Latest AGI News & Updates
Claims, timelines and debates about artificial general intelligence โ from lab roadmaps to skeptic takedowns.
69 articles
America Builds AGI, China Deploys AI
The article contrasts the United Statesโ ambition to pursue AGI and autonomous intelligent agents with Chinaโs emphasis on deploying AI in factories, hospitals, vehicles, government, and retail. It argues that the two approaches reflect different capital models: long-horizon, vision-driven funding in the U.S. versus cash-flow and commercialization pressure in China.
Anthropic discovers 'J-space' in Claude models
Anthropic researchers identified an internal 'J-space' in Claude that governs reasoning and concept representation, which emerged spontaneously during training, suggesting a potential path toward AGI.
The Rise of Recursive Self Improvement in AI
AI models are increasingly capable of modifying their own code and training processes, a phenomenon known as Recursive Self Improvement (RSI). Major labs like Anthropic and OpenAI report that AI agents are now handling significant portions of their own development and safety research.
DeepMind CEO proposes US-led Frontier AI regulatory body
Google DeepMind CEO Demis Hassabis proposes a US-led regulatory body for frontier AI models, modeled after the financial industry's FINRA. The goal is to mandate safety evaluations before public release to ensure AGI development remains secure.
Zhipu AI aims to overcome three major AGI hurdles
Zhipu AI is positioning itself as China's answer to Anthropic by focusing on advanced AGI development. The company is actively addressing three critical technical challenges to scale its intelligence capabilities.
Zhipu and MiniMax pivot to long-term AGI research
Following their IPO lock-up expiration, Chinese AI leaders Zhipu and MiniMax have issued internal letters committing to long-term AGI development over short-term commercial gains. Both companies are securing massive funding to invest in foundational model research and infrastructure.
Bezos-backed Prometheus raises $12B for AGI engineering
Prometheus, an AI company backed by Jeff Bezos, has secured $12 billion in Series B funding. The company aims to develop 'Artificial General Engineers' to automate complex technical problem-solving.
Bezos-backed Prometheus raises $12B for physical product AI
Prometheus, co-led by Jeff Bezos, raised $12 billion at a $41 billion valuation. The company is building an 'artificial general intelligence' focused on engineering physical products.
Sundar Pichai on the future of AI-driven search
Google CEO Sundar Pichai discusses the integration of Gemini models into Search and YouTube, shifting the focus from delivering links to executing tasks. The conversation highlights the industry's move toward a 'zero-click' search environment.
OpenAI hiring researchers for recursive AI self-improvement
OpenAI is hiring safety researchers with salaries up to $445k to focus on recursive AI self-improvement. The team aims to develop tools for model interpretability and defense against data poisoning.
AI's disruption: Beyond productivity to human value
AI is shifting from a productivity tool to a disruptor of human cognition and creativity, threatening to create a 'functionally useless' class. Society must urgently address wealth redistribution and the definition of human value through new institutional frameworks.
Startup raises $650M for self-improving AI systems
A four-month-old startup has secured $650 million in funding to pursue the development of recursive self-improving AI. The project aims to create systems that can accelerate their own development cycles beyond human capacity.
Researchers Launch $4B Effort for Self-Improving AI
A new venture called Recursive Superintelligence has been formed by former researchers from Google, Meta, and OpenAI. The company aims to automate the AI development process through a massive $4 billion funding effort.
Gaode Unveils AGI Embodied Tech Stack
Gaode (Amap) publicly releases its first full-stack embodied technology system aimed at AGI. The system dominates 15 global SOTA benchmarks. It signals rapid convergence as embodied AI infrastructure matures.
GPT-6 Leaked as AGI Pursuit
GPT-6 has been exposed in leaks. The model is fully geared toward achieving AGI. This represents OpenAI's bold next step in frontier AI development.
ARC-AGI-3 Resets Frontier AI Scoreboard
ARC-AGI-3, the latest iteration of the AGI benchmark, has reset the leaderboards for frontier AI models. This update challenges systems on abstraction and reasoning with novel tasks. A side note mentions tools for creating branded reaction GIFs in company Slack.
