自动驾驶的商业化落地正在全球范围内加速推进。 截至2025年5月，Waymo在美国旧金山、洛杉矶、凤凰城和奥斯汀运营的自动驾驶出租车达到1500辆，每周完成超过25万次付费出行服务；百度Apollo已 在全球部署超1000辆无人驾驶汽车，累计提供超1100万次出行服务，安全行驶里程超过1.7亿公里。 大规模落地仿佛意味着技术已经成熟，其实不然，关于自动驾驶，还有很多尚未达成共识的流派分歧。 比如，传感器方案，纯视觉和多传感器融合方案该如何抉择？系统架构上，采用模块化设计，还是拥抱新兴的端到端架构？更进一步，关于如何理解世 界，VLA和VLM孰优孰劣？ 这些悬而未决的争议，正引领着自动驾驶驶向尚未完全确定的未来。而理解这些不尽相同的技术路线，就是理解自动驾驶从哪里来、到哪里去，如何实现 技术的自我进化。 眼睛之争 纯视觉vs多传感器融合 一切始于"看见"。汽车如何感知世界，是自动驾驶的基石。在这个问题上存在着对峙已久的两大阵营，且双方至今未休。 故事最早可以追溯到2004年美国莫哈韦沙漠的一场挑战赛。 那时，美国国防高级研究计划局设立了200万美元的奖金，吸引数十支顶尖高校和科研机构参与，试图解答"如何让车辆感 ...

文丨詹詹 编辑丨杜海 来源丨正经社（ID：zhengjingshe） （本文约为600字） 原告律师也指控：特斯拉明知系统仅适用于封闭高速公路，却放任其在城市道路激活；马斯克公开宣称"Autopilot比人类更安全"的营销话术，更模糊了技术 责任边界。 最新消息是，特斯拉不服判决，已申请重新审理。他们认为事故责任完全在司机的鲁莽行为，并警告此类判决可能会阻碍安全技术的研发。这不仅是目前特 斯拉遇到的赔偿金额最大的官司，同时也是特斯拉在有关辅助驾驶案件中的首次败诉。【《正经社》出品】 CEO·首席研究员|曹甲清·责编|唐卫平·编辑|杜海·百进·编务|安安·校对|然然 声明：文中观点仅供参考，勿作投资建议。投资有风险，入市需谨慎 喜欢文章的朋友请关注正经社，我们将持续进行价值发现与风险警示 转载正经社任一原创文章，均须获得授权并完整保留文首和文尾的版权信息，否则视为侵权 近日，美国佛罗里达州法院对特斯拉Autopilot系统致人死亡案作出历史性判决，裁定特斯拉承担33%的责任，赔偿总额高达2.425亿美元，约合人民币17.49 亿元。 这是一起2019年发生在佛罗里达州的致命车祸，一辆开启Autopilot的特斯 ...

Group 1 - Baird analysts reaffirmed a "neutral" rating on Tesla (TSLA.US) and lowered delivery and regulatory credit expectations for the second half of 2025, indicating signs of weakness in its automotive business [1] - The analysts believe that market expectations for Tesla's profitability remain overly optimistic and urge the company to clarify timelines for its autonomous taxi and Optimus projects [1][2] - Tesla is currently facing multiple legal and regulatory challenges, including a collective lawsuit approved by a California judge in August 2025, which accuses the company of misleading consumers about the capabilities of its Full Self-Driving (FSD) system for eight years [1] Group 2 - The lawsuit reflects a significant shift in judicial responsibility regarding autonomous driving, as courts are beginning to hold automakers accountable for creating a "false sense of security" in their marketing, while still emphasizing the driver's duty of care [1] - A lawsuit in Florida further exposes systemic flaws in Tesla's systems, alleging that the company delayed critical safety updates and misrepresented its Autopilot system as "nearly autonomous" despite only having partial driving assistance capabilities [1] - Analysts emphasize that Tesla needs to provide transparent disclosures regarding its profitability pathways and the commercialization timeline of its autonomous driving technology to stabilize market expectations, particularly for the autonomous taxi and Optimus robot projects [2]

