multi-modal models AI News & Updates
Hark Secures $700M Series A at $6B Valuation for Secretive AI Personal Assistant Platform
Hark, an AI lab founded by serial entrepreneur Brett Adcock, raised $700 million in Series A funding at a $6 billion valuation to build an agentic AI personal assistant with custom hardware. The company plans to release multi-modal models this summer that will serve as a universal interface with the digital world, followed by dedicated hardware devices. Despite the massive funding round, Hark has revealed few details about its product, focusing on creating AI tools for everyday consumers rather than enterprise software development.
Skynet Chance (+0.04%): Development of an agentic AI system designed as a "universal interface" with autonomous capabilities across digital ecosystems increases surface area for potential loss of control. The privacy concerns mentioned around context-gathering from users' lives highlight alignment challenges in deploying powerful assistive agents.
Skynet Date (-1 days): The $700M investment and focus on GPU compute infrastructure (Nvidia B200s) accelerates deployment of autonomous agentic systems into consumer hands. However, the focus on consumer products rather than frontier capabilities research provides modest acceleration rather than dramatic timeline compression.
AGI Progress (+0.03%): Building agentic systems that can autonomously interface with diverse digital services represents progress toward general-purpose AI capabilities. The integration of multi-modal models with hardware designed specifically for AI assistance demonstrates advancement in embodied and contextual AI systems.
AGI Date (-1 days): The massive capital infusion ($700M Series A) dedicated to compute resources, top talent recruitment, and rapid product development accelerates the timeline for deploying advanced agentic systems. The company's ambition to create a universal digital interface with custom hardware suggests aggressive development pace toward more general AI capabilities.
Wirestock Raises $23M to Supply Multi-Modal Creative Data to AI Foundation Model Makers
Wirestock, a platform that originally helped photographers sell stock photos, has pivoted to become a data provider for AI labs, raising $23 million in Series A funding. The company now supplies images, videos, design assets, and 3D content from over 700,000 artists and designers to six major foundation model makers, achieving a $40 million annual revenue run-rate. Wirestock focuses on providing high-quality, annotated multi-modal data for creative AI applications like image and video generation.
Skynet Chance (0%): This news addresses data supply for AI training, which is a capability enhancement factor, but does not directly relate to AI safety, alignment, control mechanisms, or autonomous decision-making that would affect loss of control scenarios. The focus is purely on commercial data procurement for creative applications.
Skynet Date (+0 days): Improved access to high-quality multi-modal training data could marginally accelerate the development of more capable foundation models, though the focus on creative applications rather than reasoning or autonomous systems limits the impact on risk timeline. The effect on pace toward potentially dangerous AI systems is minimal.
AGI Progress (+0.02%): High-quality multi-modal data is crucial for training more capable foundation models, and this represents improved infrastructure for scaling AI systems across images, video, 3D, and potentially audio modalities. However, this is incremental progress in data supply rather than a fundamental breakthrough in AI capabilities or architecture.
AGI Date (+0 days): The availability of specialized, high-quality multi-modal datasets from a professional platform with 700,000 contributors and $40M revenue run-rate moderately accelerates the pace at which AI labs can train and improve their models. This addresses a key bottleneck (quality training data) but represents evolutionary rather than revolutionary progress in the timeline toward AGI.