Welcome to In the Loop, TIME’s new AI-focused newsletter, published twice weekly. These editions are available as stories on Time.com and via email. If you’re currently viewing this in your browser, consider subscribing to receive future editions directly in your inbox.
Who to Know: Larry Ellison
Larry Ellison may not be the first name that comes to mind among the world’s most prominent tech figures. However, this past Wednesday, the 81-year-old Oracle chairman briefly attained the title of the world’s wealthiest individual, with his net worth nearing $400 billion, surpassing Elon Musk.
Ellison’s net worth surged by $100 billion, marking the largest single-day increase ever. This remarkable gain followed Oracle’s optimistic growth projections, which highlighted hundreds of billions of dollars in anticipated revenue from AI companies utilizing Oracle’s cloud computing services. Ellison’s dramatic ascent can be linked to both the current AI boom’s market forces and his distinct corporate financing strategies.
Lower in the Stack — Oracle does not exhibit strong performance across all metrics. The company failed to meet analysts’ revenue forecasts and anticipates three more years of negative cash flow. An industry publication reported recent substantial layoffs within the company.
However, the current phase of the AI hype cycle sees AI companies developing products, rather than end-users, as the primary spenders. Consequently, companies positioned lower in the technology stack, such as chipmakers like NVIDIA and Oracle—which manages data centers offering cloud infrastructure to other AI firms—are benefiting most significantly.
Oracle’s proficiency in addressing the immense logistical hurdles involved in constructing superclusters liberates AI companies from operational pressures, positioning them as an attractive collaborator. Furthermore, their pricing is frequently more competitive than rivals. This has led to the signing of contracts valued at hundreds of billions of dollars with entities such as OpenAI and xAI. When Oracle announced these agreements and projected future revenue on Wednesday morning, the company’s stock price soared by approximately 38%.
Networking Advantages — Ellison’s robust connections with other influential global figures undoubtedly play a role. He hosted a fundraiser for Trump in 2020, was observed at Mar-a-Lago during Trump’s fundraising activities last year, and in January, became a key partner in Trump’s Stargate project, which seeks to develop $500 billion in AI infrastructure across the U.S. Ellison has also maintained a long-standing close relationship with Musk: In 2022, he reportedly offered via text message to contribute a billion dollars, or “whatever you recommend,” towards Musk’s endeavor to acquire Twitter.
Stock Buyback Surge — While numerous companies are significantly benefiting from the AI boom, Ellison’s exceptionally large profits warrant explanation. The reason stems from the substantial amount of Oracle stock he holds. Ellison possesses a 41% stake in Oracle, almost twice his ownership from fifteen years prior. This increase is attributed to him and Oracle undertaking one of the largest stock buyback initiatives in corporate history over the past decade, occasionally securing loans to fund these share repurchases.
Investors favor stock buybacks as they enhance the value of their existing shares. However, the strategy faces considerable criticism, with detractors arguing it favors immediate profits over sustained investment in infrastructure or research and development. Chuck Schumer famously described stock buybacks as “one of the most self-serving things that corporate America does.”
Given Ellison’s significant ownership of Oracle stock, his net worth fluctuates directly with the share price. This explains why his wealth momentarily surpassed Elon Musk’s this week, only to revert to approximately $376 billion the following day.
What to Know: Regulatory sandbox proposal
Ted Cruz has consistently been a strong proponent for the AI industry in Congress, advocating for minimal intervention to foster growth and innovation. This Wednesday, he proposed legislation designed to shield AI companies from regulation while they test novel technologies.
The proposed bill would establish a “sandbox” framework, allowing companies to apply for waivers from specific regulations for two-year terms, renewable for up to a decade. Cruz garnered backing from OSTP director Michael Kratsios, who attended Cruz’s hearing on Wednesday and asserted that anti-innovation regulations posed “a huge problem for our industry.”
Cruz’s bill faces significant obstacles to passage. Numerous Republican senators, such as Marsha Blackburn and Josh Hawley, are eager to enact regulations aimed at mitigating AI-related harms. Concurrently, left-leaning groups reacted with strong disapproval. The consumer rights advocacy organization Public Citizen stated that the proposal “lets companies skirt accountability, and treats Americans as test subjects.”
AI in Action
Hours following the assassination of right-wing commentator Charlie Kirk in Utah, President Trump released a video from the White House on social media, denouncing the event. However, questions quickly arose regarding its authenticity. Within minutes, allegations of the video being AI-generated circulated online, as amateur investigators highlighted anomalies such as Trump’s finger briefly vanishing and the video’s oversaturated hues.
Nevertheless, experts have expressed doubt regarding this theory. Hany Farid, co-founder of the cybersecurity firm GetReal Security, stated on LinkedIn, “We have reviewed this video and find no evidence that the audio or video is AI-generated.” He suggested the finger distortion could be a result of “localized video manipulation.”
Regardless, the public reaction to the video underscored the increasing challenge of trusting online visual content, a consequence of progress in AI deepfake technology. Fabricated videos are perceived as genuine, while authentic ones can be readily dismissed as fakes. (Trump himself previously labeled a video of a bag being tossed from the White House as “probably AI-generated.”)
Companies such as Google have introduced watermarking systems to distinguish between genuine and altered videos. Yet, thus far, viral videos appear to disseminate more rapidly than they can be authenticated. Amidst this confusion, it seems probable that individuals will increasingly adhere to their existing beliefs regarding truth and falsehood, thereby contributing to a more fragmented and contentious online environment.
What to Read
Recommended Reading: Alex Reisner, The Atlantic
Reisner reveals that AI companies have been acquiring millions of YouTube videos, including a significant number of ‘how-to’ guides, to train their artificial intelligence systems. By drawing knowledge from these skilled individuals, AI can now replicate their role in providing guidance, potentially making the original human experts redundant.