这或许是当下对打感最好的网球机器人。
头图来源：Hirdwire
2024 年巴黎奥运会，郑钦文在罗兰·加洛斯捧起金牌的那个下午，中国的网球场预订量在一周内翻了数倍。两年后，最先把这股热潮变成生意的，是一群造机器的深圳工程师。
过去一年，至少四款中国团队打造的 AI 球类机器人在 Kickstarter 上筹得百万美元级资金。庞伯特以 26255% 的超募率卷走 265 万美元；智能家居大厂 SwitchBot 推出的 Acemate 30 天入账 180 万美元；Tenniix 首战筹得近百万美元后，其母公司英汉思动力完成红杉联合领投的三轮共超亿元融资；最新收官的 AceiiLab 首批量产正在交付，单台售价约 2000 美元，早鸟价 999 美元。
但火归火，这条赛道连「卖的是什么」都还没说清。在「AI 网球机器人」标签背后，可以是打网球的人形机器人，也可以是便携发球机，也有可能是一款手机 App。
为什么网球机器人赛道突然爆火？一台真正的 AI 网球机器人到底该长什么样？带着这两个问题，我们在深圳的一片球场上见到了 AceiiLab，和它背后一思科技的联创晁光。随后我们发现，这场竞赛里硬件只是入场券，真正的关键在于你把这台机器定位成一件器材，还是一个对手。谁能让用户打得更久、离开得更难，谁才有机会提前拿到「AI 网球机器人」的定义权。
一、「球场上来了个会跑位的陪练」
见到实物前，我们对网球机器人的想象是一块会移动的大屏，或一条能挥拍的机械臂。
AceiiLab A1 被拉进球场时，看起来像一只 20 寸行李箱，下半截是带轮子的本体，上半截是可拆卸的球包，可装 120 颗球。过去是一手拎机器、一手拖球筐，现在两件行李合成一件，整机约 20 公斤，四个轮子推行。
「我们发现在设备上，用户的本质需求不是要求机器轻，而是希望自己搬运不累。选择行李箱的形式，不需要一直发力，很大程度上减少累的体感。」AceiiLab 相关负责人说，「行李箱的方式既解决了移动难题，也是一个友好的用户使用过程。」
真正开机后，我们发现 AceiiLab 的玩法和常见的传统发球机并不一样，它是一个可移动的、交互性较强的产品。「这可能是我当下使用过最好的 AI 网球机器人，移动带来的真实对打感是其他同类产品几乎不可比的地方。」一位参与到现场与 AceiiLab 对打的朋友说道。
将 AceiiLab 放在网球线上，按下主机开关后，两侧车轨支架自动展开。十几秒后，摄像头认出球场上的线，机器开始沿线巡逻。它并不满场乱跑，只认中场线、底线和边界线这几条「跑道」，像一名守规矩的球员，绝不越位。
AceiiLab A1 开机准备 ｜来源：Hirdwire
热身完毕，好戏开场。开球前需先在 App 里选择模式，目前主要有三种玩法：组合&移动、自由练习、接发球训练。
以组合 & 移动模式为例，AceiiLab A1 会按照内置的 50 多套球路组合，一边沿线移动一边按剧本喂球，正手一颗、反手一颗，深一颗、浅一颗，球速和旋转轮着变，发球间隔最快能压到半秒一颗。
自由练习则把发球选择权交还给用户，无论是隔网对打、同侧陪练，还是让它在底线自由移动，用户均可按需选择。它还能定点喂球，或站在左右边线喂出角度刁钻的球，相较于传统发球机练来练去就那几个角度，它能把整条底线都变成发球点。
接发球练习模式中，还有一个「球童模式」，你不需要按下开始键，只需竖起球拍，机器就能自动识别就绪状态并完成发球，就像有一位专业球童一直站在对面为你发球。
真正开打后的画面，和我们想象中激烈的「对拉」有些不同。所谓「对拉」的真相是，你把球打过网，球还没落地，机器就移动到落点附近，紧接着从落点处发出一颗新球，模拟「接到了你的球」从网对面给你发一颗新球。它从不真正回击你那颗球，但一来一回之间，回合感确实成立。
预测用户击球落点位置｜来源：AceiiLab
不过，这既考验用户的适应能力，也考验机器能否快速到位、精准发球。如果中间卡顿明显，对拉体验就会大打折扣。我们花了至少几十分钟，才慢慢适应和机器打球的节奏。
