Tech & AI

Researchers analyzed 120 specialized programming languages built to help engineers secure software—and discovered they rarely integrate with each other or fit into the full development process. The fragmentation leaves companies patching security gaps with disconnected tools instead of unified systems, slowing deployment and increasing costs.EN

Researchers have resolved a foundational graph theory problem that has stumped mathematicians since the 1980s. The breakthrough could accelerate algorithm design for networks, logistics, and data optimization—areas where understanding graph structure directly impacts computational efficiency and real-world system performance.EN

A new analysis of Google's hypothesis-generating AI system reveals it already exhibits creative traits, challenging the assumption that automation is purely mechanical. The finding suggests that treating AI science tools as creative partners—not just calculators—could unlock better approaches to validating and improving them, with implications for biomedical research and drug development.EN

Researchers found that soaking electron beam 3D-printed titanium parts in dilute hydrofluoric acid for two hours smooths their surfaces significantly, removing a major barrier to using additive manufacturing for aerospace and medical devices. The discovery could reduce post-processing costs and timelines for high-precision metal components.EN

A major European medical position paper links air pollution, noise, and heat to rising blood pressure rates—identifying hypertension as a sentinel marker of environmental damage. For healthcare systems and insurers, this signals growing costs from pollution-related disease and new pressure to integrate environmental factors into clinical care and public health strategy.EN

Researchers have identified a mechanochemical grinding technique that dramatically improves the performance of clay as a concrete additive—but only for certain clay types. The finding could help construction firms reduce cement use and lower carbon emissions, though success depends on selecting the right clay composition for the activation method.EN

A new analysis reveals that while perovskite-silicon hybrid solar cells promise significant efficiency improvements, cost reductions depend heavily on scaling and innovation—not just efficiency gains alone. For manufacturers and energy companies betting on next-generation solar, the real challenge isn't the technology; it's keeping production costs competitive.EN

Researchers have designed a trial framework that uses real-time biomarkers to personalize radiation doses for cancer patients while dramatically reducing the number of participants needed. The approach could accelerate treatment approval timelines and lower clinical trial costs—a critical advantage in an industry where studies often take years and millions of dollars to complete.EN

Researchers have developed a machine learning system that forecasts bus travel times while quantifying prediction confidence—a capability transit agencies and riders currently lack. The model tracks multiple buses simultaneously and accounts for how delays cascade across a route, enabling operators to make better decisions about scheduling and passenger communications.EN

Researchers have used machine learning to automatically design photonic systems that prove quantum mechanics' spooky predictions work in practice. The breakthrough brings device-independent quantum networks—unhackable communication systems immune to detector fraud—closer to commercial viability by making them easier to build and more resistant to real-world signal loss.EN

Researchers found that pulsed charging—rather than constant current—cuts battery degradation by half, potentially extending electric vehicle battery life significantly. The breakthrough could reshape EV economics by reducing replacement costs and improving long-term affordability for mass-market adoption.EN

Scientists have designed a new quantum error-correction system that uses five times fewer physical components than current industry standards while maintaining equal reliability. The breakthrough could accelerate the timeline for building commercially viable quantum computers by dramatically reducing the hardware costs and engineering complexity that currently limit deployment.EN

Scientists have released QDFlow, an open-source simulator that generates synthetic data for training machine learning models on quantum dot devices. The tool could accelerate development of quantum computing hardware by eliminating the need for expensive, time-consuming experimental data collection.EN

Researchers identified why some perovskite materials degrade faster than others when exposed to repeated heating and cooling cycles—a critical problem for tandem solar cells. By swapping one organic component for a bulkier alternative, they achieved dramatically better stability, offering manufacturers a straightforward design rule to improve product durability and reduce warranty claims.EN

A new framework combines frugal innovation with basic technology to deliver maternal health services in low-income regions where resources are scarce. The approach could help countries meet UN targets for reducing maternal mortality without requiring expensive infrastructure—a critical shift for governments and health systems struggling with limited budgets.EN

Researchers discovered that coating cellulose film with standard laser printer toner increases the electrical output of friction-based energy harvesters by 44%. The finding offers a dirt-cheap way to optimize devices that convert mechanical motion—from footsteps to machinery vibration—into usable power, potentially unlocking new markets in self-powered sensors and wearables.EN

Researchers have built an AI system that predicts exactly how printing speed, layer height, and infill density affect the environmental cost of 3D-printed parts. The tool achieves 97% accuracy and works across different materials and designs—giving manufacturers a practical way to cut electricity use and plastic waste before hitting print.EN

A year after launching an open-source platform for modeling how drugs work in the body, developers say the tool is now used by over 100 institutions—including major pharmaceutical companies and regulatory agencies. The shift toward shared, transparent software could accelerate drug approvals and cut development costs by letting firms collaborate on complex calculations that traditionally required expensive proprietary tools.EN

Researchers have developed machine learning models that automatically analyze scrap metal composition before furnace melting, eliminating time-consuming manual selection. The advancement could significantly reduce production costs and energy consumption for European steel producers, while improving competitiveness as the industry faces mounting climate regulations.EN

Researchers found that converting wood residues into hydrochar and adding it to soil cuts greenhouse gas emissions by up to 300 kg annually while improving financial returns. For pulp and paper companies, the discovery offers a dual payoff: a new revenue stream from waste valorization and measurable climate credentials that regulators and investors increasingly demand.EN

Researchers tested a web-based diabetes education portal with actual patients and identified specific design flaws that prevent self-management. The findings reveal a framework for building digital health tools that patients will actually use—critical as healthcare systems worldwide scale remote monitoring to reduce costs and improve outcomes.EN

A new study shows that combining photovoltaic cells with solar-assisted heat pumps dramatically improves performance and reduces climate impact. But the upfront cost is steep: payback periods stretch 24% longer, raising questions about subsidy models needed to make the technology commercially viable.EN

Researchers developed a camera-based method to analyze machined surfaces in real time, catching quality problems before parts leave the shop floor. The approach cuts inspection time and costs by eliminating the need to send samples to separate metrology labs, potentially reshaping how manufacturers control quality.EN

A study of 12 master's programs across Sweden found that while universities cover all major cybersecurity roles, they're unevenly preparing students for emerging threats like AI and cloud attacks. The gap could leave European employers scrambling for talent as cyber threats grow more sophisticated.EN

Engineers have developed a more precise way to control inverters that convert renewable energy onto the power grid, addressing a critical bottleneck as utilities worldwide shift away from coal and gas. The advance could reduce grid instability and allow utilities to integrate far more solar and wind capacity without expensive infrastructure upgrades.EN