效率是传统方式的3到6倍，采集员佩戴的正是自研PMEC数据采集手套——与演讲内容互为印证（经核实）。

规模上，讲者给出三个口述数字：超过100万小时的含触觉高精度人类数据，构成所谓的"omnichannel database"（全渠道数据库；官方口径称「全模态数据库」，年产近2亿条数据，小时与条两种计量口径并存，见文末标注）；数据工厂共建五座——华东宿迁（Suqian，转写作"Suzhen"）、华北天津，另有三座即将落地（经核实：公开口径为天津先导厂加宿迁/武汉/自贡/赣州新建四厂，五厂联动，与演讲一致）；此外还有面向家庭的采集设备，「有些人就想在自己家里教机器人」——数据入口从工厂铺到日常场景，讲者称之为「真正的物理智能数据飞轮」。算法层对应自研的OmniVTLA全模态大模型（转写作"Omni VTA models"，经官网核实官方名为OmniVTLA，即视觉-触觉-语言-动作）：支持预训练、后训练与终身学习，机器人自己运行产生的数据也回流训练；触觉信息按慢适应（slow adapting）与快适应（fast adapting）两类分别嵌入——恰好呼应开场的皮肤感受器分类。讲者还预告了一个即将在「月底」发布的"physics world"（物理世界）：做触觉信息的编码器，让机器人「感觉世界」，目标是few-shot乃至zero-shot的泛化。他给出的基准测试也极简：机器人能不能处理从未见过的物体——「现实是机器人的终极考题（reality is the ultimate task）」。

这一段藏着本场最重要的机制链：「以人为中心、非遥操作」的采集方式 → 数据不绑定特定机器人本体、且自带触觉真值，可跨本体复用 → 规模化形成百万小时级数据资产，叠加机器人自产数据的终身学习飞轮 → 二阶影响：具身数据的护城河属性不是「量」而是可复用、可迭代、全模态对齐 → 启发：判断一家具身公司的数据故事，先问三个问题——数据能不能换本体继续用？有没有力/触觉通道？机器人部署后的数据能不能回流？（许晋诚在CES 2026期间的公开采访中也表达了同样的三属性论，与本场互为印证。）

四、PIE第三角与第二个PIE：灵巧手、开源与生态

PIE的第三角是执行。讲者的逻辑很干脆：感知再完美、大脑再聪明，没有合格的执行层——没有灵巧手这样的「执行器」——机器人闭环就落不了地。帕西尼的做法是自研多款机器人手设计，并且开源手部设计给需要的机器人公司（口述）。可靠性是拿具体产品背书的：旗舰H13灵巧手（经核实即DexH13：四指仿生、16自由度、单手集成约1140个ITPU多维触觉单元与AI手眼相机，2026年CES全球发布，并出现在NVIDIA CES 2026 Keynote中）经历了「超过两万小时的使用测试，之后依然是一只非常有前景的手」（口述）。讲者还举例说有机器人公司用帕西尼的手设计弹钢琴（转写作"Chestnut robots / Mister Chestnut"，对应的公司与人物未能确认，保留原文，不排除为其他公司名或人名的误识）。客户名单同样是口述：LG、三星、京东、比亚迪，以及手机领域的公司（“the owners"处语义不明，保留原文）。

最后讲者把PIE升了一层：第一个PIE（感知-智能-执行）解决机器人与物理世界交互的能力；第二个PIE——平台（Platform）、集成（Integration）、生态（Ecosystem）——解决开放生态，「让每个想做自己机器人的人，都能用自己的技术真正开发成功」。收尾的双关也在此：「bring your piece, build our pie」——带上你那一块（piece），一起把这张饼（pie）做大。愿景落点是：人人拥有自己的个人机器人，把人类从劳动中解放出来，「让机器人去做更高层级的事」。

这一层的机制链：开源灵巧手设计 + 平台化生态 → 降低下游机器人公司的执行层门槛，把帕西尼从传感器供应商变成「物理AI的触觉基建」 → 二阶影响：生态位的护城河从硬件毛率转向数据与标准 → 启发：观察指标是外部机器人公司采用其开源手与数据接口的程度、DexH系列出货量，以及"physics world"发布后的零样本泛化演示。

