CrowdStrike’s $290 million acquisition of Onum aims to strengthen real-time data threat detection amid escalating cyber threats, while Lyft reports three profitable quarters, expanding globally with autonomous ride innovations. Additionally, AI’s influence continues to grow, with Google’s Jeff Dean predicting it could surpass human breakthroughs within 20 years.
▶️ Management and Leadership
CrowdStrike Buys Onum to Boost Real-Time Data Threat Detection
CrowdStrike acquired Onum for about $290 million to enhance real-time data threat detection, supporting its AI-native security vision amid rising cyber threats and complex M&A environment.
- CrowdStrike announced its acquisition of data observability startup Onum for approximately $290 million.
- The deal was completed within three months, focusing on real-time pipeline detection and threat analysis.
- CrowdStrike’s M&A strategy emphasizes acquiring companies at the right stage to minimize integration risk and maximize value.
Lyft Achieves Profitability and Expands Globally with Autonomous Ride Innovation
Lyft, under CEO David Risher since 2023, reports profitability, global expansion, and innovation in autonomous rides, aiming to improve customer experience and reduce surge pricing impacts.
- Lyft CEO David Risher reports three consecutive profitable quarters after years of losses, with increased active riders and drivers, and international expansion via acquisition of FreeNow.
- Risher emphasizes a company purpose focused on customer obsession, operational excellence, and global growth, moving away from a domestic-only, B2C model.
- Lyft has worked to eliminate surge pricing, citing rider dissatisfaction, and is partnering with autonomous vehicle companies like Mobileye to develop driverless rides; autonomous vehicles are expected to reduce driver pay and insurance costs but may alter pricing dynamics.
Excel Remains Top Skill in Silicon Valley Job Listings Despite AI Rise
Despite AI’s rise, Excel remains the top skill in Silicon Valley job listings, with 531,000 mentions, crucial for data management amid ongoing talent wars and high compensation offers.
- Excel, released in 1985, is the most cited skill in 531,000 recent tech job postings, surpassing Python (67,000) and SQL (60,000)
- AI-related skills like machine learning and AI appear less frequently, with 31,000 and 25,000 mentions respectively
- Tech firms are offering up to $200,000 premiums and $1 million packages for AI and machine learning experts
Grindr CEO Warns of AI VC Bubble Threatening Innovation and New Companies
Grindr CEO George Arison warned of a “VC bubble” forming around AI applications, risking the destruction of promising companies; he cited SoftBank’s failed investments and noted incumbent firms’ efforts to limit new competition.
- Grindr CEO George Arison stated a “VC bubble” is forming around AI, particularly on the application side, with excessive investment driven by herd behavior.
- Arison warned that many “great companies” could be destroyed due to the VC frenzy, citing SoftBank’s $9 billion investment in WeWork and $375 million in Zume, both of which failed.
- He noted that incumbent AI companies aim to suppress investment in new startups to protect their market position and highlighted rapid innovation among competitors like Elon Musk’s Grok and Claude Code.
TikTok Offers US Salaries Up to $949K Amid 1800+ Open Roles
TikTok offers US salaries up to $949,349 for key roles in AI, e-commerce, and search, with software engineers earning $144,000–$301,158, amid over 1,800 open positions despite regulatory uncertainties.
- TikTok has over 1,800 open roles in the US, focusing on AI, e-commerce, and search.
- Salary offers for foreign hires range from $65,000 (finance representative) to $949,349 (head of product and design).
- Software engineers at TikTok earn between $144,000 and $301,158; data scientists between $85,821 and $283,629; machine learning scientists between $168,000 and $390,000; marketing managers between $85,000 and $430,000.
▶️ Technology
Jeff Dean Warns AI Could Surpass Human Breakthroughs in 20 Years
Google’s Jeff Dean avoids AGI debates due to definitional variability; AI models now outperform humans in many tasks and may accelerate scientific progress within the next two decades.
- Jeff Dean, Google’s chief scientist, avoids AGI discussions due to diverse definitions and problem complexity.
- Current AI models outperform humans in most non-physical tasks but lack human-level expertise across all domains.
- Dean states AI is already close to surpassing human breakthroughs in certain scientific and engineering domains within 5-20 years.
Legal Tech Duopoly Battles AI and Startups Amid Heavy Investments
Legal tech’s duopoly faces increased competition from startups and AI innovations; Thomson Reuters and LexisNexis invest billions, integrating models like GPT-3.5 to enhance legal research and workflows.
- Legal tech giants LexisNexis and Thomson Reuters have historically dominated, with Relx reporting $1.2 billion in legal segment revenue in H1 2025, up 9% YoY.
- Investment in legal software is at an all-time high, with startups like Vlex and Harvey challenging incumbents; Vlex’s $1 billion sale aims to integrate into legal ops platforms.
- Thomson Reuters has accelerated AI adoption by acquiring Casetext and expanding its applied research team to 260 members, aiming to leverage large language models like GPT-3.5 for legal workflows.
AI’s Impact on Culture: Creativity, Trust, and Society
AI automates and expands cultural production, creating new forms and challenges, including risks to originality, societal trust, and the nature of shared human narratives.
- The article explores AI’s impact on culture, creativity, and societal narratives, emphasizing increased cultural output and unpredictable new forms.
- AI-generated content, including podcasts, art, and storytelling, may lead to formulaic art and diminish human originality, but also enables new creative possibilities.
- AI’s influence raises moral, spiritual, and trust issues, with concerns about the erosion of shared societal stories and the potential for manipulated realities.
Rise of Local LLMs Driven by Privacy, Hardware Advances, and Cost Savings
Rising adoption of local LLMs is fueled by privacy concerns, hardware/software improvements, and cost savings, enabling users to run models like Llama and Gemma 2 on personal devices.
- Growing interest in running LLMs locally driven by distrust of big tech, data privacy concerns, and regulatory changes, especially in Europe.
- Hardware advancements (e.g., GPUs with up to 32 GB VRAM, M2 MacBook Pro) and software developments (e.g., llama.cpp, ggml stack) enable local inference.
- Quantization techniques reduce model precision to lower resource requirements with minimal accuracy loss, facilitating deployment on consumer hardware.
WhatsApp Warns of Targeted Attack Exploiting Device Sync Vulnerability
WhatsApp warned of a targeted attack exploiting CVE-2025-55177, a flaw allowing unauthorized device sync, potentially used in sophisticated operations against specific users.
- WhatsApp disclosed CVE-2025-55177, allowing incomplete authorization of linked device synchronization, potentially enabling arbitrary URL content processing on target devices.
- The vulnerability may have been exploited in sophisticated, targeted attacks, similar to a recent zero-click flaw patched by Apple (CVE-2025-43300).
- Attackers likely used these flaws in highly targeted operations, possibly involving commercial surveillanceware vendors, against specific individuals.