A MONSTER Weekend for Tesla and SpaceX ⚡️

好的，以下是内容的中文翻译： 发言人首先测试了FSD重新激活功能，据信该功能已在14.3.9版本推出，目前在14.3.10版本上。轻踩刹车后FSD解除，系统重新激活，屏幕上显示了一条新消息，发言人觉得“太酷了，太酷了，太酷了”，并赞扬了特斯拉AI团队。 **星舰14号飞行任务汇报：** * **入轨：** 星舰首次成功进入轨道，部署了26颗星链V3卫星。 * **星链V3：** 这些卫星通过激光链路、地面站和用户终端实现了联系。它们将在为客户提供服务之前进行检查并提升轨道，服务质量可能要到明年第二季度，当大约一千颗V3卫星投入运行后，才会发生有意义的变化。发言人指出，关于星链速度更快的说法为时尚早。 * **助推器问题：** 助推器上的一台猛禽发动机在初始爬升过程中关闭。热分离成功。超重型助推器在返回点火中重新点燃了33台计划中的猛禽发动机中的31台，随后在主推力着陆点火中点燃了13台中的11台，在溅落在海湾之前减少到5台，然后是3台发动机。着陆点火关闭后，飞行终止系统（FTS）被触发，作为安全演示。 * **星舰问题：** 在上升过程中，一台猛禽真空发动机提前关闭。星舰通过让剩余的五台发动机燃烧更长时间，仍然达到了亚轨道轨迹。随后，它重新点燃了一台猛禽海平面发动机以进入轨道。 * **轨道持续时间：** 名义上预计飞行六圈，但由于上升过程中发动机问题，为预防起见，仅完成了两圈。 * **脱轨与着陆：** 星舰重新点燃了一台猛禽海平面发动机进行首次脱轨点火，随后再入大气层，并重新点燃了所有三台猛禽海平面发动机进行着陆点火和翻转机动，最终精确溅落在太平洋。 * **容量：** 罗布（Rob）指出，SpaceX在一个小时内增加的星链容量比其前1.5年的发射总量还要多。 * **股价与变现：** 发言人猜测，增加的容量是否会转化为公司的现金流，以及是否会出现新的服务层级或更高的费用以获得更快的服务。 * **星舰41号着陆：** 马库斯·豪斯（Marcus House）表示，星舰41号不会被回收，在接触水面时发生了爆炸，这与星舰40号在13号飞行任务中的着陆形成了对比。 * **发动机性能（助推器）：** 在13号飞行任务中，助推器着陆时13台发动机中只有8台重新点燃，而在14号飞行任务中，13台中有11台成功点燃。 * **15号飞行任务预测：** 发言人不认为15号飞行任务会尝试捕获，预计将是另一次轨道插入尝试，目标是飞行超过两圈并实现100%的发动机功能。如果尝试捕获，星舰（而非助推器）的可能性更大。 * **星链未来：** 埃隆·马斯克（Elon Musk）重申，从长远来看，星链越来越有可能承载地球上大部分的IP流量，有效地成为互联网。 * **快速重复使用：** 发言人强调，捕获两艘飞行器对于快速重复使用和低成本至关重要。 * **发射频率：** 埃隆·马斯克表示，星舰距离实现每小时一次的飞行还有两到三年。目前的发射间隔显示，最近两次飞行之间分别为63天和66天。 * **载荷成本：** 14号飞行任务部署了26颗V3卫星，估计重52吨。每次发射1.1亿美元，这大约是每吨210万美元，已经接近猎鹰9号的每吨成本范围（15吨载荷，3000万美元成本，约为每吨200万美元）。这是一艘完全一次性的星舰，载荷不足其运载能力的一半。 * **历史巧合：** 星舰精确地在猎鹰1号火箭发射18年后的同一天进入轨道。埃里克·伯杰（Eric Berger）建议“轨道日”应该成为公司永久性的节日。 * **15号飞行任务展望：** 发动机挑战将影响SpaceX对15号飞行任务的计划，很可能涉及另一次带有溅落的测试运行。 * **捕获时间线：** 发言人认为，要到2027年初才能实现捕获两艘飞行器，这将是发射速度的一个重要转折点。 * **飞行持续时间：** 飞行持续了大约三个小时，而非预期的九个小时。 **特斯拉自动驾驶出租车/赛博出租车：** * **注册车辆：** 特斯拉在德克萨斯州共注册了546辆用于自动驾驶出租车网络的车辆，其中126辆（占23%）是赛博出租车。 * **车队增长：** 9月份新增了232辆车，车队规模几乎翻倍。 * **活跃车辆：** 据传，每天有70-100辆车活跃，发言人因此认为特斯拉正在为未来的某个事件做准备，可能是10月10日的X收购事件。 * **赛博出租车注册：** 仅29天内就注册了126辆赛博出租车，超出了发言人的预期。 * **最新更新：** 特斯拉发送通知称赛博出租车车队已翻倍，提供更多乘车服务。 * **新功能：** 一项新的自动驾驶出租车功能是：如果乘车前曾打开过行李箱，那么在乘车结束时会自动打开行李箱。经确认，该功能仅在车门打开后才会触发。 * **客户体验（轶事）：** 乔什（Josh）报告了一次负面体验：自动驾驶出租车将他们放在距离目的地几个街区的地方，客服支持无效且无法看到目的地，退款请求被拒绝。 * **埃隆的商业计划：** 埃隆·马斯克表示，赛博出租车将很快在佛罗里达州、内华达州和其他州投入商业运营，并于明年年中在加利福尼亚州运营。 **Grok模型：** * **Grok 4.7：** 埃隆·马斯克承认4.7版本可能未能达到预期，但他强调其在当前和未来的实际工程方面的卓越表现。 * **对比：** 这与Anthropic的模型形成对比，后者对软件工程师来说非常出色。 * **特斯拉/SpaceX使用：** 特斯拉和SpaceX的工程师已经在使用Grok模型。 * **企业应用：** 发言人认为，由于Grok专注于实际工程和来自特斯拉/SpaceX的训练数据，它在企业应用方面具有显著优势。Grok预计在软件使用案例方面也会有所改进。 **Optimus：** * **手部设计挑战（谣言）：** 报告，主要来自“The Information”，表明Optimus的手部设计极其复杂，以至于目前的组装设备无法可靠地制造。此外，据报道，特斯拉的员工拒绝穿戴动作捕捉服进行训练，担心工作被取代。 * **怀疑：** 发言人对动作捕捉的说法表示怀疑，指出特斯拉已在很大程度上转向仅通过摄像头收集数据。 * **手部机械结构：** Optimus的手部采用肌腱驱动结构，前臂每条手臂有25个执行器，每个执行器使用三根细控制线，定位精度达到0.08毫米。 * **制造问题：** 汽车级设备难以达到手部电缆通道和关节组件所需的亚毫米级公差，导致大量的返工。触觉传感器也不可靠。特斯拉的解决方案是可更换的手套式模块化传感器层。 * **组件：** 这只手需要100多个螺丝和微型部件，而成熟的工业供应商尚未能稳定批量供货。 * **产量提升：** 据报道，特斯拉已将Optimus的产量增加了十倍，从第二季度每周几十台增至8月份的每周几百台，目标是年底前每周超过1000台，长期目标是每周20000台（每年从弗里蒙特生产100万台）。 * **监管担忧：** 联邦通信委员会（FCC）于7月禁止新的外国制造人形机器人在美国市场销售，尽管宁波等供应商的单个部件目前不在禁令范围内。对部件的更严格限制可能即将到来。 * **商业策略：** 早期的Optimus V3单元将租赁给精选的仓库客户，而非直接销售，以提供灵活性。全面商业销售预计不会早于2027年末。 * **发言人意见：** 发言人警告不要对规模抱有过高期望，但考虑到挑战，他认为“每周数百台”的报告是看涨的。他们重申，特斯拉选择设计一个尖端的手部，因此遇到挑战是预料之中的。 **Roadster发布会：** * Roadster发布活动因户外活动恶劣天气预测而改期至10月15日。 **FSD在欧盟：** * 在10月6日的TCMB会议上，不会对特斯拉的受监督FSD进行投票；该投票至少推迟到12月。 * 瑞典交通管理局建议投反对票，除非特斯拉移除速度偏移功能。其他国家如法国也有自己的担忧。发言人指出，这与他们最初的预测一致，即各国将拖延。 **总结：** * 该视频是第一部分，第二部分将于明天发布。 * 特斯拉股票收盘价为357.45美元（下跌3.94%）。 * SpaceX股票收盘价为145.47美元（下跌2.16%）。 * 纳斯达克100指数（NDX）下跌1.08%。 * ZocDoc被提及为该视频的赞助商。