Group 1: Tesla Legal Issues - Tesla is involved in multiple lawsuits and regulatory disputes, including compensation for a fatal accident related to its autonomous driving technology and a jurisdictional dispute with the SEC [1] - Tesla has filed a motion in the Southern District of Florida to overturn or reconsider a ruling in a product liability and wrongful death case, which could result in a compensation of approximately $242.5 million if upheld [1] - A jury previously found Tesla partially liable for a 2019 fatal accident in Florida involving a Model S vehicle operating in enhanced Autopilot mode [1] Group 2: xAI and Competition with OpenAI - Elon Musk's AI startup xAI has filed a lawsuit against a former engineer for allegedly stealing trade secrets related to the Grok chatbot and moving to competitor OpenAI [2] - The lawsuit highlights the ongoing talent and technology competition between Musk and OpenAI, with Musk also suing OpenAI for deviating from its original mission [2] - xAI has also filed a lawsuit against OpenAI and Apple, accusing them of monopolizing the AI chatbot market on Apple devices [2] Group 3: Tesla Stock Performance - Analysts believe the legal issues have limited impact on Tesla's stock price, referencing a past incident in 2018 that did not alter the company's long-term trajectory [3] - On a recent trading day, Tesla's stock price fell by 3.5% to close at $333.87, failing to break the critical technical level of $340 [3] - Year-to-date, Tesla's stock is down approximately 14%, but it has seen a cumulative increase of 68% over the past 12 months [3]

事故发生后，遇难者家属和幸存者将特斯拉公司与车主共同告上法庭，但后来与车主达成和解，特斯拉 成为惟一被告。直至今年8月1日陪审团裁定，特斯拉应对这起事故承担33％的责任，并判令该公司向受 害者家属及伤者支付约2.43亿美元的赔偿金，包括4260万美元的补偿性赔偿金以及2亿美元的惩罚性赔 偿金。 此前，特斯拉在该案件中坚称，事故的惟一过错方是车主本人，因其分心驾驶使得视线离开了道路，最 终导致惨案的发生；且该公司在用户协议中也清楚写明了驾驶者必须全程保持警惕。最终，陪审团裁 定，驾驶者本人被判定承担67％的过错责任，但由于未被列为被告，相关部分无需赔偿；特斯拉方面承 担33％的责任，并被处以顶格罚款。 为什么特斯拉在这次自动驾驶事故中收到天价罚单？主要有以下几个原因：首先，陪审团认为，在事故 发生时，特斯拉的Autopilot系统未能及时干预，甚至未对前方静止车辆发出警告或采取紧急制动，未预 见道路上存在的风险。而且，Autopilot系统允许驾驶员长时间脱手方向盘，但缺乏有效的生物识别监 控，无法充分保障驾驶安全，存在技术上的不合理性。 当地时间8月1日，美国佛罗里达州迈阿密联邦法院的陪审团裁定，美国电动汽车 ...

Core Viewpoint - General Motors (GM) is accelerating its efforts in autonomous driving by recalling approximately 1,000 former Cruise engineers to refocus on passenger vehicle autonomous driving research, aiming for L3 to L5 levels of automation [1][4]. Group 1: Strategic Shift - GM is shifting from a previous focus on Robotaxi platform expansion to a more pragmatic approach centered on passenger vehicle Advanced Driver Assistance Systems (ADAS), emphasizing safety redundancy, cost constraints, and production rhythm [1][4]. - The new Chief Product Officer, Sterling Anderson, is driving this strategic pivot, leveraging his experience from Aurora, a company known for its fully autonomous truck services [3][6]. Group 2: Technical Integration - The recall of former Cruise talent is intended to integrate the advanced perception, planning, simulation, and safety systems developed in complex urban environments into new passenger vehicle production platforms, thereby shortening engineering ramp-up time and preventing the loss of critical capabilities [4][6]. - GM's previous autonomous driving strategy faced challenges, leading to a pause in the expansion of Cruise's Robotaxi services due to regulatory pressures, financial losses, and operational vulnerabilities [8][9]. Group 3: Future Outlook - The decision to restart ADAS research reflects GM's confidence in the technology and a strategic prioritization of "deliverable, regulatory-compliant, and sustainable" passenger vehicle automation over fully autonomous driving in the short term [9]. - By focusing on L2+/L3 systems across more vehicle models and scenarios, GM aims to build a foundation of safety data and engineering capabilities for future advancements in higher-level autonomous driving [9].