120 颗球，按正常 3.0 水平的对打节奏，大概能打 30 分钟左右。由于每一颗球都是新球，整体消耗速度也会比较快。「不管是新手还是熟手，都能获得不同的体验感。」
「反应不过来」是我们早期采访多位参与到和 Aceii 对打的用共有的第一感受。整体对打过程中，我们明显感受到，这并非一款适合 0 经验小白无门槛上手的产品。它更有效地地方，在于巩固正反手和移动步伐，基本堪比早期训练过程中，同侧训练中，模拟教练手抛球。它需要用户了解自己的水平，也清楚如何设置相关发球参数，是一款需要与用户调试配置、相互适应的产品。据介绍，这款产品希望能够满足 2.0 至 4.0 水平用户的需求。
二、教练、球童、损友，它全包了
制造业有一条隐秘的鄙视链：卖产品的不如卖耗材的，卖耗材的不如卖服务的。过去五十年的发球机厂商，始终停留在链条最底端，只做一锤子买卖，卖完就结束了。
「大部分发球机厂商的目标就是把机器卖出去。但卖出去对别人是终点，对我们却是起点。」在 AceiiLab 相关负责人看来，发球机买回家容易吃灰，原因无非两点，价值上限太低，替代不了教练；或者是使用频率太低，新鲜感一过就闲置。AceiiLab 想造的是一个反向飞轮，设备越好玩，用户越常打；越常打，越进步；越进步，越好玩。硬件解决「打得起来」，软件负责留住用户。
飞轮的第一圈，从一张成绩单开始。每轮打完，你都能看到这一场的训练数据：接球成功率、最高连击、平均球速、最高球速。还有一张落点图，把你和机器的每颗回球都钉在球场上，界内界外，一目了然。
更关键的是，它还在「教」你。机器会录下你的击球视频，从拍面、稳定性、时机、动力链、准备动作五个维度分析，告诉你这一轮的表现如何，并与过往数据进行对比。
当你把数据提交 AI 分析后，系统会指出弱点所在，不仅有原因、建议，还会推送相关视频课程，帮你更好地调整状态。
在 AceiiLab 的设想里，「AI 教练的一天」应该是这样的：你昨天反手随挥偏移不够、扭胯动力链不足，今天就练这两点，空挥比对标准动作，练完机器再考你一遍。一套完整的教练流程，被原样搬进 AceiiLab 系统。
不仅如此，练球之外，它还想增添更多趣味性。打球时，你可以开启球童模式，挥手即出球；也可以切换游戏模式闯关，把枯燥训练变成一场主动挑战。嫌一个人打球无聊，还能把精彩瞬间分享进社区，顺手约好友来练球。「流量世界已经被算法算尽，变量只在物理世界。一台在球场上狂奔的网球机器人，天然引人围观。」相关负责人说道。
这些设计，不少来自市场最直接的反馈。北美路测时，一位用户打了半小时，问出一个团队没预料到的问题，「我是不是永远赢不了它？」他们由此意识到，机器可以无限打下去，但人不行。人需要关卡，需要输赢，需要一个能被打败的对手。
于是有了闯关，甚至有了机器的败北规则：球落地时，如果它没跑进 50 公分以内，就算它输。因为一台有机会认输的机器，比一台永远正确的机器，更像球友。
而所有这些数据攒到最后，会变成用户离不开机器的理由。「当机器存着你几年的成长数据，甚至存着你最好朋友的陪练数据，即使竞品便宜 500 块，你也不会换。」相关负责人说道。当参数可以被追平，模具可以被抄走，唯独用户投入的时间，谁也搬不走。
我们也由此看到，未来硬件公司的壁垒，或许不在于产品参数，而在于品牌向用户关系的转化。AceiiLab 正在一个具体的场景里，用 AI 和数据持续交付服务。
三、在最卷的供应链上，长出一个新物种
一台网球机器人，为什么只沿底线跑？为什么不接球？这些看起来是技术问题，但在早期阶段其实都是产品选择问题。要理解这些选择，得先认识做选择的人。
AceiiLab 背后的一思科技，是一支服务机器人老兵攒起来的队伍。创始人刘礼谦在大疆做过结构工程师，在蓝胖子做过仓储机器人，在库卡做过系统工程师，后来转向商用清洁和割草机器人，当过割草机器人公司虎鲸的 CEO，拿过当年的海外众筹冠军。这支三十多人的团队，在服务机器人行业里摸爬滚打了许多年。
正因为做得久，他们看出了无论服务机器人还是网球机器人，其实都活在一套具体的规则里。割草机认草坪的边界，网球机器人认球场的线，都是规则，场地是标准的，球路是抛物线。而网球几乎是真实世界里最守规矩的运动，这是他们早期看好这个方向的原因之一。