现实是终极考题：这场演讲确定与不确定的边界

把三场连起来看，硅谷101这场年度大会无意中拼出了物理AI的完整切面：天文学那场讲AI如何消化远超人类注意力的数据洪水（天文篇），半导体那场讲AI如何开始制造运行自己的芯片（半导体篇），而帕西尼这一场补上了最后一段：AI如何触摸并改变物理世界。讲者没有回避行业最核心的怀疑——机器人复现动作不等于理解动作——他的答案是把「手感」工程化：一颗芯片铺满全身、百万小时含触觉的人采数据、一只经过两万小时考验的手。这个答案成立的前提是：触觉数据确实能像文本之于大语言模型一样形成规模效应，且「以人为中心」的数据能跨本体泛化——前者尚无公开的第三方基准，后者是全行业遥操作路线之争的胜负手，演讲都未给出可验证的数据。对读者，可操作的观察清单有三条：DexH13及后续手型的实际出货与采用；五厂数据产能（官方口径年产近2亿条）是否有下游付费验证；“physics world"发布后的零样本抓取演示是真实任务还是演示场景。确定的是方向——「世界不是由像素组成的」这句话，大概率会成为物理AI这一轮叙事里被反复引用的注脚；不确定的只是，谁先把触觉做成基建。

ASR转写校正表

本场转写为英文演讲+节目中文旁白，识别质量整体良好。按技能规范，结合上下文与公开资料校正如下；附录中仅机械应用「确认」级校正，「疑似/未确认」项保留原文：

转写原文（修正版）

以下为整份转写稿全文，已按上表「确认」级校正机械替换（公司名各拼写→PaXini、Omni VTA→OmniVTLA、Suzhen→宿迁），保留 [起-止] 时间戳；「疑似/未确认」项（NDFactor、Chestnut、omnichannel、the owners 等）保持原文。原始文件：SoundAnalyze/transcripts/机器人也需要“手感”：帕西尼谈触觉、数据与物理AI｜硅谷101年度线下大会.txt（音频1160秒，10段，全文无截断）。

[0s - 120s] 过去几年，机器人大脑越来越聪明，但要真正走进现实世界，还差一样很重要的东西，就是像人类一样能感知物体的触感，理解接触时的受力变化。那么，如何让机器人真正拥有感知和理解物理世界的能力？接下来，就让我们回到硅谷幺零幺的Alignment二零二六大会现场，看看Michael徐如何理解Physical AI的发展方向，以及机器人还需要突破哪些技术瓶颈。I’m uh Dr. Michael Xu, and today’s topic is actually the pie for physical AI. And later I will explain what is the pie, the PIE why is it. Maybe for everyone it’s like just a food, but for us it’s actually a very important thing for robots. So for uh our company name, actually a lot of people will uh have a very interesting part is that they are they were asking what is our name? Why our company name is PaXini? Uh, it’s uh because the in the human skin, uh we have four different kind of nerves. The one is responsible for the texture and the vibration. A fast vibration is called Pacinian corpuscle. So we want to make the sensor skin a very similar and likely like a Pacinian corpuscle. So that’s why we make we make the name is called PaXini. And the sensor is actually for the the finger is very important. So uh just before we start, we can actually ask everyone to imagine. For example, your left hand and the right hand, you can hold a a cup full with waters, and the other hand hold it with a egg. We want to ask how human actually know why we react with this kind of object? It’s actually based on the tactile sensors. Why we are why the the egg we holding on is not sliding or slipping, or we are we are not breaking.

[120s - 240s] Just because we have a lot of tactile sensing information, so that’s why our brain can process it. So the tactile sensor is very important for for our human body for an interact with the physical world. Very short introduction for our company. We are a global leader in embodied perception driven intelligence robotic company. And currently, we have the total funding raise over 600 million dollars. And our newest valuation of the companies are over 3 billion dollars. And for our products, actually we have the number one market share in the world. We are not only using on the robots. Actually, we have in the automobile in other field. So let me introduce a little bit about our shareholders. Our shareholders is actually for example Samsung, LG, BYD, this kind of big company corporation. They are our shareholders. So we they are not only our shareholders. We also help them to build their own robots or their own automation machines. So for the future robots, actually we see this kind of big corporation. They are very focusing on the physical AI, and they realize that the tactile sensing information is a super important for the for the robots itself. When we start to realize the the interaction with the physical world, we want to ask how does the intelligence truly existing in the physical world? And this turns out we want to have more question coming up, is if the AI can think, but can it understand the physical world? For example, this graph over here, a human is doing the cleaning the floors, and the robot is cleaning the floors too. But when the robots is doing exactly the same thing, the the action, but does it realize or

