Maria Anselmi

“The central dynamic of the fusion race is clear: the United States laid the scientific groundwork; China is positioning to win the industry… The nation that leads in fusion will secure significant economic advantages, ensure its energy independence, and maintain its technological edge in critical areas, including AI and national security.” So concludes the Special Competitive Studies Project - SCSP’s Commission on Scaling Fusion Energy in a report released today (see link below) that urges the United States to make rapid fusion energy commercialization a top priority: 1️⃣ The president should declare fusion energy to be a national security priority through an executive order. 2️⃣ The Department of Energy should support fusion energy research, development, and demonstration with a $10 billion investment. That would fund new infrastructure to close technical gaps, public-private partnerships that de-risk private sector fusion designs, and a new tier in DOE’s Milestone-Based Fusion Development Program to demonstrate  power plants. 3️⃣ The federal government should take strategic actions to support efficient regulations, lead in using digital design tools like AI, and develop a fusion supply chain and workforce. Why this big push? Among the report’s findings: ▶️ “China, recognizing fusion’s strategic importance, is mobilizing massive state resources not only to pursue scientific breakthroughs but also to build the infrastructure, supply chains, and industrial capacity required for deployment at scale — a proven strategy Beijing has already used to lead sectors like solar panels and advanced batteries.” ▶️ "Since 2023, China has mobilized over $6.5 billion across its fusion enterprise, almost three times the funding appropriated to the Department of Energy’s Fusion Energy Sciences (FES) Program in roughly the same period of time." ▶️ “China is producing ten times more fusion PhD graduates than the United States, a direct function of government funding.” ▶️ “The United States’ scientific community identified the important infrastructure in 2020… but DOE has not executed this plan. China, on the other hand, is backing its fusion goals with massive infrastructure projects.” Read the report here: https://lnkd.in/ePfY7j-K #FusionEnergy #China

285

31 Kommentare

Rust Foundation Executive Director & CEO Rebecca Rumbul announced today on stage at #rustconf 2025 the launch of the Foundation’s new initiative: the Rust Innovation Lab. The Rust Innovation Lab is designed to provide comprehensive support—including governance, legal, networking, marketing, and administrative resources—for funded open source projects, ensuring their long-term sustainability and community-led growth. We are proud to welcome Rustls as our inaugural hosted project under the Rust Innovation Lab. Read the full announcement: https://lnkd.in/gSkZVb9n #opensource #rustconf25 #rustls #rustfoundation

135

2 Kommentare

Decentralized AI training has gone from theory to reality. Over the past year, projects like Nous Research, Prime Intellect, Pluralis Research, templar, and Gensyn have shown it’s possible to: ▪️ Coordinate GPUs globally for model training ▪️ Reduce communication overhead with novel optimizers ▪️ Introduce crypto-economic incentives and verification layers Why it matters:  ▪️ Access: Taps into unused GPUs worldwide, opening advanced AI development to more researchers.  ▪️ Openness: Challenges centralized labs’ control over foundational models.  ▪️ Foundation: A critical step toward building a fully onchain AI stack. Our new Galaxy Research report, Decentralized AI Training: Architectures, Opportunities, and Challenges, explores:   ▪️ Architectures powering decentralized training protocols  ▪️ The technical and economic problems they aim to solve  ▪️ Opportunities and risks as the field scales Decentralized training is no longer just a concept. It’s happening now. The big question: can it deliver practical advantages over centralized labs? 📖 Read the full report by Lucas Tcheyan: https://lnkd.in/eA4KVeaG 

69

2 Kommentare

We were at the SDMX Global Conference in Rome, presenting Documind together with Bank of Greece - showing how GenAI can assist in generating structural metadata for SDMX. Why it matters: metadata modelling has traditionally been slow and resource-intensive; with AI, it can become faster, smarter, and more accessible, opening the door for wider adoption of SDMX. Here are some insights from the conference: 🤖 This is no longer experimentation — AI is becoming a core interface between humans and statistical data, lowering technical barriers and accelerating analysis. We saw live examples of how GenAI can draft analytical reports (ILO) or navigate SDMX datasets (BIS). 🛡️ Standardisation is the backbone of trust in the AI era — Across all the world from National Statistical Agencies to Central Banks, case studies showed how SDMX ensures that data is consistent and interoperable. In an age where AI can generate any output, only standardized and trusted data guarantees credibility and comparability. 🌍 Cross-domain interoperability is moving from vision to reality — UNICEF linking SDMX to geospatial systems, Eurostat advancing statistical ontologies, and ECB showcasing semantic search all point to the same trend: data silos are dissolving. combining economic, social, and environmental datasets is essential to tackle today’s multidimensional policy challenges. ⚙️ Modernisation and automation are no longer optional — FAO, Bank Indonesia, and Statistics Netherlands shared how SDMX-based automation reduces manual burden and improves quality assurance. With shrinking resources and increasing demand for timely data, automation is becoming the only sustainable way to keep statistical systems relevant. Thanks to George Verouchis for holding our presentation! Let us know your opinion on these topics in the comments below! #SDMX #SDMXconference #doneai

36