一思科技办公室上方的天台球场以及办公室一角 来源：Hirdwire
但两类机器又有一个根本的不同。割草机器人和用户之间是雇佣关系，草割干净就行，用户不关心机器怎么割的。它服务的是结果，价值的天花板，最多就是一份雇工的工资。网球机器人正好相反，它服务的是过程。打球的人在乎每一拍的球速多少，落点在哪，动作标不标准。机器每一秒都在感知你、回应你，这是伙伴关系。他们最终想实现的，也正是让用户随时随地都能与高水平的教练和球友对战。
想清楚了要做什么，他们的下一个问题是怎么做，在普通球友掏得起的价格里，把价值做到最大。
首先要过成本这一关。团队的办法是站在成熟产业的肩膀上，机器的底盘直接采用扫地机器人打磨了十年的移动方案，轮子直接用行李箱的轮子，移动上采用类似车道线保持的解决方案。
省下来的钱，换成了时间和价格空间。但省钱不等于凑合，该下重注的地方，他们选择自己造，比如说这台机器最贵的零部件，机器人的「眼睛」。
市面上现成的双目摄像头模组基线太短，最多能看清球落在哪，看不清球是怎么飞过来的。一旦看不清轨迹，机器就无法提前跑位，所谓「对手感」更无从谈起。为此团队自研了超长基线双目，球飞出去 10 毫秒，它就算出了第一个落点；飞行过程中不断观测不断调整，50 毫秒，整条运动轨迹了然于胸。户外光线的干扰也被压到最低，逆光、眩光、夜场的灯，都不影响它看球、判球。
另一处重注下在算法上，这也是全公司人数最多的团队，主要解决运动算法分析能力。他们已经在做多模态的肢体与动作分析；在模拟对手方面，目前的用户建模仍以规则式为主，也就是把教练的经验知识提炼成步伐、手势等六个维度的参数，写进规则，已经能模拟出诸如「纳达尔的左侧上旋球」这样的打法训练方式。
眼睛和算法是这台机器上他们花钱最大的两个地方。但真正定义这台机器的，是他们选择不做的地方。毕竟，做加法只需要钱，做减法需要判断，而这支团队对 AI 网球的判断，都写在一张「不做」的清单里。
四、一张「不做」清单，看懂消费级 AI 网球机器人的取舍题
在各类产品形态衍生的今天，我们看到 Aceii 选择了很多「不做」：不做满场跑的机器人，只沿底线移动；不做花哨功能，用移动发球串起体验；不一次性做完所有需求，首先努力满足 2.0 到 4.0 水平的进阶球友。每个「不做」背后，都有依据。
其中，我们最好奇的是，为什么不造一台类人形机器人来打网球？晁光说：「人形机器人打网球，验证的是机器人的能力；我们做的是网球机器人，交付的是打球的体验，是真正能够让消费者愿意买的产品。在平整场地上，轮子的效率高于双足，就像人用洗衣机的效率高于手搓。」
而选择不做接球的背后，则藏着一个训练观念。在 Aceii 看来，发球和接球是两件事，发球决定训练的节奏，机器为了追求接球把自己的节奏打乱，整体的训练价值便打了折扣。更何况，接球成功率一旦不够高，场面就会倒过来，不是机器陪人练球，而是人在遛机器。
还有选择底线移动的原因，则是源于市场需求。据行业统计，2.5 以上的目标用户七八成的击球是底线对拉。守住底线，就守住了七八成的体验。
那么，一台处处克制的机器，为什么偏偏在「移动」上不计成本？
在他们看来，移动是 AceiiLab 和传统发球机的分界线。一台放在原地的机器，和人的交互一上午只有一次，开机，设置，然后各干各的。而一台会移动的机器，一秒就交互一次，你打左它追左，你打右它追右，你被它看见了，它在回应你。这也带来了交互纵深。此外，移动带来了对手感，带来了多点喂球、隔网对拉、球童这些玩法，也带来了一种微妙的情感连接。我们在现场也感受到早期使用用户对于「移动」的对打体验，颇受好评「能在移动中发球，沉浸感真人感更足」。
把这些「不做」倒过来看，AceiiLab 的选择路径就显影了。中国硬件公司走了三十年的老路，把一个品类做便宜，再被后来者做得更便宜，做一代产品，打一场仗，毛利一代薄过一代，直到无仗可打。AceiiLab 想走的是另一条，先用一个被彻底重构的发球体验把机器卖出去，靠量产降本，靠移动和交互留住人，再让攒下的数据变成服务。
