- NATURAL 20
- Posts
- Weaponized AI: China’s Global Play
Weaponized AI: China’s Global Play
PLUS: SoftBank Plans $1T Robot Investment, Scale AI Eyes $25B Valuation and more.

Find out why 1M+ professionals read Superhuman AI daily.
In 2 years you will be working for AI
Or an AI will be working for you
Here's how you can future-proof yourself:
Join the Superhuman AI newsletter – read by 1M+ people at top companies
Master AI tools, tutorials, and news in just 3 minutes a day
Become 10X more productive using AI
Join 1,000,000+ pros at companies like Google, Meta, and Amazon that are using AI to get ahead.
Today:
Weaponized AI: China’s Global Play
xAI Acquires X For $33B
OpenAI Nears $40B Funding Round
SoftBank Plans $1T Robot Investment
Scale AI Eyes $25B Valuation
China's "Weaponized" Open Source AI and US Tech Collapse…
China is rapidly advancing in robotics and AI, often open-sourcing its work. A viral robot dancing video took a turn when it was shown wielding axes—amazing, but unsettling. China’s strategy is clear: release powerful, free AI models to undercut U.S. companies, then profit from selling cheap AI-enabled hardware like drones and robots.
This open-source blitz mirrors China’s broader tactic—copy, scale, and dominate through low prices. Tools like DeepSeek and Jarvis VAA show real progress, even in games like Minecraft. While the U.S. leads in software, China is gaining fast, threatening to reshape the global AI infrastructure and economy.
Elon Musk’s AI company xAI has acquired his social media platform X (formerly Twitter) in a $33 billion all-stock deal. Musk aims to merge xAI’s technology with X’s wide user base to create smarter digital experiences. Despite past financial losses, Musk sees the move as a step toward building a powerful, AI-driven platform. The integration also echoes his earlier merger of Tesla and SolarCity under a similar leadership structure.
Why it matters
Massive data access: xAI now controls data from over 600 million X users—fuel for AI training.
Platform power: Integration gives xAI a ready-made distribution channel for AI tools like Grok.
Influence consolidation: It shows how tech giants might consolidate platforms to accelerate AI development and reach.
OpenAI is close to securing a $40 billion funding round led by SoftBank, with $30 billion directly from SoftBank and $10 billion from others, including Microsoft. The deal hinges on OpenAI fully becoming a for-profit company by year’s end. The partnership includes a joint venture in Japan and $3 billion in annual spending by SoftBank on OpenAI tools, aiming to spread AI adoption and possibly speed up progress toward general AI.
Why it matters
Massive funding boost: A $40B injection accelerates OpenAI’s development and infrastructure.
Global expansion: The Japan joint venture shows strategic localization of AI enterprise tools.
AGI ambition: SoftBank’s belief in near-term artificial general intelligence pushes the global AI race forward.
SoftBank plans to invest $1 trillion to build AI-powered robot factories across the U.S., aiming to help manufacturers facing worker shortages. These factories will be located in industrial parks and produce general-purpose robots capable of handling varied tasks. The move expands SoftBank’s existing AI initiatives, including the Stargate data center project and investments in Skild AI. This massive investment signals a bold bet on AI transforming labor and manufacturing in the U.S.
Why it matters
Scale of investment: A $1 trillion bet shows unmatched confidence in AI’s industrial future.
Robot workforce: It accelerates the shift toward AI-driven automation in manufacturing.
U.S. focus: Highlights the U.S. as a key arena for AI infrastructure, jobs, and innovation.
🧠RESEARCH
Video-R1 is a new method that helps large language models better understand and reason about videos. It solves problems with time-based understanding and lack of training data by using a new algorithm and combining video with image reasoning data. It outperforms even top commercial models on key video benchmarks.
UI-R1 uses rule-based reinforcement learning to help AI better predict actions on mobile screens. With a small but high-quality dataset and a new reward system, the model improves accuracy on both familiar and new tasks. It even matches
This survey reviews how large language model agents are built, how they work together, and how they improve over time. It organizes recent research, highlights key challenges, and explores real-world uses. The paper offers a clear roadmap for understanding and developing intelligent agents that can adapt and act toward goals.
OlymMATH is a new math benchmark made to truly test how well language models can reason. It includes 200 tough, Olympiad-style problems in both English and Chinese. Early results show that even top models struggle with the hardest questions, revealing gaps in current AI math skills and evaluation tools.
VBench-2.0 is a new benchmark for testing how realistic AI-generated videos truly are—not just visually, but in terms of physics, logic, and human understanding. It evaluates five key areas like human likeness and commonsense. This sets a higher standard for future video models aiming for real-world accuracy.
🛠️TOP TOOLS
Ecrett Music - AI-powered music composition platform designed specifically for content creators, including video producers, game developers, podcasters, and other digital content makers.
Hubble AI - Designed to make artificial intelligence (AI) accessible and integrable into various business operations without the need for coding expertise.
GhostCut - AI-driven video authoring tool designed to enhance localized video marketing and simplify professional video production.
Neeva AI - Search engine that stood out for its commitment to user privacy and an ad-free experience.
Echo Labs - AI-powered tool designed to transform the way digital content is created, particularly in the realm of video production.
📲SOCIAL MEDIA
an Ghibli avatar of me explains the Ghibli trend while the history of the world scrolls by in Ghibli images...
— Wes Roth (@WesRothMoney)
4:25 AM • Mar 29, 2025
🗞️MORE NEWS
Scale AI may seek a $25 billion valuation in a share buyback offer, more than doubling last year’s figure, amid rising demand for AI tools. U.S. labor officials are also investigating the company.
Zendesk is testing AI agents built with OpenAI to handle full customer service conversations. These agents reason through problems, adapt responses, and follow company rules—cutting setup time and aiming for 80% automation.
Midjourney and NYU created a new AI training method that boosts creative writing variety by 23% with only a small drop in quality. It compares how different responses are to increase originality in storytelling.
Perplexity is partnering with firmly.ai to let users shop directly within its AI platform. This integration helps buyers complete purchases faster and gives merchants instant access to a large audience without complex setup.
Researchers at Singapore Management University created AgentSpec, a rule-based system that keeps AI agents from making unsafe or unintended decisions. It adds safety checks during tasks and works across popular agent frameworks like LangChain and AutoGen.
A new book excerpt reveals that Sam Altman was briefly fired as OpenAI CEO in 2023 due to concerns over secretive behavior, including owning a company fund. Internal backlash led to his swift return.
OpenAI has loosened content rules for ChatGPT’s new image generator, now allowing depictions of public figures and sensitive topics like racial traits and symbols in educational contexts. The shift aims to increase user control and reduce blanket refusals.
What'd you think of today's edition? |
Reply