top of page
Urban Clouds
    • Aug 30

👉Issue #81 of QX Snapshots is out!




Welcome to QX Snapshots - a weekly recap of the key news on emerging technologies. In this newsletter, you will get a "digest" of the latest info on Enterprise Blockchain, AI, Quantum Technology. Hope it brings you value :)


If you are interested in emerging technologies and you like what you read, consider subscribing to our newsletter on LinkedIn!


Every Friday we share key snapshot news on AI, Quantum Technologies, Metaverse, and Enterprise Blockchain/DLT.


Join our ~500 trusted readers at U.S Bank, Eurasian Development Bank, EY, Morgan Stanley, BNY Mellon, Microsoft, Verizon, Facebook, ServiceNow, ConsenSys, Polygon Labs, Duke University, Sky, Informa, Freshfields Bruckhaus Deringer, Santander, and others.





[Blockchain] Natixis and DekaBank Pioneer Germany's First Blockchain-Based Covered Bond; Hong Kong's Tokenization Sandbox Attracts Major Banks; Centrifuge Debuts Institutional Asset Lending on Coinbase’s Base. Natixis Pfandbriefbank and DekaBank issued Germany's first digital registered covered bond on the SWIAT blockchain, with a €100 million volume. The blockchain-based process enables same-day settlement, improved tradability, and reduced risks. SWIAT, a joint venture of DekaBank, LBBW, and SC Ventures, provided the blockchain infrastructure. DekaBank aims for €1 billion in issuance volume on SWIAT by 2025. Meanwhile, the Hong Kong Monetary Authority launched its "Project Ensemble" tokenization sandbox, attracting major institutions like HSBC. The sandbox aims to develop a comprehensive tokenization market architecture for faster, scalable transactions. HSBC completed three proof-of-concept tests, including digital bond purchases using tokenized deposits, interbank transfers, and electronic bills of lading settlements. The project, designed to future-proof its financial system, also involves Hang Seng Bank, Ant Digital Technologies, and Global Shipping Business Network. Also, Centrifuge has launched an institutional real-world asset lending market on Coinbase's Base network, using Morpho Vaults. The market allows borrowing against three types of tokenized U.S. Treasury bills: Centrifuge's Anemoy fund, Midas' mTBILL, and Hashnote's USYC. This initiative aims to provide instant liquidity without redeeming the underlying T-bills. It's the first use of Ethereum-powered Coinbase Verifications in a permissioned lending market. The collaboration showcases the growing trend of tokenizing traditional financial instruments on blockchains, offering new utility for these digital assets. And, Bloomberg added Polymarket data to its Terminal. Polymarket is the world's largest prediction market. It is a decentralized information markets platform where people trade real-money markets on the outcomes of the most highly debated current events and follow the odds to garner accurate insights about the future. It is built on Polygon.



[AI] OpenAI Unveils AI Strawberry and Orion; Magic Boosts AI for Software Development; Google Research on Game Engines with AI.  OpenAI is developing a new AI called Strawberry (formerly Q*) to enhance its chatbot capabilities and to develop Orion. Strawberry can solve complex math and programming problems, and reason through subjective topics. It's being integrated into OpenAI's upcoming Orion large language model and may launch in chatbot form this fall. This development is part of OpenAI's strategy to maintain its lead in the competitive AI market, where its revenue has tripled to $283 million monthly. The company is also seeking additional funding and exploring ways to reduce costs. Apple and Nvidia are reportedly in talks to invest in the company. Meanwhile, Magic has developed an AI model with a 100M token context window, capable of understanding 750 novels worth of text. This breakthrough enhances code synthesis and AI agent functionality. The company introduced HashHop, a new evaluation method for long context models. Magic's LTM-2-mini model is significantly more efficient than competitors in processing and memory usage. The company is partnering with Google Cloud and NVIDIA to build new AI supercomputers, Magic-G4 and Magic-G5. Magic has raised $465M in funding and is expanding its team to focus on inference-time compute, supercomputing, and AI safety. They aim to revolutionize software development with advanced AI capabilities. Also, Google Research has introduced GameNGen, an AI-powered game engine that generates real-time Doom gameplay using neural networks. The system processes previous frames and player inputs to create new, visually coherent frames at 20 FPS. GameNGen was trained on hundreds of hours of AI-simulated Doom gameplay. To solve issues with frame-to-frame cohesion and visual fidelity degradation in Stable Diffusion, researchers implemented extended sequence training with Gaussian noise corruption and a self-correcting image process. This innovation allows for stable, high-quality image generation over long periods. 