在这条路上，克制不是保守，而是一种耐心。这条路能不能走通，现在下结论还早。首批量产机今年 7 月才开始交付，订阅的续费率、课程的完课率、社区的活跃度，都要等真实用户给出数字。
时间站在哪一边，球场上的用户会给出答案。
本文由 AI图鉴自动抓取并翻译，版权归原作者所有；如需查看原文请点下方链接。
It's probably the best tennis robot in the day.
Source: Hirdwire.
At the 2024 Olympic Games in Paris, during the afternoon when Zheng Chin Wen held the gold medal at Roland Garlos, China's tennis court bookings doubled in one week. Two years later, the first to turn the tide into business was a group of Shenzhen engineers who built machines.
Over the past year, at least four AI-type robots built by the Chinese team raised millions of dollars on KickStarter. Pompert took $2.65 million at an over-mobilization rate of 26,255 per cent; the Acemate, launched by SwitchBot, an intelligent home company, recorded $1.8 million over 30 days; Tenniix, whose parent company, after almost a million dollars in the first war, managed to complete three rounds of co-financed Redwood; and AceiiLab, the newest receiver, was in the process of delivering its first production at a single price of approximately US$ 2000 and early bird prices of US$ 999.
But the fire is on fire, and this track is not even clear about what it sells. Behind the "AI tennis robot" label, it could be either a tennis human robot, a portable transmitter, or possibly a cell phone, App.
Why is the tennis robot track suddenly on fire? What does a real AI tennis robot look like? With these two questions, we saw Aceiilab on a field in Shenzhen, and the technology behind it. Then we found that hardware in this race is just a ticket, and the real point is whether you have this machine as an instrument or as an opponent. Who can play longer, and who can leave harder, and who can get a "AI tennis robot" in advance?
One, "There's a run-off in the field."