The speaker begins by testing the FSD reactivation feature, believed to have launched in 14.3.9 and now on 14.3.10. Tapping the brake disengaged FSD, and the system reactivated, showing a new message on the screen, which the speaker found "so, so, so cool" and praised the Tesla AI team. **Starship Flight 14 Debrief:** * **Orbital Insertion:** Starship successfully reached orbit for the first time, deploying 26 Starlink V3 satellites. * **Starlink V3:** These satellites achieved contact via laser links, ground stations, and user terminals. They will undergo checkouts and raise orbits before serving customers, likely not meaningfully changing service until Q2 next year when around a thousand V3s are operational. The speaker notes claims of faster Starlink are premature. * **Booster Issues:** A single Raptor engine on the booster shut down during the initial climb. Hot staging was successful. Super Heavy relit 31 of 33 planned Raptor engines for the boost backburn, then 11 of 13 for the high thrust landing burn, reducing to five, then three engines before splashing down in the Gulf. The flight termination system (FTS) was triggered upon landing burn shutdown as a safety demonstration. * **Ship Issues:** A single Raptor vacuum engine shut down early during ascent. Starship still achieved its suborbital trajectory by burning the remaining five engines longer. It then reignited a single Raptor sea level engine for orbital insertion. * **Orbit Duration:** Nominally, six orbits were expected, but only two were completed due to engine issues during ascent, as a precaution. * **De-orbit & Landing:** Starship relit a single Raptor sea level engine for its first de-orbit burn, re-entered, and relit all three Raptor sea level engines for a landing burn and flip maneuver before splashing down on target in the Pacific Ocean. * **Capacity:** Rob noted that SpaceX added more Starlink capacity in an hour than in the first 1.5 years of launches. * **Stock Price & Monetization:** The speaker speculates on whether increased capacity will translate to cash for the company and if new tiers or higher costs will emerge for faster service. * **Ship 41 Landing:** Marcus House stated Ship 41 would not be recovered, showing an explosion upon water contact, contrasting with Ship 40's landing on Flight 13. * **Engine Performance (Booster):** On Flight 13, only 8 of 13 engines relit for the booster landing, whereas on Flight 14, 11 of 13 lit successfully. * **Flight 15 Predictions:** The speaker does not expect a catch on Flight 15, anticipating another orbital insertion attempt with a goal for more than