AI Roundup: MIT’s Brain-Inspired Chips, AI Weather Forecasts, and Warehouse Intelligence

Today’s top stories in Artificial Intelligence and Machine Learning feature new research from MIT on energy-efficient, brain-inspired computing. In other news, the World Meteorological Organization has endorsed using AI to improve weather forecasts, and Gather AI is strategically shifting its focus to AI-powered warehouse intelligence.

Sources:

WeRide and Uber Debut Autonomous Robotaxi Service in Riyadh

WeRide and Uber have launched an autonomous robotaxi service in Riyadh, Saudi Arabia, introducing self-driving vehicles to the Uber platform in the country for the first time. Initially, the service will operate with a safety operator present in the vehicle. The first public routes are established between Roshn Front and Princess Noura University. This initiative aligns with Saudi Arabia’s Vision 2030 for smart mobility. WeRide had previously secured the first robotaxi autonomous driving permit in Saudi Arabia, and this launch follows a similar service introduction on the Uber platform in Abu Dhabi in late 2024.

Sources:

Grubhub & Avride Launch Robot Delivery Pilot in Jersey City with Wonder

Grubhub is expanding its partnership with autonomous vehicle startup Avride to pilot a robot delivery program in Jersey City. The service will initially cater to customers of the restaurant chain Wonder, marking Grubhub’s first autonomous delivery venture outside of college campuses. The pilot will commence with five of Avride’s advanced delivery robots, which leverage AI-powered autonomous driving technology. Following this initial phase, Grubhub plans to expand its autonomous delivery services to other markets nationwide.

Sources:

Generative AI to Dominate Smartphones, Projecting 43% of Shipments by 2027

Generative AI is poised to revolutionize the smartphone market, with projections indicating that GenAI-enabled devices will constitute 43% of all shipments by 2027. This technological evolution is shifting the industry’s focus from hardware specs like screen size and camera quality to intelligent, context-aware user experiences driven by on-device AI. Experts predict this could disrupt the traditional app-centric model, paving the way for more adaptive interfaces and new device form factors integrated with wearables or ambient computing. As AI becomes more deeply embedded in operating systems like Android, it will enable features unconstrained by physical hardware by offloading processing to the edge. However, this growth introduces significant privacy and data security challenges, as these AI systems rely on vast amounts of personal data.

Sources:

InsurGPT: A New Generative AI Platform Tailored for the Insurance Industry

The company Roots has launched InsurGPT, a generative AI platform specifically trained to handle the complexities of the insurance industry. This addresses the shortcomings of generic models in understanding the sector’s unique nuances. Co-founder Chaz Perera noted that insurers have struggled for decades to automate the vast amounts of unstructured data in underwriting and claims, a problem costing the industry billions. Unlike generic AI, InsurGPT is designed for insurance-specific data and workflows. Its AI Agents can also navigate the complex legacy software common in insurance firms, enabling implementation without a complete overhaul of existing IT infrastructure.

Sources:

Anthropic Deepens Google Cloud Partnership with Massive TPU Deal for AI Scaling

Anthropic has announced a major expansion of its partnership with Google Cloud, securing access to up to one million of Google’s Tensor Processing Units (TPUs). This multi-billion dollar deal will bring over a gigawatt of new computing capacity online by 2026, aimed at scaling Anthropic’s AI research and the development of its Claude models. Anthropic highlighted the superior price-performance and efficiency of the TPUs as key reasons for the decision. This move underscores the escalating demand for specialized AI compute infrastructure and strengthens the collaboration that began in 2023.

Sources:

Google Achieves Verifiable Quantum Advantage with ‘Quantum Echoes’ Algorithm

Google’s Quantum AI division has reported a significant breakthrough using their new ‘Willow’ quantum chip, which features a 105-qubit array. Researchers successfully executed a verifiable quantum advantage algorithm named ‘Quantum Echoes,’ an out-of-order time correlator (OTOC) algorithm. This new technique is reported to be 13,000 times faster than the most advanced classical supercomputers. The experiment involved computing a molecule’s structure by analyzing the ’echo’ of a signal sent through the quantum system. The results, validated against traditional Nuclear Magnetic Resonance (NMR), confirmed their accuracy and demonstrated a major step toward practical applications in medicine and materials science. This marks the first time a quantum computer has run a verifiable algorithm that surpasses supercomputer capabilities.

Sources:

IBM Hits Milestone: Quantum Algorithm Runs on Standard AMD Hardware

IBM has successfully run a key quantum computing algorithm in real-time on standard Advanced Micro Devices (AMD) field-programmable gate arrays (FPGAs). This achievement is a significant step toward creating more practical and cost-effective hybrid quantum systems. The algorithm is designed for dynamic quantum error correction, a critical challenge in building stable quantum computers. By using widely available AMD hardware instead of expensive custom units, IBM is paving a path to more accessible quantum computing. According to Jay Gambetta, IBM’s VP of Quantum, this implementation is ten times faster than required for live error correction, putting IBM’s development a full year ahead of its 2029 roadmap for its Starling quantum system.

Sources:

Percona has released its inaugural State of Open Source Database Management Report, providing key insights into the enterprise data infrastructure landscape. The report highlights the growing adoption of multi-database environments, the need for unified management strategies, the impact of AI and cloud-native solutions, and challenges from changing open-source licensing models. It also emphasizes the role of technologies like Kubernetes in enhancing the scalability of cloud-native databases. A key finding reveals that 78% of technical professionals view PostgreSQL as crucial for their organization’s AI/ML initiatives.

Sources: