MIT Unveils TX-GAIN, The Most Powerful University AI Supercomputer
MIT’s Lincoln Laboratory has launched the TX-Generative AI Next (TX-GAIN) computing system, now the most powerful AI supercomputer at any U.S. university. Optimized for generative AI, TX-GAIN is set to drive innovation in critical research areas such as biodefense, materials discovery, and cybersecurity. With a peak performance of two AI exaflops, the supercomputer will support a wide range of research and development projects at Lincoln Laboratory and across the MIT campus. Researchers are already utilizing TX-GAIN to model complex protein interactions for biological defense, demonstrating its significant impact on scientific breakthroughs. The system is housed in an energy-efficient data center in Holyoke, Massachusetts.
Sources:
TDK and Hokkaido University Pioneer Analog Reservoir AI Chip for Real-Time Learning
TDK Corporation, in collaboration with Hokkaido University, has developed a prototype of a reservoir AI chip that utilizes an analog electronic circuit to emulate the cerebellum. This new chip is designed for reservoir computing, a model that processes time-series data with high speed and low power consumption, making it ideal for edge AI applications. Unlike traditional deep learning models that require massive computations, reservoir computing uses natural phenomena, which can lead to reduced power consumption and latency. TDK plans to exhibit a demonstration device featuring the analog reservoir AI chip at CEATEC 2025 in Japan. This development is aimed at accelerating the commercialization of reservoir computing devices for the edge AI market.
Sources:
Waymo Extends Autonomous Vehicle Testing in NYC Through 2025
New York City regulators have extended Alphabet’s Waymo autonomous vehicle testing permit until the end of 2025. This extension allows Waymo to continue operating up to eight Jaguar I-Pace vehicles with human safety operators in Manhattan and Downtown Brooklyn. The decision to extend the pilot program comes as Waymo demonstrates its technology’s ability to navigate New York’s complex urban environment. Despite the extension, Waymo still needs to secure separate licenses from the city’s Taxi and Limousine Commission to carry passengers or operate a commercial robotaxi service. The regulatory landscape for fully driverless operations in New York State remains undeveloped, with relevant legislation still in committee.
Sources:
Diligent Robotics Bolsters AI Advisory Board with Top Experts
Diligent Robotics, a leader in humanoid robotics for healthcare, has announced the expansion of its AI Advisory Board. The new members are Siddhartha Srinivasa, a Robotics Professor at the University of Washington, and Zhaoyin Jia, a Distinguished Engineer specializing in robotic perception and autonomy. The advisory board will collaborate with Diligent Robotics to advance the capabilities of embodied AI, with a particular focus on ensuring their robots can work seamlessly with people. This initiative aims to push the boundaries of how robots interact, adapt, and learn in real-world environments.
Sources:
Mira Murati’s Thinking Machines Lab Debuts ‘Tinker’ API for LLM Fine-Tuning
Thinking Machines Lab, a startup founded by former OpenAI CTO Mira Murati, has launched its first product, an API named Tinker, designed for fine-tuning large language models. Currently in private beta, Tinker allows AI and security researchers to experiment with algorithms and data using Python without the complexities of distributed training. The API provides access to models such as Alibaba’s Qwen-235B-A22B and Meta’s Llama-3.2-1B. The company, founded in February 2025, includes researchers from prominent AI organizations like OpenAI, Google, Meta, and Mistral. While the API is initially free, it is expected to transition to a usage-based pricing model in the near future.
xenoBrain Integrates Anthropic’s Claude AI for Personalized Economic Forecasting
The economic forecasting platform xenoBrain has partnered with Anthropic to integrate its conversational generative AI, Claude. This collaboration has resulted in a new feature called ‘Personalized AI Commentary,’ which launched today. The tool is designed to help businesses translate predictive insights into actionable strategies by generating explanatory commentary that is customized to a user’s specific operational knowledge. By merging xenoBrain’s predictive analytics with Claude’s natural language processing, the system now offers not only forecasts on corporate performance and market trends but also provides contextual insights and strategic recommendations. This enhanced functionality aims to provide businesses with faster and more context-aware decision-making capabilities.