two orbits and 100% engine functionality. If a catch is attempted, it's more likely for the ship than the booster. * **Starlink Future:** Elon Musk reiterated that Starlink is increasingly probable to carry the majority of Earth's IP traffic long term, effectively becoming the internet. * **Rapid Reusability:** The speaker emphasizes that catching both ships is crucial for rapid reusability and low cost. * **Launch Cadence:** Elon Musk stated Starship is two to three years away from hourly flights. The current cadence shows 63 and 66 days between recent flights. * **Payload Cost:** Flight 14 deployed 26 V3 satellites, estimated at 52 tons. At $110 million per launch, this is roughly $2.1 million per ton, which is already in Falcon 9's cost-per-ton territory (around $2 million per ton for 15 tons at $30 million). This was a fully expendable Starship carrying less than half its capacity. * **Historical Coincidence:** Starship reached orbit exactly 18 years to the day after the Falcon 1 rocket. Eric Berger suggested "orbit day" should be a perpetual company holiday. * **Flight 15 Outlook:** Engine challenges will influence SpaceX's plans for Flight 15, likely involving another test run with splashdown. * **Catch Timeline:** The speaker believes catching both ships won't happen until early 2027, which will be a major inflection point for launch pace. * **Flight Duration:** The flight lasted about three hours, instead of the anticipated nine. **Tesla Robotaxi/Cybertaxi:** * **Registered Vehicles:** Tesla has 546 total vehicles registered in Texas for the Robotaxi network, with 126 (23%) being Cybercabs. * **Fleet Growth:** 232 vehicles were added in September, nearly doubling the fleet. * **Active Vehicles:** Anecdotally, 70-100 vehicles are active daily, leading the speaker to believe Tesla is preparing for a future event, possibly the X takeover event on October 10th. * **Cybercab Registration:** 126 Cybercabs were registered in just 29 days, exceeding the speaker's expectations. * **Recent Update:** Tesla sent a notification saying the Cybercab fleet has doubled, with more rides available. * **New Feature:** A new Robotaxi feature auto-opens the trunk at the end of a ride if it was opened before the ride, confirmed to only open after a door is opened. * **Customer Experience (Anecdote):** Josh reported a negative experience where a Robotaxi dropped them off a few blocks from the destination, support was ineffective and couldn't see the destination, and a refund request was denied. * **Elon's Commercial Plans:** Elon Musk stated Cybercab will soon operate commercially in Florida, Nevada, and other states, and in California by mid-next year. **Grok Models:** * **Grok 4.7:** Elon Musk admitted 4.7 might have been a letdown relative to expectations but emphasized its excellence for real-world engineering, both current and future. * **Contrast:** This is contrasted with Anthropic's models, which are excellent for software engineers. * **Tesla/SpaceX Use:** Tesla and SpaceX engineers are already using Grok