Core Viewpoint - An investor has filed a lawsuit against Tesla and CEO Elon Musk, alleging that the company exaggerated the effectiveness of its autonomous driving technology prior to the launch of its Robotaxi service [1] Group 1: Lawsuit Details - The lawsuit was filed in the federal court in the Western District of Texas, claiming that Tesla misrepresented the safety of its autonomous vehicles, increasing the likelihood of regulatory scrutiny [1] - The plaintiff, Denise Morand, aims to represent investors who purchased Tesla stock since April 19, 2023, seeking compensation for alleged violations of securities laws [4] Group 2: Impact on Stock Price - Following negative media coverage of traffic violations during the Robotaxi service trial, Tesla's stock price dropped by 6.1% over two trading days, closing at $327.55 on June 25 [1] - Tesla's stock price slightly decreased by 0.068% to $309.05 on a recent Monday [2] Group 3: Previous Legal Issues - Tesla was recently ordered to pay $243 million in a separate case related to a fatal accident involving its Autopilot system, marking its first significant legal defeat concerning its driving assistance technology [5] - The company has faced ongoing pressure on its stock price due to various factors, including controversies surrounding Musk's role in the Trump administration and a slowdown in new vehicle sales [5] Group 4: Executive Compensation - Tesla approved a temporary stock reward plan worth approximately $30 billion to incentivize Musk to remain with the company, emphasizing the belief that his leadership will create shareholder value [5]

Core Insights - Tesla's Optimus humanoid robot project is undergoing a leadership change as Milan Kovac, the project engineering lead, has announced his departure, with Ashok Elluswamy taking over [1][4][5] - The transition occurs as the Optimus project is entering a critical phase for mass production, raising industry interest in Tesla's robotics strategy [1][5] Group 1: Leadership Change - Milan Kovac has decided to leave Tesla, citing personal reasons and a desire to spend more time with family, while assuring that this decision will not impact the project's progress [4] - Ashok Elluswamy, the new lead, has extensive experience in autonomous driving algorithms and previously served as the director of Autopilot software [4] Group 2: Project Timeline and Goals - Tesla plans to deploy 1,000 units of the Optimus robot in its factories by the end of 2025, focusing on tasks such as battery handling and material transport [4] - The company aims for large-scale production and external sales by 2026, targeting an annual production of 1 million units [4] - Elon Musk has indicated that the capabilities of the Optimus robot, including speed and precision, have significantly improved, with expectations for unprecedented growth rates in the future [4][5] Group 3: Strategic Importance - The advancement of the Optimus project is crucial not only for the robotics market but also for creating an ecosystem that integrates Tesla's energy and autonomous driving businesses [5] - The success of the project under Elluswamy's leadership will be a key indicator of Tesla's AI strategy effectiveness, particularly in achieving the 2026 production goals [5]

Group 1 - The core issue highlighted is the gap in safety standards and regulatory frameworks for the intelligent driving industry compared to the aviation sector, particularly following the Xiaomi SU7 accident [2][3] - The Xiaomi SU7's removal of LiDAR to reduce costs resulted in inadequate detection capabilities at high speeds, raising questions about the maturity of safety standards in the intelligent automotive sector [2][3] - The article emphasizes the need for a robust regulatory framework for AI technologies, drawing parallels with the stringent safety measures in the aviation industry, such as the extensive certification processes for aircraft [2][3][4] Group 2 - The intelligent driving industry has not learned from the aviation sector's experiences, leading to regulatory lag and insufficient self-regulation among companies [3][4] - The concept of "standard arbitrage" is discussed, where companies exploit outdated regulatory standards to minimize technical responsibilities, prioritizing short-term market competition over long-term user safety [3][4] - The article calls for a discussion on whether to adopt progressive regulation similar to traditional vehicles or to enhance standards by learning from the aviation industry's preventive regulatory systems [4] Group 3 - The ethical implications of AI technology are explored, highlighting the lack of accountability in AI developments compared to the swift regulatory actions taken in cloning technology [5][6][7] - The article points out the absence of a unified global framework for AI ethics and governance, which poses significant challenges in assigning responsibility for AI-related incidents [7][8] - It suggests the establishment of a comprehensive governance framework for AI, akin to a "Manhattan Project," to address ethical challenges and ensure safety in high-risk applications [9][10][11] Group 4 - The proposed governance framework includes strict certification systems, risk assessment mechanisms, and the introduction of an "ethical black box" to trace decision-making processes in AI systems [9][10] - The article advocates for a combination of market incentives and accountability measures to encourage companies to prioritize ethical standards in AI development [10][11] - It emphasizes the need for international cooperation in AI governance to prevent regulatory arbitrage and ensure a cohesive approach to ethical standards across borders [10][11][12]