Data Acquisition & Strategy - Tesla's FSD free trial is a data collection initiative targeting 1.5 million HW4 vehicle owners [1] - The trial aims to gather 1.5 billion kilometers (approximately 932 million miles) of real-world driving data in 30 days [1] - The data volume surpasses Google's Waymo's data collection efforts [1] - Tesla is using massive amounts of data to influence regulations and train its FSD model [1] Technological Implications - HW3 vehicle owners are excluded, signaling a shift in the Robotaxi landscape [1] - Tesla's AI flywheel is poised to accelerate [1]

Financial Data and Key Metrics Changes - Revenues for Q3 2025 reached $13 million, an increase of 81% year-over-year and 48% sequentially, marking one of the strongest quarters in the company's history [6][14] - Non-IFRS gross margins for Q3 2025 were 56%, down from 74% in Q3 2024, reflecting increased investments and a higher share of large-scale projects with AI customers [15][16] - Net profit for Q3 2025 was $0.1 million, a significant decrease from $4.2 million in Q3 2024, primarily due to a high financial income in the previous year [17][18] - Adjusted EBITDA for Q3 2025 was $1.2 million, slightly down from $1.4 million in Q3 2024 [17] Business Line Data and Key Metrics Changes - The company saw a 26% increase in paying customers and a 17% increase in average revenue per customer during the quarter [6] - The product suite is evolving from a single product proxy business into a diversified multi-product data infrastructure platform, with significant growth in AI-centric products [11] Market Data and Key Metrics Changes - Strong growth was noted in the AI vertical, offsetting declines in other segments, with one large-scale AI customer contributing approximately $3.5 million in revenue [15] - The company is experiencing significant traction with major global e-commerce platforms in Asia, which are placing repeat and expanding orders [7] Company Strategy and Development Direction - The company prioritizes long-term relationships and market share capture over short-term margins and profitability, with a focus on AI-driven growth [3] - Strategic investments are being made to expand capacity and enhance infrastructure, which are expected to drive long-term growth despite short-term margin pressures [8][10] - The company aims to shift towards higher value products, which will improve unit economics and margins over time [9] Management's Comments on Operating Environment and Future Outlook - Management expressed confidence in the growing demand for AI services and the company's positioning within the data industry, anticipating a revenue of approximately $12 million for Q4 2025, representing a 62% year-over-year growth [12][19] - The volatility in demand from leading labs is expected to stabilize as models transition from research to production [10] Other Important Information - The company's cash position as of September 30, 2025, was approximately $24.6 million, supporting ongoing strategic investments [18] Q&A Session Summary Question: Can you talk about the large project for Dataset delivery? - Management clarified that the demand is for a combination of a scraper and a dataset, with high customer satisfaction and retention [22][25] Question: Do you see usage being higher or lower once R&D customers have developed their models? - Management indicated that while initial data needs are high during the education stage, usage may become more predictable in the production stage [26][27] Question: What is the pipeline to sell this new Dataset delivery solution to other customers? - Management confirmed that there are already other customers leveraging this product, with additional prospects in the pipeline [30] Question: How should we think about the gross margin recovering as revenue scales? - Management projected that gross margins could improve significantly if third-party vendor reliance decreases, with a potential gross margin of around 70% if all solutions were in-house [31][33] Question: When will unit economics improve? - Management noted that as demand stabilizes and infrastructure efficiency increases, unit prices may rise, benefiting overall margins [36][37]

Core Insights - The article discusses the advancements in humanoid robotics, particularly focusing on the TWIST2 system, which integrates portability, scalability, and full-body control capabilities, overcoming the limitations of traditional motion capture systems [1][3][4]. Group 1: TWIST2 System Overview - TWIST2 is a new humanoid robot remote operation and data collection system developed by teams from Amazon FAR, Stanford University, and UC Berkeley, which allows robots to perform complex tasks autonomously [3][4]. - The system eliminates the need for expensive motion capture equipment, enabling quick deployment and operation [4][14]. Group 2: Key Innovations of TWIST2 - The system features a low-cost detachable neck module priced at $250, which provides egocentric vision capabilities when combined with a $400 stereo camera [9][14]. - TWIST2 utilizes a combination of PICO 4U VR devices and two leg motion trackers, with a total hardware cost of approximately $1,000, allowing for easy setup in just one minute [14][17]. - The system captures full-body human movements at a frequency of 100Hz, significantly improving data collection efficiency compared to previous systems [14][24]. Group 3: Data Collection Efficiency - TWIST2 allows a single operator to control the entire data collection process, achieving high efficiency with the ability to complete 100 successful demonstrations in just 20 minutes [22][25]. - The system's total data flow frequency exceeds 50Hz, with a latency of less than 0.1 seconds, enhancing the operator's ability to perform precise tasks [24]. Group 4: Autonomous Control Capabilities - TWIST2 introduces a hierarchical visual motion strategy framework that enables robots to execute tasks autonomously based on visual input, without relying on simplified speed commands [26][30]. - The framework has been successfully tested, allowing robots to perform complex tasks such as precise object manipulation and dynamic balance [32][35]. Group 5: Cost-Effectiveness and Accessibility - The total cost to set up a complete TWIST2 system is around $1,650, making high-quality humanoid robot research more accessible [37]. - The research team has made the system, data, and models open-source, positioning TWIST2 as a foundational infrastructure for humanoid robotics research [39].

