AI and Human Collaboration Boosts Self-Driving Lab Performance by 150%
Researchers at Argonne National Laboratory and the University of Chicago have developed a new shared control framework for self-driving laboratories, where human scientists and AI collaborate on the discovery process. This innovative approach has already yielded a 150% increase in mix conducting performance for MIECPs over previous methods. In this model, the AI advisor suggests potential actions, but the human researcher makes the final decision. The team’s future goal is to create a tighter feedback loop where the AI can learn directly from human choices, refining its own decision-making models to better mimic expert intuition.
Sources:
G&J Pepsi Expands Autonomous Vehicle Fleet with Cyngn’s DriveMod Technology
G&J Pepsi, the largest independent Pepsi bottler in the U.S., is significantly expanding its use of autonomous vehicles by placing a new purchase order for Cyngn’s DriveMod Tugger program. This expansion demonstrates growing confidence in Cyngn’s autonomous technology for automating material handling and boosting productivity in large-scale distribution centers. Following a successful initial deployment, the new autonomous vehicles will be rolled out across additional G&J Pepsi facilities serving customers in Ohio and Kentucky.
Sources:
US Air Force to Decommission NIPRGPT GenAI Platform, Migrates to GenAI.mil
The U.S. Department of the Air Force is decommissioning its experimental generative AI chatbot, NIPRGPT, by December 31, well ahead of its planned 2026 sunset. The move comes after the Department of Defense launched its new enterprise AI platform, GenAI.mil. All NIPRGPT users are being directed to migrate their data and workflows to the new DoD-wide platform or other approved systems. Launched in 2024, NIPRGPT provided Air Force personnel with a sandbox to experiment with generative AI on the Pentagon’s non-classified network.
Kubernetes 1.35 ‘Treenetes’ Released with Stable In-Place Pod Resource Scaling
The release of Kubernetes v1.35 ‘Treenetes’ introduces several significant features, headlined by the general availability of in-place pod resource adjustments. This allows administrators to vertically scale CPU and memory allocations for running pods without a restart, a critical enhancement for AI training workloads and edge computing that eliminates downtime. Other key updates in this release include the graduation of mounting OCI images as volumes to stable, enabling better versioning for non-code artifacts. Kubernetes 1.35 also deprecates the IPVS proxy mode and introduces constrained impersonation as an alpha security feature to prevent nodes from being maliciously impersonated.
Sources:
GitHub Enterprise Cloud Launches Data Residency in Japan
GitHub has announced the general availability of data residency in Japan for its GitHub Enterprise Cloud customers. This new capability gives enterprises greater control over their data, allowing them to store their code and repository information within Japan to meet specific compliance and data sovereignty requirements. The service leverages Microsoft Azure’s secure and highly available global data center infrastructure. GitHub has confirmed that it plans to support additional regions for data residency in the future.
Sources:
QubitSolve Awarded $1.2M NSF Grant for Quantum CFD Software Development
QubitSolve Inc. has secured a $1.197 million Small Business Innovation Research (SBIR) Phase II grant from the National Science Foundation (NSF). The funding will accelerate the development of its computational fluid dynamics (CFD) software designed for quantum computers. The grant will help evolve the software from a 2D prototype to a 3D minimum viable product, with a commercial release targeted for late 2027. QubitSolve’s Variational Quantum CFD (VQCFD) algorithm aims to solve complex Navier-Stokes equations for simulations that are intractable for even the most powerful classical supercomputers, with applications in aerospace, defense, and energy.
Sources:
Solana and Aptos Fortify Blockchains Against Future Quantum Computing Threats
The Solana and Aptos blockchain ecosystems are proactively implementing measures to defend against future security threats from quantum computers. The Solana Foundation is collaborating with post-quantum security firm Project Eleven to test post-quantum digital signatures on its testnet, ensuring quantum-resistant transactions can be processed without disrupting the network. Similarly, the Aptos network is evaluating proposal AIP-137 to prepare for quantum advancements. The core concern is that a sufficiently powerful quantum computer could break current cryptographic standards, enabling attackers to derive private keys from public keys and forge transactions.
Sources:
Riverlane Unveils Hardware Decoder for Real-Time Quantum Error Correction
Riverlane has published a paper in Nature Communications detailing its Local Clustering Decoder (LCD), a hardware-based solution for real-time, scalable quantum error correction. This marks a significant milestone, as the LCD is the first decoder to achieve the real-time speed, high accuracy, and adaptive performance in hardware necessary for building fault-tolerant quantum computers. Implemented on an FPGA, the decoder can execute a full decoding cycle in under a microsecond. It is designed for the surface code architecture used by many leading quantum computers and can adapt to changing noise conditions as systems scale.
Sources:
Belden and QSECDEF Partner to Advance Quantum-Secure Networking Standards
Belden Inc. has formed a strategic partnership with Quantum Security & Defence (QSECDEF) to accelerate the development of quantum-secure connectivity. QSECDEF is a global organization dedicated to building a quantum-secure supply chain through procurement standards and certification frameworks. In this collaboration, Belden will lend its expertise in network infrastructure to help establish practical standards for quantum-safe networking, aiming to enhance the resilience of critical systems and prepare enterprises for a secure transition to quantum technologies.
OpenPOWER and ChipFoundry Announce Winners of Microwatt Open-Source Hardware Challenge
The OpenPOWER Foundation and ChipFoundry have announced the winners of the Microwatt Design Challenge, a hackathon designed to foster the development of the open-source Microwatt core using a fully open-source toolchain. The competition drew over 300 team registrations, with winning designs including a System-on-a-Chip (SoC) generator and a minimal hardware debugger. These winning designs will be fabricated into silicon, marking a major step toward creating a complete open-source pipeline for POWER hardware, from digital design to physical chip manufacturing.