Good morning from OWITH.ai: the podcast that gives you only what's important to hear in the AI and tech world. The conversation begins with the recent excitement surrounding SpaceX's IPO, poised to make history with a staggering valuation between $1.5 trillion and $2 trillion. Alongside potential IPOs from OpenAI and Anthropic, this event is seen as a monumental milestone for the venture capital industry, potentially introducing $5 trillion in value. However, insights from Nizar Tarhuni of PitchBook suggest that while these IPOs are significant, they don't represent a watershed moment for the VC market. Despite the buzz, a mere 700 out of 11,000 venture investors have stakes in these companies, indicating that 80% of active VCs won't directly benefit from these liquidity events. These companies stand out due to their enormous valuations and public offerings that dwarf most others from the past two decades. Their need for capital, especially for computing power, surpasses what private markets can typically support. Hans Tung from Notable Capital observes that top-tier limited partners are already involved with these firms and will likely pursue more direct access to future key AI players, funneling capital towards a select group of winners. Mo Jomaa from CapitalG compares IPO windows to cyclical roller coasters, emphasizing that while they open and close, the investments made in preparation for IPOs remain timeless. SpaceX prioritizes growth stories because they tend to attract more investor interest than mere stock purchases. Transitioning now to other notable developments in technology and business... The focus shifts to key advancements in the space economy, scientific superintelligence, and hiring trends at Anthropics. Bridgit Mendler, CEO of Northwood, emphasizes the increasing mainstream impact of the space economy due to advancements in launch capabilities and satellite manufacturing. She underscores the importance of ground infrastructure in connecting Earth and space, drawing parallels to cloud infrastructure for startups. Mendler argues that as data throughput increases, so does the space economy's value, signifying a shift towards broader adoption beyond niche markets. Concerns are also raised about U.S. science funding cuts, particularly in healthcare and biotechnology. Industry leaders warn that defunding risks leaving the U.S. trailing its global rivals. Geoffrey von Maltzahn of Lila Sciences highlights the urgency of investing in scientific superintelligence—a concept where AI systems continuously engage in the scientific method across various fields. On hiring trends within Anthropics, Boris Cherny outlines three desirable traits for applicants: generalism with interdisciplinary knowledge, low ego to foster collaboration, and empiricism to learn from data-driven feedback. These traits align with a culture of humility and adaptability within teams. Meanwhile, other tech news includes potential layoffs at Microsoft's Xbox division, Google's new AI model for faster text generation, FBI actions against fake consulting domains linked to China, OpenAI's potential price reductions, Amazon's significant loan for AI infrastructure development, and legislative efforts in Canada to restrict social media access for minors. Moving now to challenges faced by CEOs in governing AI... The discussion turns to challenges CEOs encounter in AI governance, focusing on Anthropic's release of its Mythos-Class Fable 5 model. A financial services CEO expressed concerns over this development, emphasizing complications rather than potential weaponization. This reflects a broader issue for CEOs: navigating ever-evolving AI regulations and the autonomy of companies like Anthropic in setting these standards. Anthropic's Claude Fable 5 model presents both opportunities and challenges, automatically blocking certain responses to maintain safety but also limiting capabilities for researchers and developers. Its lack of transparency is notable; users receive no indication when content is compromised. Concerns about data retention also arise; unlike previous models offering zero data retention, Claude Fable 5 retains data for 30 days, leading companies like Microsoft to limit its use. Cybersecurity remains a pressing concern. Bezalel Eithan Raviv from Lionsgate Intelligence Network criticizes inadequate security measures for such AI models. He argues these models should be treated like currency or weapons due to their potential misuse by cybercriminals, underscoring the need for regulation in this rapidly advancing field. In closing with some business news... The conversation touches on various topics: a federal judge overturned Trump’s $100,000 H-1B visa fee affecting AI firms like OpenAI and Nvidia; Meta launched a training program for data center technicians amid U.S. data center expansion; inflation rose to 4.2%, driven by gasoline prices linked to geopolitical tensions; and markets showed mixed performances with some indices rising while others declined. Additionally noteworthy is the release of the Fortune Crypto 100 list ranking influential companies within the digital asset ecosystem—featuring both crypto-native firms and traditional financial giants—and MasterCard's launch of a new protocol enabling AI agents to transact with each other. Overall reflections capture complexities within AI governance and highlight significant business developments across various sectors as we continue observing these dynamic changes within the industry landscape.Support the show Thanks for listening! Follow us on Twitter, Instagram and Linkedin