Huang: AGI Here, China AI Fastest
NVIDIA CEO Jensen Huang tells Lex Fridman AGI is already achieved, China leads AI evolution via implicit open-source, global programmers will hit 10B, and unveils NVIDIA's trillion-dollar AI factory strategy with extreme co-design.
ARCANA: Multi-Agent Framework for ARC-AGI-2 Reasoning
ARCANA is a collaborative multi-agent framework designed to solve complex ARC-AGI-2 tasks under strict constraints. It utilizes a combination of object-centric perception, latent program policy, and reflective feedback to improve reasoning efficiency.
Zhipu AI Launches 'Touch High' AGI Research Initiative
Zhipu AI founder Tang Jie announced the 'Touch High' initiative, shifting focus from short-term commercialization to long-term AGI research. The company will prioritize four key areas: long-horizon tasks, autonomous agent systems, self-evolving models, and AI safety.
Measuring AI Intelligence Beyond Human Capability
This research proposes a new paradigm for evaluating AI models that have surpassed human capability by using relative measurement instead of static benchmarks. The framework utilizes model-generated challenges to create an adversarial psychometric rating system that scales alongside agent capabilities.
From AGI to ASI: Pathways and Future Transitions
This research report explores the theoretical transition from human-level Artificial General Intelligence (AGI) to Artificial General Superintelligence (ASI). It identifies four primary development pathways and discusses the potential for a series of transformative societal changes rather than a single event.
LeCun's Vision: The Race for Latent World Models
Yann LeCun is heavily betting on world models as the next frontier in AI. Leading visual AI teams are already actively researching and implementing these complex latent space architectures.
DeepSeek targets AGI in $10bn funding round
DeepSeek founder Liang Wenfeng has confirmed that the lab is prioritizing AGI development over immediate revenue generation. The company plans to continue its open-source model strategy as it seeks its first major external funding round.
DeepSeek Targets AGI in Massive $10 Billion Funding Round
DeepSeek is currently raising $10 billion in a new funding round to accelerate its pursuit of Artificial General Intelligence (AGI). Management has signaled to investors that the company will prioritize long-term research breakthroughs over immediate commercial monetization.
AI Cognition Evolves Unevenly Across Generations
Researchers developed a psychometric framework to assess generative AI cognitive profiles against human norms, revealing top performance in verbal and working memory but failures in perceptual reasoning. The AIQ Benchmark tracks evolution across six generations and two model families, showing asymmetric gains with language bias over visual tasks. This indicates scaling alone may not achieve balanced AGI.
Musk-Anthropic alliance ushers heavy AI era
Elon Musk allies with Anthropic, marking large model wars entering 'heavy industry era'. AGI dominance demands efficient hardware, cheap sustainable energy, and chip manufacturing autonomy.
Hassabis: AGI Via Learning & Memory
DeepMind CEO Demis Hassabis stresses continuous learning and memory over context window expansion for AGI. He advocates RL for reasoning, small end-side models, and AI-driven scientific breakthroughs. Shares insights on startup strategies amid AGI timeline.
Gaode Launches ABot: First AGI Embodied Tech Stack
Gaode has released ABot, the world's first full-stack embodied technology system designed for AGI. It achieves 15 state-of-the-art (SOTA) results across key benchmarks. The system builds a continuously evolving closed loop for embodied intelligence.
Yann LeCun Raises $1B for Physical World AI
Yann LeCun, Metaโs former chief AI scientist, has raised $1 billion for his new startup AMI. He argues human-level AI requires mastering the physical world, not just language. AMI aims to prove this vision.
Amazon $50B OpenAI Bet Tied to IPO, AGI
Reports indicate Amazon's potential $50 billion investment in OpenAI depends on OpenAI's IPO and AGI progress. The deal's viability hinges on these milestones. Sourced from The Information via 36Kr and Sina Finance.