Before seeing the object, our imagination for tennis robots is a big screen that moves, or a swinging arm.
When Aceii Lab A1 was pulled into the field, it looked like a 20-inch suitcase, the lower part of the body with a wheel, the upper part was a detachable bag, with 120 balls. It used to be a machine with one hand, a tow basket with one hand, and now it's two pieces of luggage combined, about 20 kilograms, four wheels.
"We find that the essence of the user's needs on the equipment is not to require the machine to be light, but rather to move it yourself. The choice of the form of the suitcase does not require a constant effort to significantly reduce the feeling of fatigue. AceiiLab, the person in charge, said, "The way in which the suitcases are used solves the problems of movement and is a friendly user usage."
And when it really turned on, we found that Aceii Lab's play was different from the usual traditional players, and it was a mobile, interactive product. "This is probably the best AI tennis robot I've ever used, and the real sense of hit that moves is almost comparable to that of other products. One of the people who was involved in the scene with Aceii Lab said to his friend who was beaten.
Place the AceiiLab on the tennis line and automatically expand the tracks on both sides after pressing the mainframe switch. Ten seconds later, the camera recognized the line on the field and the machine began patrolling along the line. It is also dissatisfied with the fact that it only recognizes the "runway" of the midfield, the bottom line and the boundary line, like a disciplined player, and never exceeds its position.
Aceiilab A1 ready to go. Source: Hirdwire
It's time to warm up. There are currently three main types of games: combination & movement, free practice, and catch training.
In combination & mobile mode, for example, AceiiLab A1 will follow the built-in 50-set combination, moving along one side by feeding the ball by the script, one by one, one by one, one by one, one by one, one by one, one by one by one, the speed of the ball and the wheel will change and the speed of the ball will be pressed to half a second.
Free exercise gives the choice to the user, who can choose as needed, whether by cutting off the net, working side by side or by moving around the bottom line. It can also feed the ball, or it can feed the ball at the right and the right edges, and it can turn the entire bottom line into a point of play, compared to the traditional player practice.
There's also a "boy" mode in the game practice mode, and you don't have to press the start button, you just have to put up the bat so the machine can automatically recognize the state of readiness and finish the ball, like a professional boy who's been standing across the street.
It's a little different from what we think about the "yes" that really started. The truth is that you hit the ball across the net and before it landed, the machine moved near the drop point, then sent a new ball from the drop point and simulated a new ball from across the net. It never really hit your ball back, but between and between, it does.
Source of projected user strike point location: AceiiLab
However, this tests both the user ' s adaptability and the rapid availability and precision of the machine. If Carton was clear in the middle, it would be a big discount on the pull experience. It took us at least a few minutes to adjust to the pace of playing with machines.
120 balls, playing the rhythm at a normal 3.0 level, can play about 30 minutes. Because each ball is new, the overall rate of consumption will be faster. "Whether beginner or an expert, you get different experiences."
"Reactional failure" is the first feeling that we shared with Aceii in our early interviews. As a whole, we clearly felt that it was not a product that was suitable for zero experience without threshold. It is more effective in consolidating the positive backhand and moving pace, which is largely comparable to early training, where a simulation of a coach throws a ball. It requires users to know their level, and it is clear how to set the relevant firing parameters, a product that needs to be configured and adapted to users. It was described that the product was expected to meet the needs of users at 2.0 to 4.0 levels.
Two, coach, ballboy, loss friend. It's all covered.
There is a hidden chain of stigmatization in the manufacturing sector: products are sold less than waste materials, and materials are sold less than services. Over the past 50 years, players have stayed at the bottom of the chain, selling only a hammer, and it's over.
"The goal of most players is to sell the machine. But selling it to others is the end, and for us it is the beginning. In the view of AceiiLab's responsible officials, it is easy to eat ash at home for two reasons, because the upper limit of value is too low to replace the coach; or because the frequency of use is too low and the sense of freshness is idle. What AceiiLab wants to build is a reverse wheel, and the more fun the equipment is, the more frequent the users are fighting, the more progressive the more progressive the more fun it is. Hardware solves "getting it together" and software keeps users.