[240s - 360s] Really understand how it is doing. When we are doing the cleaning on the floor, we actually need to use our hand, our tactile sensors to feel the floor and pressing the nozzle of the of the clean machine and on the floor to actually clean it to to do the vacuum. But the robot really really understand the action. Actually, it’s a it’s a doubt. So intelligence without the physical understanding is incomplete. And if we want to know the the robots from knowing to doing and from doing to doing reliably, it’s also a big cross. It’s a it has a very big gap. So we want to ask, for example, we want to know the next generation of the AI. They will not just understand the the instruction. They will not just react for the prompt. They actually need to interact with the world and make the things happens or change. And not only that, actually the world is not made of pixels. We are not just using the cameras to see everything. So reality is not just what we see. It’s also what we touch, we feel, and we change. And to understand the world, actually we need to for our companies, we actually push AI to understand the reality and then to push the AI or push the now now we call AI, but I think the the video just say you you are now saying SI right, super intelligence. Are you saying super intelligence? I don’t know. Are you AI now? Especially in maybe outside of US. Okay, back to the slides. So AI must to know to go beyond the the observation, not only the scene, but to react or to interact with it. And from the observation to interaction, we have actually a lot of these kind of experiments because we want to have.

[360s - 480s] The robots from the passive sensing to really active, really actively understanding the physical world. So only that we can let the robot to learning from the realities. So as you can see over here, the the the tiny videos over here. When the robot is actually doing the the manipulations with the tactile sensor, it it actually feels how it grab the nozzle or how it grab the thin object or the the light bulbs. And not only that, actually intelligence, it actually needs a lot of more than knowledge. Knowledge is like for example the text. It needs actually more than text. It actually needs more like experience. What is the experience? It it actually is a human data. And a robot may know what is a cup is. You can watch it. You can see a cup is a surrounding cylinders. But it does it how know how to grab it or interact with it. Maybe maybe it doesn’t know. So the experiments actually teaches us or teaches robots how to act in it, do the action or manipulation with it. So more about that. If we want to have more experience, and we actually need to build in the experience layer for for robot or for AI, and physical AI, because we need to sense the world. So if you see this this there’s a procedure here. It’s actually one of our procedure in in our data factory. The human observe and interact and then pick up the the bottles, and we have the datas to interact with this kind of object. So that’s why we build this kind of foundation for the AI data or robot data, and then for machines to learn from the physical world. So this is actually very important layers. And for the every interaction is actually

[480s - 600s] It’s an opportunity to to learn from the physical world. So not only the data factory in one of the place, we also want to let the robot learn from everywhere. So actually, we have this kind of devices for some people. They just want to teach robots in their houses or in their families. So we want to have the every opportunity to that can teach the the robot, and it’s it’s actually the real physical intelligence flywheel of the data itself. And turns out the physical interaction into intelligence also one of the key points. So you can see over here, it’s actually one of our glove, and we teach actually to the robots. And this is actually not the not a teleoperation. We use the data actually to train a very little models, and then it actually has very precise the action for the for the result. So turns out the perception is very very important for the the tactile signal itself, and also interaction with every every object. And learning turns out to be a higher successful way for each kind of the mission for the robot itself. And if we have this kind of experiments, and then to let the robot to learn, and how can we do the benchmark, or how can we evaluate this kind of a this kind of a mission is fulfilled with our needs, human needs. So we need to actually build a benchmark. And for a very simple way is that can the robot handle with the what is never encountered before? For example, the object it never see, the object it never handled before. Can it actually? If the robot can handle with it, it actually can exist in and the intelligence in the real world.

[600s - 720s] So eventually, we want to ask: reality is the ultimate task for the robots, and the only major thing is to let the robot and the physical AI grounded in measurable and real-world interaction. That’s why back to our pie. So pie is how intelligence meets the realities. With the pie, I want to describe more about our pie. A lot of people asking, why is pie? Is it apple pie, pineapple pie? What kind of pie is it? It’s actually built from perception, intelligence, and executions. For every robot, it actually needs these three things, and one pie is actually can create infinity possibilities. With our PaXini’s pie, we can create more than usual for these kind of robots. And first, I want to describe describe the perception itself. The perception is actually built from our own sensor chips. As you can see over here, this is Dr. John’s hand, and Dr. John actually teaching and using the tactile sensing ability on the hand itself. And this is actually a very sharp needle on the fingertip when it pinching, when the robot hand is pinching. And if the robot actually having this kind of sensing ability, it actually can react and interact with human human being. And with this kind of sensor chip architecture, we can build everything. So now, what you see on the on the fingertip of the the robot sensors and tactile sensors, and also the sensor on the automobile, maybe for other different different kind of applications, we actually we can build everything from this architecture of the chip, the fused chip, fused technology. Now we actually have different kind of products on the on the robots. For example, in the joint, we have the torque sensors. We have the NDFactor for the six-axis force force torque sensors. We actually have the tactile

