top of page
Urban Clouds
    • Apr 12

👉 Issue #61 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, and Metaverse. 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 +300 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] Circle and BlackRock Enhance Digital Asset Liquidity, Ledger and Ondo Finance Boost Institutional Trading, DigiFT’s Tokenized Money Market Fund. Circle introduces a smart contract allowing BlackRock's BUIDL Fund investors to transfer shares to Circle for USDC, providing a near-instant, 24/7 off-ramp. This functionality brings the benefits of tokenized assets, such as speed, transparency, and efficiency, to BUIDL investors. The launch of BUIDL and Circle's smart contract functionality marks a milestone in the integration of digital and real-world assets, offering a glimpse into the future of finance powered by blockchain technology. Meanwhile, Ledger Enterprise TRADELINK, a trading network technology with clearing and settlement capabilities integrated with Ondo Finance's OUSG and USDY. This represents a significant advancement for institutional participants, enabling effective operations in the evolving digital assets space. The OUSG provides liquid exposure to short-term U.S. Treasuries and USDY is a tokenized note secured by short-term US Treasuries and bank demand deposits. Also, DigiFT launched the 'DigiFT USD Money Market Fund Token' (DMMF), a yield-generating solution for stablecoins holders. It represents a direct interest in a money market fund managed by a licensed Singapore-based fund manager, investing in high-quality, short-term money market instruments and debt securities. Issued on Ethereum and Arbitrum, DMMF is designed to transform cash management for stablecoins holders seeking treasuries or cash solutions to generate yield. And, BACS (Blockchain Arbitration and Commerce Society) has launched operations in Dubai's DIFC to support the growing blockchain ecosystem. The organization focuses on risk mitigation, dispute resolution, knowledge sharing, and building trust in the blockchain industry, as more startups gravitate towards DIFC. BACS aims to elevate Dubai's profile as a destination for blockchain projects by offering arbitration and dispute resolution services, strengthening the legal framework for emerging technologies. The organization's presence in Dubai aims to facilitate global expansion and attract innovative companies. 


[AI] Mistral Launches Mixtral 8×22B, Intel Unveils Gaudi 3 AI accelerator, new Meta AI Chip Family, and Vals AI testing standardization. French AI startup Mistral released Mixtral 8×22B, a powerful new frontier LLM featuring a 65,000-token context window and 176B parameters. The model uses a sparse mixture of experts (SMoE) approach, optimizing performance and cost by combining specialized models for tasks. Mistral joins a wave of companies releasing or teasing model updates this week, challenging the notion that only big tech giants can develop top-performing models. Meanwhile, Intel unveiled a comprehensive enterprise AI strategy, introducing the Intel Gaudi 3 AI accelerator. This accelerator delivers better performance and power efficiency than Nvidia H100 at a lower cost. Gaudi 3 will be available to OEMs, broadening enterprises' AI data center market offerings. Intel also announced new Gaudi accelerator customers and partners, aiming to create an open platform for enterprise AI and industry leaders to accelerate the deployment of secure generative AI systems. Also, Meta released the next generation of its custom Meta Training and Inference Accelerator (MTIA) AI chip family, delivering a 3x performance improvement over the previous generation. The new chip currently serves Meta's recommendation and ranking models, with plans to expand support to GenAI models like Lllama. This development is crucial for Meta's effort to reduce reliance on costly Nvidia GPUs and optimize their systems. Also, Vals.ai is creating an independent, standardized testing system for evaluating the performance of AI services across various industries. The goal is to address the pressing need for unbiased testing to help businesses understand which models best suit their needs. Vals.ai's initial report revealed that leading models struggled with tax questions, highlighting the importance of domain-specific testing. The startup's approach has garnered interest from the legal community and secured pre-seed funding.


[Quantum Technology] NTT Demonstrates Sub-Millisecond Data Center Links Between U.S. and U.K. with All-Photonics Network; Quantum Breakthroughs with Tokyo-based Quanmatic and First IBM Quantum System One. NTT and NTT DATA demonstrated hyper low-latency connections between data centers in the U.S. and U.K. using All-Photonics Network (APN) technology. The goal is to transform geographically distributed IT infrastructure into a single data center equivalent. Data centers in the U.K. and U.S. were connected using APN equipment, achieving latency of less than 1 millisecond and jitter of less than 0.1 microseconds, significantly outperforming conventional networks. The APN technology refers to a network infrastructure that relies entirely on photonic (light-based) technology for the transmission and routing of data, as opposed to traditional electronic or electro-optical systems.  APN can play a crucial role in the development and deployment of quantum communication networks. Meanwhile, Quanmatic Inc., a Tokyo-based firm, has solved NP-hard optimization problems exceeding 100 million bits using quantum annealing technology, surpassing previous records. The company's success in leveraging quantum computing, classical optimization, and artificial intelligence reflects a significant advance in practical quantum computing applications. Quanmatic's collaboration with ROHM Co., Ltd. to introduce factory floor allocation products for complex LSI production environments further illustrates the real-world impact of its technologies. Also, IBM and Rensselaer Polytechnic Institute (RPI) have unveiled the world's first IBM Quantum System One on a university campus. The 127-qubit system aims to advance quantum computing research, train the next generation of quantum experts, and establish the Capital Region as a hub for quantum innovation. The collaboration will help RPI develop new quantum curricula and educational programs, preparing the future quantum workforce. 


