Coinbase is testing AI agents that appear on Slack and email, with CEO Brian Armstrong predicting these agents will soon conduct more on-chain transactions than humans. He also forecasts that AI agents will outnumber Coinbase employees shortly, highlighting the company's focus on integrating AI technology into its operations. The initiative underscores the growing role of AI in the crypto industry.
SK Hynix's stock rose 3.4% after beginning mass production of its 192GB SOCAMM2 memory module for Nvidia’s Vera Rubin AI platform. The module aims to reduce power consumption and eliminate memory bottlenecks. Nvidia's stock was down 1.4% in pre-market trading. The production start marks a significant step in AI hardware development as of April 2026.
This Web3 news article is unrelated to AI technology. It discusses the best hybrid mattresses for various needs, highlighting how combining foam and springs creates a balance of comfort and support. The article emphasizes extensive testing to identify top options for couples, back pain, and other preferences, but does not mention specific companies, amounts, or dates related to AI or Web3 developments.
Marvell's stock surged up to 6.3% premarket after reports of talks with Google to develop two AI chips. One is a memory processing unit for Google’s TPU, and the other is a new TPU for AI inference. The partnership aims to strengthen Google’s AI hardware capabilities amid growing industry demand.
A longitudinal study revealed that loneliness in older adults is more strongly associated with memory lapses, including immediate and delayed recall, than with overall cognitive decline. The research highlights the impact of social isolation on memory function, emphasizing the importance of social connections for cognitive health in aging populations. The findings underscore loneliness as a key factor in age-related memory issues.
Chinese tech workers are being instructed to train AI agents to replace them, sparking debate among early adopters. A GitHub project called Colleague Skill, launched earlier this month, enables users to distill colleagues’ skills and traits to create AI doubles. This development highlights growing concerns over job security and ethical implications in AI adoption within China’s tech industry.
Google shares increased following reports of discussions with Marvell on developing custom AI chips for cloud and AI workloads. The collaboration aims to create new TPU designs and memory processing units to enhance AI performance and efficiency. This move aligns with a broader industry trend of hyperscalers building in-house chips to reduce reliance on Nvidia GPUs.
