.png){kind=logo}
What’s New Today: China breaks ground with the world’s first mass-produced hybrid stochastic AI chip designed for high efficiency and real-world applications.
Fast-Track Insights: Sweden’s Lovable, an AI no-code platform, is nearing a $1.5 billion valuation with a $100 million funding round—just six months after launching.
Here’s a quick rundown of the biggest tech headlines making waves today. Let’s dive into the day’s top stories, from China’s new AI chip to Lovable’s rapid rise in the global startup scene.
China’s researchers at Beihang University have introduced the world’s first large-scale “non-binary” AI chip using Hybrid Stochastic Number (HSN) computing. By combining traditional binary logic with probabilistic methods, this chip delivers enhanced energy efficiency, fault tolerance, and in-memory computing capabilities. Manufactured on 110 nm and 28 nm SMIC processes, it’s already being used in aviation, smart displays, and industrial control systems—sidestepping U.S. export restrictions.
Lovable, a Swedish AI-driven no-code platform, is targeting a $100 million funding round at a $1.5 billion valuation just six months after launch. It enables non-technical users to build production-grade apps via natural language, leveraging OpenAI, Gemini, and Anthropic models. With $60 M ARR, half a million users, and 85 % retention, Lovable is outpacing traditional dev tools.
Indian students at Harvard are enduring a “rollercoaster” of uncertainty as Trump‑era visa restrictions, a $2.2 billion funding freeze, revoked SEVP certification and a hesitant job market threaten their academic and professional futures. Some are considering returning to India or migrating elsewhere, yet peer networks and university support help them persist amid turmoil.
In 2025, AI’s habit of “hallucinating”—confidently generating false or fabricated content—undermines its credibility. Gaps in training data and model complexity fuel these errors. Hallucinations risk spreading misinformation in critical areas like medicine, research and journalism. Tackling them requires better datasets, model transparency, design that flags uncertainty, and human oversight to restore trust.
Russia’s Ministry of Digital Development proposes empowering courts and law enforcement to confiscate coins and impose fines on unlicensed crypto mining. Penalties range from 100,000–400,000 ₽ for individuals and up to 2 million ₽ for corporations. The draft law also mandates mining infrastructure operators report activities, and criminalizes unauthorized crypto payments, signaling a tougher stance on illicit digital asset activity.