We're being surveilled!A lot of data is being collected from your everyday purchases when you use credit cards. This could potentially lead to further taxation or even denial of service.@TheDesertLynx https://t.co/t4Xb4LrMBY ...

Core Viewpoint - AppLovin is facing an SEC investigation regarding its data collection practices, leading to a significant drop in its stock price, which fell over 16% this week [1][3][4] Investigation Report - The SEC is investigating AppLovin for potentially illegal data collection practices aimed at targeting customers for advertisements, with allegations of collecting personal data from mobile applications like TikTok [3] - The outcome of the investigation could result in fines and impact future revenue generation if the company is found to be targeting advertisements more precisely than allowed [3] Stock Performance - AppLovin's stock initially tanked due to the investigation news but later recovered some losses before declining again amid a broader market sell-off [4] - Despite the recent drop, AppLovin's shares have increased by 300% over the past year, with a revenue growth of 77% to $1.26 billion last quarter and a 266% increase in revenue over the last five years [4] Valuation Concerns - AppLovin's stock is trading at a high price-to-sales (P/S) ratio of 37, significantly above the S&P 500 average of 3.4, indicating high growth expectations to maintain current stock levels [5] - The competitive landscape of digital advertising, dominated by large tech companies, presents both opportunities and risks for AppLovin, especially in light of the SEC investigation and short selling reports [6]

Core Viewpoint - AppLovin's shares fell by 14% due to a report of a potential investigation by the U.S. Securities and Exchange Commission regarding its data collection practices [4] Company Summary - AppLovin is facing scrutiny from the U.S. Securities and Exchange Commission, which is investigating the company's data collection processes [4] - The decline in share price reflects investor concerns over regulatory actions and potential implications for the company's operations [4]

Core Points - Meta will start using conversations with its AI chatbot to personalize ads and content on Facebook and Instagram, with notifications to users beginning on October 7, 2025, and the changes taking effect on December 16, 2025 [1][2] Group 1: Implementation Details - Interactions with Meta AI, whether through voice or text, will be used similarly to how the platform currently tracks likes, shares, and posts to determine user interests [2] - The changes will initially roll out in most regions, with Europe and the UK receiving updates later due to stricter data and privacy regulations [4] Group 2: User Control and Privacy - Users cannot opt out of data collection when using Meta AI, but they will have control over the level of personalization of ads and content on their feeds [4] - Conversations about sensitive topics, such as religious views, sexual orientation, political views, and health, will not be used for ad targeting [5] Group 3: User Engagement - More than 1 billion people use Meta AI across its family of apps, including Facebook, Instagram, and WhatsApp [3] - The goal of the announcement is to ensure transparency and provide users with ample notice regarding the upcoming changes [3]

Humanoid Robotics Industry Overview - Humanoid robots are seen as the future of labor, potentially automating roughly 50% of global GDP currently attributed to physical labor [3] - Humanoid robots are considered the "gold standard" for data collection, essential for building and maintaining digital twin models [4] - Big tech companies are actively strategizing and investing in humanoid robots to secure a piece of the data pie that will be generated over the next 20 years [5][6] - The bottleneck for humanoid robot adoption is not manufacturing, but deployment, including strategy, integration, workforce preparation, and inventory digitization [15][16][17] - A 10 to 15-year supply-demand imbalance is expected for humanoid robots in the lowest-hanging fruit use cases, with companies willing to pay $100,000 per bot annually over a four to five-year lifespan [18] Investment Strategies & Market Analysis - The speaker believes that Tesla is the only publicly traded company worth considering in the humanoid robotics space, despite acknowledging issues with the Optimus program [20][21][22] - The US is likely to ban Chinese robots due to data security concerns, creating separate markets for China and the rest of the world [23][24] - The speaker emphasizes the importance of distribution in the humanoid robotics market, believing that companies with commercially viable relationships and the ability to deploy robots at scale will control and own the data [39][40][41] Social Investing Methodology - The speaker advocates for "social investing," which involves detecting changes in the world and connecting them to companies that will be affected before the information becomes widely disseminated [47][48] - The speaker uses various data sources, including TikTok comments, web traffic data, and institutional credit card data, to identify trends and assess consumer demand [55][56][59] - The speaker exits a trade when the information that motivated the investment becomes widely known among other investors [52][53] - The speaker is a concentrated, levered investor, typically holding 5 to 15 names at a time and allocating 5% to 30% of their portfolio to high-conviction trades [68][69]

User Acquisition - American firms are rapidly expanding their user base [1] Data Accumulation - These firms are accumulating vast amounts of data [1]

Data Collection Review - The Labor Department's Office of Inspector General is initiating a review of the challenges the BLS faces in collecting data [1] - The review will focus on challenges and mitigating strategies for collecting PPI and CPI data [2] - The review will also focus on collecting, reporting, and revising monthly employment data [2] Context and Response - The review follows the president's move to fire the BLS commissioner last month [2] - The Labor Secretary stated the department is committed to finding solutions to data problems, including modernization for improved transparency and accuracy [2] Timeline - No specific timeline for the review has been laid out [3]