[Metaverse] Siemens Sets Course for Industrial Metaverse with Focus on Interoperability; Samsung and Wilder World's Metaverse Integration; NASA's VR for Gateway. Siemens has outlined a strategy for building the foundation of the industrial metaverse, focusing on interoperability, standardization, data integration, and ecosystem collaboration. The strategy views the industrial metaverse as a solution for real-world problems through digitizing industrial processes, with data remaining integral to ensuring efficient operation and maintenance. This strategy is critical for companies involved in the industrial metaverse, as it will shape the future of digital interaction with industry and other aspects of life. Meanwhile, Samsung Electronics is creating an interoperable smart TV environment powered by Polygon PoS. Samsung's TX1 project aims to bring metaverse experiences to over 200 million users, enabling wider access to an on-chain world. The initiative includes a partnership with Wilder World to integrate its metaverse gaming platform into the Samsung ecosystem, offering NFTs and expanding web3 accessibility. Also, NASA uses virtual reality (VR) to prepare astronauts for living and working on Gateway, the first human space habitat beyond Earth's orbit. VR enables astronauts to familiarize themselves with the station's layout, functionality, and tasks as part of the Artemis campaign, which aims to send humans to Mars. This application highlights the broader use of VR in interplanetary exploration plans. 


[General technology] Google Pledges $1 Billion to Strengthen Pacific Subsea Internet; NASA’s Advanced Solar Sail Technology and Collaborative Robotics Raises $100m. Google has announced a $1 billion initiative to enhance subsea internet infrastructure in the Pacific, building two new internet links and upgrading existing infrastructure. The project aims to improve reliability and reduce latency. Google's cloud data centers feature a custom chip called the Titanium offload processor to optimize hardware operations and speed up network traffic computations. Meanwhile, NASA's Advanced Composite Solar Sail System, set to launch this April, will test a next-generation solar sail technology using lightweight composite booms made from flexible polymer and carbon fiber materials. The 12U CubeSat, built by NanoAvionics, will deploy an 860-square-foot solar sail to demonstrate the new boom technology and perform orbital maneuvers. Successful deployment and operation of the lightweight composite booms could enable larger-scale missions to the Moon, Mars, and beyond, supporting solar sails up to 21,500 square feet. The technology also has potential applications in early warning systems for solar weather and constructing habitats on the Moon and Mars. Also, Collaborative Robotics, founded by former Amazon Robotics chief Brad Porter, has raised $100 million in a funding round led by General Catalyst, valuing the company at over $500 million. The startup, which has 35 employees, develops robots that can move boxes and carts in various settings, potentially replacing conveyor belt systems. Despite a brief slump in 2022, venture capital interest in robotics startups has rebounded, with investors recognizing the potential of AI to enhance robot flexibility and adaptability. Collaborative Robotics' early customers include Mayo Clinic, and the company sees significant demand across industries where people move carts, boxes, and totes.









FEATURED:  Trustless Audits without Revealing Data or Models


By: Suppakit Waiwitlikhit, Ion Stoica, Yi Sun, Tatsunori Hashimoto, Daniel Kang


“There is an increasing conflict between business incentives to hide models and data as trade secrets, and the societal need for algorithmic transparency. For example, a rightsholder wishing to know whether their copyrighted works have been used during training must convince the model provider to allow a third party to audit the model and data. Finding a mutually agreeable third party is difficult, and the associated costs often make this approach impractical.

In this work, we show that it is possible to simultaneously allow model providers to keep their model weights (but not architecture) and data secret while allowing other parties to trustlessly audit model and data properties. We do this by designing a protocol called ZkAudit in which model providers publish cryptographic commitments of datasets and model weights, alongside a zero-knowledge proof (ZKP) certifying that published commitments are derived from training the model. Model providers can then respond to audit requests by privately computing any function F of the dataset (or model) and releasing the output of F alongside another ZKP certifying the correct execution of F. To enable ZkAudit, we develop new methods of computing ZKPs for SGD on modern neural nets for simple recommender systems and image classification models capable of high accuracies on ImageNet. Empirically, we show it is possible to provide trustless audits of DNNs, including copyright, censorship, and counterfactual audits with little to no loss in accuracy.”


Read more: here.




bottom of page