Improving the speed and energy-efficiency of AI agents
Reports and Proceedings
Updates every hour. Last Updated: 22-Oct-2026 12:17 ET (22-Oct-2026 16:17 GMT/UTC)
“Murakkab” is a new automated system that streamlines the design of agentic workloads for AI applications and optimizes their deployment for customers, reducing computation and cost while boosting energy efficiency.
MIT researchers developed a computational technique for designing buildings, bridges, and other structures that can easily be built on time and budget while also using less material.
By leveraging a classical mathematical function, researchers at the Simons Foundation’s Flatiron Institute developed a new method that enables molecular dynamics simulations to run 2.5 to seven times as fast — all without sacrificing accuracy. For the most popular software package for molecular dynamics simulations, GROMACS, the researchers achieved a fivefold speed increase in simulations run at high accuracy.
A new study published in The Open Construction & Building Technology Journal by Dr. Carlos Alejandro Diaz Schery of the Department of Industrial Engineering, Pontifical Catholic University of Rio de Janeiro, Gávea, Rio de Janeiro, Brazil — together with co-authors Flávia de Souza Costa Neves Cavazotte and Rodrigo Goyannes Gusmão Caiado — makes a pointed observation about why so many Building Information Modelling (BIM) projects fall short of expectations: the problem lies less in the software and more about how organizations are led through the change. Despite over 120,000 published studies on BIM deployment, experts estimate that between 60 and 70 percent of BIM efforts fail to deliver the benefits that were expected. Schery et al.’s study argues that the missing piece is a clear understanding of which leadership behaviors are needed at each stage of an organization’s BIM journey — and that this gap is particularly consequential in Latin America, where cultural dynamics, resource constraints, and uneven institutional support make the challenge more complex than in Europe or North America.
A research paper by scientists from Tsinghua University presented a novel text sequence stimulation paradigm that combines periodic visual stimulation with orthographic information and elicits distinct occipital and occipitotemporal scalp response patterns relative to conventional brightness flicker.
The new research paper, published on Jun. 15 in the journal Cyborg and Bionic Systems, developed a novel text sequence stimulation paradigm that dramatically improves user comfort without sacrificing performance.