First lap of the wheel, starting with a report card. After each round, you can see the training data for this game: catch success rate, hit the highest, average and highest. There's also a drop-off map that nails every ball you and the machine to the field, inside and out.
More importantly, it's teaching you. The machine will record your hit video, which will tell you how this round works and compares it with past data, from a five-dimensional analysis of film, stability, timing, power chain, preparation for action.
When you submit the data to AI analysis, the system identifies the weaknesses, not only for a reason, but also for a related video course to help you adjust your state better.
In the idea of AceiiLab, "Ai Coach's Day" should be like this: you didn't have enough swings yesterday, you didn't have enough twirling power chains, and today you're going to practice these two things, you're going to wave to standard, and you're going to do the machine again. A complete train-the-trainer process was moved into the Aceiilab system.
More than that, it wants more fun than practice. When you play, you can open up the kid's mode, wave your hand, or you can switch the game mode off and turn dry training into a proactive challenge. It's boring to play alone, and it's a good moment to share it with the community and meet friends to practice. "The flow world has been calculated, and the variables are only in the physical world. It's a tennis robot that runs on the field, and it's natural to look around. The blogger says:
Many of these designs come from the most direct feedback from the market. At the North American Road Survey, a user played for half an hour and asked a question that a team didn't expect, "Do I never win?" They realized that the machine could go on indefinitely, but not human. People need a level, they need a win or a win, they need a rival who can be defeated.
So there's the break-in, and there's even the machine's losing rule: when the ball falls, if it doesn't get within 50 centimeters, it loses. Because a machine with a chance to lose is more like a ballmate than a machine that is always right.
And all this data will end up being the reason why users can't drive a machine. "When the machine saves your growth data for several years, even your best friend's companion data, even if it's 500 bucks cheap, you won't change it. The blogger says: When parameters can be pursued, molds can be copied, and no one can move them except for the time that the user has invested.
We also see that the barriers to future hardware companies may lie not in product parameters, but in the transformation of brands into user relationships. AceiiLab is delivering services on a continuous basis using AI and data in a specific scenario.
III. A new species emerges in the most rolling supply chain
A tennis robot. Why run down the bottom line? Why not catch the ball? These appear to be technical issues, but at an early stage they are essentially product selection issues. To understand these options, it is important to know who makes the choices.
The technology behind Aceii Lab is a team of veterans who serve robots. The founder, Liu Rei-Kin, was a structural engineer in Daeji, a storage robot in Blue Fats, a systems engineer in Cooka, then a commercial cleaning and weed-cutting robot, who was the CEO of the lawn mower robot killer whale, and who was the champion of the year abroad. This team of over 30 people has been crawling around in the service robot industry for years.
It's because of the long hours they've been doing, they've seen that whether they serve robots or tennis robots, they actually live in a set of specific rules. The lawn mower's lawn boundary and the tennis robot's line to the field are rules, the site is standard and the path is parabolic. And tennis is almost the most disciplined sport in the real world, and that's one of the reasons why they looked in that direction earlier.
The skyball court above the tech office and the corner of the office.
But there is a fundamental difference between the two types of machine. The business relationship between the lawn mower robot and the user is that the lawn mower is clean and the user does not care how the machine is cut. It serves the result, the value of the ceiling and, at most, the wages of an employee. Tennis robots, on the contrary, serve process. The players care about the speed of each shot, where it falls, and it's not standard. Every second the machine senses you, responds to you, it's a partnership. What they eventually want is for users to be able to fight high-level coaches and ballmates at all times and everywhere.
Their next question is how to maximize value at a price that ordinary ballmates can afford.
