Ali Azhar

Scientists often wait months to get their hands on some of the world’s most powerful research instruments, such as state-of-the-art microscopes. At facilities like Argonne National Laboratory’s Advanced Photon Source, these machines allow researchers to use extremely bright X-rays to study materials at microscopic scales. Once they get access to such instruments, every hour counts. […]

A new study from Google and MIT finds that AI is accelerating the scientific process, with a significant increase in productivity among scientists who use AI. The paper found that AI usage is widespread in science, and that the average scientist saves nearly 30 hours per month using AI. However, there are some caveats to the AI trend, which could negatively impact the future of scientific discovery.

For all the talk about a shortage of AI computing power, and the need for more powerful chips and bigger data centers, there’s another side to the story. A significant amount of existing capacity isn’t being used. While tech companies race to build massive new data centers, expensive GPUs in other facilities can sit idle […]

Google just released the new Gemini 4 Argon, which it claims can generate up to a million tokens in a single run, a big jump for long code migrations and multi-step agent work. Competitors top out around 128,000 output tokens per request, so based on that claim, Argon's 1 million is roughly 8x higher.

Enterprise data platforms are expanding, but the files companies have accumulated over the years are not keeping pace. A new Komprise survey found that only 4% of organizations have created schema for more than 70% of their unstructured file data and made it accessible to lakehouse and analytics teams. Just 24% have even made it past the halfway point so far.

A more powerful AI model isn't necessarily the answer to unreliable business analytics. New research from AtScale found that leading AI models answered fewer than 40% of business questions correctly when querying databases directly. But when AtScale added its semantic layer, accuracy improved by 1.7x to 2.4x across all seven models tested.

CIOs need to know how to calculate the costs associated with computing. Historically, CIOs were able to determine an approximate cost of providing employees with the tools necessary to complete their job functions simply by looking at the price of a laptop or workstation. Because a typical office computer uses relatively little electricity compared to the initial investment required to acquire it, the cost of electricity rarely influenced purchasing decisions. AI workloads change this dynamic. As employees (i.e., engineers, developers, designers, etc.) increasingly require GPU-enabled workstations and other high-performance equipment that consume far more electricity than typical enterprise computers, the cost of electricity has become a significant factor in determining total ownership costs.

HP has replaced its customer data platform with Databricks CustomerLake. A customer data platform, or CDP, is the tool marketers use to pull customer information into one place. HP says it now with CustomerLake it builds audiences in 2 hours instead of 2-3 months, and that its CDP costs dropped by up to 40%.

CERN, the world's largest particle physics laboratory, has spent decades collecting enormous amounts of data from particle collisions. With the recent rapid growth in AI capabilities, CERN wants to develop a foundation model that can make better use of all that data.

The library is designed to target a specific problem: solving large sparse systems of equations. These calculations are used extensively in engineering simulations.