[720s - 840s] Sensor or for the fingertip, for the palm, and for the robot skin, even for the robot sole, and this kind of technology actually brings out we can let every robot have the fully the whole body physical sensing abilities. Based on this kind of technology, it actually has higher potential for even unknown needs, and because now recently we have a lot of customers, they also ask us they want to have different kind of sensors on their robots, and this is never heard before the application, so it’s actually astounding for for this kind of needs from every robot companies. And for the intelligence itself, we build actually the data factory. The data factory is actually a very major project for PaXini Technology, because now we can have more than one million hours we the high precision human datas with tactile sensor with tactile sensing information, and it’s called omnichannel database. If every anyone is interesting, you can just check it online on our website, and with this kind of datas we can create for example this kind of application for for example in the warehouse, in different kind of family use, and also in the factories. And with the technology itself, we build actually five different kind of factories. One is in the East China Suqian（宿迁）, and then North China Tianjin, and also we have three mores coming up. With the datas we can have the we can let the robot to learn from it, and so how to let let the robot learn from it? We need to have the algorithm, so that’s why we create this kind of a multi system algorithm for the robot itself, and a lot of people.

[840s - 960s] Here, you maybe is an engineer of robotics. You must know about the robot training. We can have the data to doing the pre-training, post-training, and even with a lifelong learning. So this kind of data is actually very important for a robot itself. And for that, we actually create the flywheel algorithm models for the robot model itself to actually to prove the data itself is very useful. So it’s called OmniVTLA models. With these models, you can see we can use the data from the data factory, and we can also use the data from the robot itself. When the robot generate the data itself, we can still using the data for lifelong learning. And the model itself can be the very important part for the for the robot to do maybe a zero shot. And now, so that’s why we actually create the so-called physics world. Physics world will actually release very soon, maybe next the end of this month. And this kind of technology actually to digest or make the encoder of the tactile sensing information, and then to the the robot to feel the world. And then we can have a very few shot or zero shot technology for for the robots. And as you can see over there, we are not give up the language, but we are still using the language. We will still using the vision itself, and also we will using the robot the interact the information for itself. And not only that, the major part is the tactile sensing. So we can we will have the slow adapting tactile information, and then the fast fast adapting tactile sensing information for for the robot to embedding the tactile sensing information for the model and algorithm itself. And more than that, actually we use these models to use on

[960s - 1080s] Different kind of applications. I will describe a little bit more about the applications, and about the PIE, the last part of the E, the execution. The execution is actually one of the major thing for for the robot because if you have the perfect perception, sensing ability, you have the brilliant brain, but you don’t have a very qualified execution layer. For example, you don’t have the robot whole body robots, you don’t have the execution any factor. For example, the robot hand, then you are actually far away from the from the finishing or closed loop of the robot to doing the the mission in the reality. So that’s why we create this kind of execution layer to have different kind of robot hand design, and the robot hand design is actually a very major part for many many different kind of robot company. For example, the Chestnut robots, they are just using the Mister Chestnut using the his hand to play the piano. It’s actually one of the design of the the robot hand itself. So for PaXini, we actually we can open source the hand design for the robot company they need, and for the design itself, we actually we have a lot of improvement in reliability. For example, one of the the major and famous hand is actually our H13. It actually have a lot of testing. We have over twenty thousand hours of the usage, and after that, still it’s a very promising hand. And not only that, we have a lot of different kind of applications with with our customers. For example, the LG, Samsung, and also JD.com, BYD, and then also with some of the the companies in the in the phone area, the owners. Finally, we want to have not only the PIE.

[1080s - 1160s] Pie for the perception, intelligence, and execution. We also want to build the pie for the platform, for the integration, for the ecosystem itself. So we have two different kind of pie. One is to have the robot to have the capability to react with this kind of a physical world. The other pie is to have the very good open ecosystem for everyone who wants to have their own robots to really successfully develop with their own technology. We want to actually have the final part of our of our presentation. We want to have the one pie for the one future, and then have infinite possibilities. In the last part, is we want everyone to bring your piece and build build your own pie or build our pie. And for the technology itself, we want to have everyone to have your own own robots and personal robot, and then to that the to free the the labors and then to have more fun into the physical world and then that the robots to do the higher things or missions. So we hope you guys enjoy my presentations. Thank you very much.