First of all, there is the cost. The team's approach is to stand on the shoulders of a mature industry, where the machine's chassis is based directly on a 10-year-old field-cleaning robot, where the wheel is based directly on a wheel in a suitcase, and where a solution similar to that maintained by a driveway is applied.
Saved money for time and price space. But saving money is not the right place, and they choose to build their own parts, like the most expensive parts of the machine, the robot's eyes.
The off-the-shelf double-eye camera model baseline is too short to see where the ball falls at most and how it flew. Once the trajectory is not clear, the machine cannot run ahead of time, much less the so-called "enemy sense". For this, the team studied its own ultra-long baseline twos, and the ball flew out of 10 milliseconds, and it came out of its first drop point; constant observations were made during the flight, at 50 milliseconds, and the entire track was in the chest. Outdoor light interference is also kept to a minimum, and the reverse light, the dazzling light, and the lights of the night field do not affect their viewing or sentencing.
Another is focused on algorithms, which are also the largest corporate team, mainly addressing motor algorithm analysis. They are already doing multi-modular body and action analyses; in the case of simulations of opponents, current user modelling is still dominated by rules, i.e. six dimensions of the instructor's experience, such as pace, gestures, etc., and rules, which can already be modelled on a method of training, such as “Nadal's left swing ball”.
Eyes and algorithms are the two places where they spend most money on this machine. But what really defines this machine is where they choose not to do it. After all, it takes money to add, it takes judgment to make deductions, and the team's judgment about AI tennis is in a list of "no."
iv. A list of no-dos to understand the consumer-class AI tennis robot's trade-offs Topic
Today, we see that Aceii has chosen a lot of "no-dos": robots that don't run full-time, moving down the bottom line; no fancy functions, with moving-out balls, to experience; and no one-time needs to do all the work, first trying to meet the goalmate at 2.0 to 4.0 levels. There's a basis behind every "no" thing.
Among them, we'd better wonder why we don't build a human robot to play tennis? "The human robot plays tennis, and it's the robot's ability; what we do is we do tennis robots, and we deliver the game experience, and it's a product that really allows consumers to buy." Wheels are more efficient than two feet on the ground, as is the use of washing machines by humans.
And there is a training concept behind choosing not to catch. In Aceii, it appears that two things are going to be true: the pace of training is determined by the pitch, and the whole training value is reduced by the machine's disruption of its rhythm in pursuit of the catch. Moreover, if the catch success rate is not high enough, the scene will turn upside down, not with the machine, but with the machine.
There is also a choice of bottom-line movement that is driven by market demand. According to industry statistics, 70 per cent of the target users above have a bottom line. You hold the bottom line, you hold the 70%.
So, why is a machine that's always restrained?
In their view, the movement is the line between Aceii Lab and the traditional players. A machine that is placed in place, which interacts with humans only once a morning, opens, sets, then works separately. And a moving machine that interacts once in a second, you hit it left, you hit it right, you're seen, it's responding to you. This also brings with it a deep interaction. In addition, mobility brings a sense of rivalry, multi-baby feeding, net-to-penetration, child games, and a delicate emotional connection. We also felt at the scene that the early use of the user's experience of "move" was highly appreciated, as it was "showing the ball in the movement and getting more immersed."
If you don't do this, the AceiiLab chooses the path. The Chinese hardware company has been on its way for 30 years, making a product cheaper, making a generation of products, fighting a war, and a Maori generation thinning over a generation until there is no war. AceiiLab wants to go another way, sells the machine with a thoroughly re-engineered playoff experience, sells the machine with a volume drop, moves and interacts to keep people, then turns the saved data into a service.
On this road, restraint is not conservative, but patience. It's too early to conclude that this road is clear. The delivery of the first batches began in July of this year, and the rates of renewal of subscriptions, completion of courses, community activity are waiting for the real users to give the numbers.
On whose side the time is on, the user on the field will give the answer.
Auto-fetched and translated by AI Atlas. All rights belong to the original author.