AGI

David Luan, head of Amazon’s Artificial General Intelligence (AGI) lab, is leaving the company less than two years after joining via an acqui-hire of his startup, Adept. Luan plans to “cook up something new,” focusing on teaching AI new capabilities as AGI development progresses. His departure follows a reorganization of Amazon’s AGI division, now led by Peter DeSantis. The FTC is scrutinizing acqui-hire deals, including Amazon’s acquisition of Adept, due to potential circumvention of regulatory oversight. Amazon declined to comment.

OpenAI finalized its strategic recapitalization, reinforcing its structure where a nonprofit foundation controls its for-profit PBC. The OpenAI Foundation now holds a $130 billion equity stake in the for-profit arm, ensuring alignment with beneficial AI development. Microsoft maintains a significant stake, valued at $135 billion. Key changes include Microsoft’s broader AGI pursuits and OpenAI’s ability to collaborate with other partners. Microsoft’s IP rights are extended to 2032, excluding OpenAI’s consumer hardware. Both companies emphasize continued collaboration and building impactful products.

OpenAI has restructured, reinforcing the nonprofit’s oversight and establishing the OpenAI Foundation with a $130 billion stake in its for-profit PBC. This channels commercial success towards safe AI development, earmarking $25 billion for global health and AI resilience. A revised partnership with Microsoft includes a $135 billion valuation, granting Microsoft independent AGI pursuit with expert oversight verifying AGI declaration. OpenAI gains flexibility, able to release open-weight models, service U.S. government clients on any cloud, and co-develop select products. The revenue-sharing model remains until AGI validation.

A report indicates a 72.2% drop in DeepSeek’s monthly downloads in Q2 2025. However, this decline is attributed to DeepSeek’s integration into third-party platforms and its focus on open-source LLM development for AGI advancement. Users are being diverted through these channels, impacting direct app downloads. While specialized AI tools like “AI+office” and “AI+education” platforms are gaining traction, DeepSeek remains a key AI player, despite facing challenges like delayed updates and user experience concerns.

OpenAI has paused operations for employees this week, partly in response to Meta’s aggressive recruitment of its researchers. Leadership is actively working to retain staff and is reassessing compensation. The company is refocusing on its primary mission of achieving AGI, even as competitors like xAI push forward with new models.

Sam Altman recounts OpenAI’s evolution from its experimental 2015 origins to leading AI innovation with ChatGPT’s explosive growth (300M+ weekly users). Emphasizing AGI development’s urgency, he acknowledges governance challenges during rapid scale-up, including his 2023 leadership crisis. Balancing speed and precision, OpenAI prioritizes iterative deployment over perfection while tackling safety research proactively. With AGI within reach, Altman shifts focus to superintelligence’s transformative potential and systemic risks, striving to reshape technical frontiers responsibly amid escalating industry competition and existential decision-making.

OpenAI launched Deep Research, an agentic AI feature enhancing ChatGPT’s multi-stage analytical workflows by autonomously synthesizing vetted online sources. Operational in complex domains like financial modeling and supply chain risk, it achieved 26.6% problem-solving accuracy across 3,000 cross-disciplinary questions (vs. 9.4% for competitors) and 72.57% on the GAIA benchmark. While outperforming prior models in rigor and documentation, limitations persist in resolving conflicting data and probabilistic reasoning. Initially available to Pro-tier users, deployment excludes EU jurisdictions, raising compliance concerns for sectors requiring calibrated confidence thresholds.

DeepSeek, a Beijing-based AGI startup, announced plans to open-source five production-grade code repositories, including training frameworks and inference engines, positioning itself as a transparent alternative to Silicon Valley secrecy. Despite its rapid growth—offering GPT-4-level tools freely while monetizing enterprise APIs—the company faces scrutiny over data ties to Chinese state-linked entities, U.S. procurement bans, and IP disputes. Analysts debate whether its “daily unlocks” strategy truly fosters open collaboration or disguises proprietary advantages, as geopolitical tensions intensify in the west-east AI rivalry.

The ARC Prize introduced the ARC-AGI-2 benchmark, targeting AI’s ability to solve novel puzzles with human-like adaptability and efficiency. The 2025 global competition offers $1 million in rewards for systems surpassing 85% accuracy while managing computational costs. Unlike prior benchmarks focusing on brute-force capabilities, ARC-AGI-2 stresses symbolic interpretation, compositional reasoning, and contextual application—areas where current models like OpenAI’s o3 remain inefficient. With human performance at $17/challenge versus o3’s $200/try, the new standard underscores economic viability as a critical milestone toward practical AGI development.

A multidisciplinary coalition warns that OpenAI’s proposed shift to a Delaware public benefit corporation (PBC) risks undermining critical safeguards for responsible AGI governance. Critics argue the change weakens the original nonprofit’s authority by introducing profit-sharing incentives, diluting accountability mechanisms, and increasing commercial influence over AGI development. While OpenAI defends the restructuring as necessary for competitiveness, opponents emphasize that the organization’s founding commitment to prioritize societal benefit over shareholder returns must remain legally enforced. The debate centers on maintaining structural independence and robust oversight amid AGI’s unprecedented global impact potential.