[Quantum Technology] Fujitsu and Osaka University Quantum Advances; Quantum Computing for Neurosciences; Ankaa-2 Available on Amazon Braket. Fujitsu and Osaka University's QIQB developed technologies for an efficient quantum computing architecture, demonstrating that a quantum computer could solve a material energy estimate in 10 hours, compared to five years for a classical computer. This was achieved using only 60,000 qubits, far fewer than previously thought necessary. The breakthrough involves improved phase angle accuracy and automatic generation of efficient qubit operation procedures. This marks a significant step towards quantum advantage, potentially accelerating innovations in superconductors, materials, and drug discovery. The achievement demonstrates how quantum computers could surpass classical computers in the early fault-tolerant quantum computation era, expected around 2030. Meanwhile, researchers used D-Wave's quantum computer to analyze the relationship between prefrontal brain activity and academic performance, focusing on South Korean CSAT scores. The study employed functional near-infrared spectroscopy (fNIRS) to measure brain signals during cognitive tasks and applied quantum annealing algorithms to identify patterns. Key findings showed that verbal fluency, recognition, and Corsi block-tapping tasks were strongly associated with higher CSAT scores. This research opens new avenues for enhancing educational strategies and demonstrates the potential of quantum computing in cognitive neuroscience. Also, Amazon Braket has launched Rigetti's 84-qubit Ankaa-2 superconducting quantum processor. This device offers improved gate operation times, increased two-qubit gate fidelities, and a square lattice qubit arrangement. Unlike other Braket devices, Ankaa-2 is available for customer circuits throughout the day, allowing for uninterrupted experiments globally. It supports Braket Hybrid Jobs with parametric compilation for improved performance. The device expands Amazon's commitment to reducing technology risk and lowering barriers to entry in quantum computing exploration through a consistent interface and pay-as-you-go pricing. Users can access Ankaa-2 via the Amazon Braket console and GitHub repository for examples and algorithms.



[General technology] Boom's XB-1 Edges Closer to Supersonic Speeds; new QR Code Phishing Campaign Discovered; Meta Warns of Threat Actor on WhatsApp. Boom's XB-1 supersonic demonstrator completed its second successful flight in Mojave, California. Key achievements include retracting and extending landing gear, activating a digital stability augmentation system, and evaluating wing aerodynamics. The 15-minute flight reached 10,400 feet and 277 mph. This flight, piloted by Tristan Brandenburg, advances the program towards supersonic flight targeted for year-end. With FAA authorization to exceed Mach 1, Boom plans about 10 more flights before reaching supersonic speeds. XB-1's progress supports the development of Overture, Boom's supersonic airliner, marking a significant step towards reviving mainstream supersonic travel. Meanwhile, a new QR code phishing campaign exploits Microsoft Sway to steal credentials, primarily targeting users in North America and Asia. The attacks use legitimate cloud applications to appear credible, with a 2,000-fold increase in traffic to Microsoft Sway phishing pages since July 2024. The campaign uses QR codes hosted on Sway to redirect users to phishing sites, employing adversary-in-the-middle tactics to capture credentials and 2FA codes. Some attacks use Cloudflare Turnstile to evade detection. Researchers note that QR codes pose challenges for defenders as they bypass email scanners and often lead users to scan with less secure mobile devices. Also, Meta's security teams blocked WhatsApp accounts linked to APT42, a threat actor, posing as tech support agents. The campaign targeted political and diplomatic officials in five countries, including some associated with the Biden and Trump administrations. While no evidence of compromised accounts was found, Meta shared findings with law enforcement and industry peers as a precaution. The attackers used basic phishing tactics to steal credentials. Meta encourages vigilance, especially for public figures and ahead of the US election, and recommends using privacy settings and reporting suspicious activity. 






Comments

bottom of page