models. * **Enterprise Adoption:** The speaker believes Grok has a serious edge in enterprise adoption due to its focus on real-world engineering and training data from Tesla/SpaceX. Grok is expected to improve for software use cases as well. **Optimus:** * **Hand Design Challenges (Rumors):** Reports, primarily from "The Information," suggest Optimus's hand is so intricate that current assembly equipment cannot reliably build it. Also, Tesla's workforce is reportedly refusing to wear motion capture suits for training, fearing job displacement. * **Skepticism:** The speaker is skeptical about the motion capture claim, noting Tesla has largely moved to camera-only data gathering. * **Hand Mechanics:** The Optimus hand has a tendon-driven architecture with 25 actuators per arm in the forearm, using three thin control cables per actuator for a positional accuracy of .08 millimeters. * **Manufacturing Issues:** Automotive-grade equipment struggles with the sub-millimeter tolerances needed for the hand's cable channels and joint assemblies, leading to significant rework. Touch sensors are also unreliable. Tesla's response is a replaceable glove-type modular sensor layer. * **Components:** The hand requires over 100 screws and micro components, which mature industrial suppliers haven't consistently delivered at volume. * **Output Increase:** Tesla reportedly increased Optimus output tenfold, from dozens per week in Q2 to several hundred per week by August, targeting over 1,000 per week by year-end, with a long-term goal of 20,000 per week (1 million units per year from Fremont). * **Regulatory Concerns:** The FCC banned new foreign-made humanoid robots from the U.S. market in July, though individual components from suppliers like Ningbo are currently outside the ban's scope. Tighter restrictions on components may be coming. * **Commercial Strategy:** Early Optimus V3 units will be leased to select warehouse customers rather than sold outright, offering flexibility. Full commercial sales aren't expected before late 2027. * **Speaker's Opinion:** The speaker cautions against high expectations for scale but finds the "hundreds per week" report bullish, considering the challenges. They reiterate that Tesla chose to design a cutting-edge hand, making challenges expected. **Roadster Unveil:** * The Roadster Unveil event was rescheduled to October 15th due to severe weather predictions for an outdoor event. **FSD in EU:** * There will be no vote on Tesla's supervised FSD at the October 6th TCMB meeting; it's delayed until at least December. * The Swedish Transportation Authority recommended voting no unless Tesla removes the speed offset feature. Other countries like France have their own concerns. The speaker notes this aligns with their initial prediction that countries would drag their feet. **Closing:** * This video is Part 1, with Part 2 coming tomorrow. * Tesla stock closed at $357.45 (down 3.94%). * SpaceX stock closed at $145.47 (down 2.16%). * The NDX was down 1.08%. * ZocDoc was mentioned as the video's sponsor.