Sources:
Shortcut Launches ‘Korey’ AI Agent to Orchestrate Software Development
Shortcut has launched an artificial intelligence agent named Korey to orchestrate the planning, tracking, and coordination of software projects. According to Shortcut’s CEO, Kurt Schrader, the Korey agent can take ideas described in natural language and transform them into structured, build-ready plans, complete with detailed specifications and acceptance criteria. The agent is designed to break down projects into actionable tasks, monitor dependencies, and provide summaries of work status. It pulls context from various sources like projects, comments, and GitHub activity to ensure the generated plans are accurate and actionable. As Korey learns a team’s workflows, it aims to function like a product manager, helping software engineering teams to better track their progress. The company plans to extend Korey’s capabilities to other tools and repositories that software engineering teams use.
Sources:
Vectra AI Acquires Netography to Boost Cloud-Native Network Observability
Threat detection and response firm Vectra AI Inc. has acquired Netography Inc., a company specializing in cloud-native network observability. Netography’s platform, Fusion, analyzes network metadata such as flow logs and DNS records without requiring agents or network taps, allowing for rapid deployment in hybrid and multi-cloud environments. With this acquisition, Vectra AI aims to expand its coverage of hybrid enterprises and improve the accuracy of its attack signal detection by incorporating context from cloud control planes and data planes provided by Netography’s technology. The integration of Netography into the Vectra AI Platform is expected to provide organizations with both frictionless, cloud-native observability and AI-driven attack signal clarity.
Sources:
Harvard Achieves Quantum Breakthrough with Continuous Operation Computer
Researchers at Harvard University have developed the first quantum computer capable of continuous operation, a significant milestone in the field. The machine ran for over two hours without restarting, a substantial improvement over previous quantum computers that could only operate for milliseconds or seconds. This breakthrough addresses the critical issue of ‘atom loss,’ where qubits escape the system, leading to computational failure. The new system utilizes an “optical lattice conveyor belt” and “optical tweezers” to replenish lost qubits in real-time, injecting 300,000 atoms per second into the 3,000-qubit system. This development, published in the journal Nature, could pave the way for fault-tolerant quantum computers and more advanced atomic clocks and sensors. Some researchers believe this innovation could lead to quantum machines that can run indefinitely within the next three years.
Sources:
Google Acquires Atlantic Quantum to Accelerate Quantum Hardware Scaling
Google has acquired Atlantic Quantum, a startup founded at MIT that specializes in developing highly integrated quantum computing hardware. The acquisition is aimed at accelerating Google’s efforts to build large-scale, error-corrected quantum computers. Atlantic Quantum’s modular chip stack, which combines qubits and superconducting control electronics within the cold stage, is expected to help Google Quantum AI more effectively scale its superconducting qubit hardware. This move is part of Google’s broader strategy to invest in the future of quantum computing and follows their recent progress with the Willow chip. The integration of Atlantic Quantum’s team and technology is anticipated to speed up Google’s quantum computing roadmap.
Sources:
Silicon Nuclear Entanglement Marks Breakthrough for Scalable Quantum Computers
Researchers have demonstrated quantum entanglement between two atomic nuclei in silicon separated by about 20 nanometers, a conceptual and practical breakthrough for scaling quantum computers. This achievement, published in the journal Science, allows for the integration of long-lived nuclear spin qubits into the existing architecture of standard silicon chips. The method utilizes electrons as a channel to create quantum entanglement between the nuclei, overcoming a significant hurdle in connecting well-shielded quantum bits. This approach could enable the construction of quantum computers that use reliable nuclear spins for computation, leveraging the mature fabrication processes of the semiconductor industry. The ability to entangle nuclei at this scale opens a pathway to building larger and more robust quantum processors.
Sources:
Wikidata Embedding Project Launches as Open-Source AI Knowledge Alternative
A new open-source initiative, the Wikidata Embedding Project, has been launched to provide an alternative to the AI knowledge models developed by large tech companies. The project aims to democratize access to AI knowledge by creating publicly available data embeddings from Wikidata. This initiative is positioned as a community-driven effort to foster more transparent and accessible AI development. The project’s launch is seen as a significant step towards creating a more open and collaborative AI ecosystem. The availability of these open-source embeddings is expected to enable more developers and researchers to build and innovate in the AI space.
Sources: