Ricerca per:

AI Coding Agents Are Becoming Part of Everyday Development

AI coding agents are quickly becoming a normal part of software development. According to JetBrains’ 2026 Developer Ecosystem Survey, around 90% of professional developers used AI coding agents at work at least once a week between May and July 2026. The survey included more than 15,000 professional developers worldwide.

Unlike traditional code assistants, AI agents can handle much larger parts of a task. They can explore a codebase, write and modify files, run tests, and suggest solutions to problems. Developers can then review the results and decide what should be kept or changed.

One interesting change is the growing popularity of Claude Code, which has reportedly overtaken GitHub Copilot among professional developers. This does not necessarily mean that traditional coding assistants are disappearing. Instead, it suggests that developers are becoming more comfortable giving AI tools more responsibility.

However, AI coding agents are not replacing developers. Their output still needs to be checked, especially when a project is complex or security is important. A developer must understand the code and be able to recognize when an AI-generated solution is wrong.

The future may bring an interesting challenge: a significant increase in productivity could come together with a decline in software quality. If developers can produce code much faster, but that code is harder to maintain, less reliable, or less carefully designed, the industry will need to find a new balance. How to benefit from the speed of AI without sacrificing quality could become one of the most important challenges for software development in the years ahead.

Artificial Intelligence and Health: From Diagnosis to Drug Discovery

https://www.sifweb.org/sif-magazine/articolo/intelligenza-artificiale-e-salute-dalla-diagnosi-alla-scoperta-di-nuovi-farmaci-2025-03-27

Artificial Intelligence (AI) is a technology that allows computers to process huge amounts of data, recognize patterns, and make highly accurate predictions. Through advanced models such as machine learning and deep learning, AI systems can learn from data and continuously improve their performance. In medicine, this innovation is transforming the way diseases are diagnosed, monitored, and treated. AI is currently used to analyze medical images, clinical records, and scientific data, helping doctors detect abnormalities more accurately and predict the risk of certain diseases. This supports earlier interventions and more personalized treatments. Robotic surgery and remote patient monitoring are also benefiting from AI, making some procedures less invasive and improving long-term therapy management. In the pharmaceutical field, AI is speeding up drug discovery by rapidly analyzing large databases of chemical compounds and biological information. This makes it possible to identify promising molecules much faster than traditional methods and even discover new uses for existing drugs, a process known as “drug repurposing.” Some AI-assisted drugs are already undergoing clinical trials, while researchers are also using AI to identify new antibiotics against resistant bacteria.                                                                                          

An additional technical aspect, only briefly mentioned in the original text, is the use of artificial neural networks in medical AI. These systems are trained on extremely large datasets — for example thousands of X-rays, MRI scans, or patient records — and learn to recognize specific patterns associated with diseases. In medical imaging, this allows AI to detect small anomalies, such as tumors or early signs of neurological disorders, sometimes with a level of accuracy comparable to specialist physicians. AI is also playing a key role in the development of predictive and personalized medicine. By combining genetic information, clinical history, lifestyle factors, and laboratory results, AI systems can help predict how likely a patient is to develop certain diseases or respond to a specific treatment. This could lead to therapies that are increasingly tailored to the individual patient, improving effectiveness while reducing side effects and unnecessary treatments.

Despite its enormous potential, AI cannot replace doctors. Medical decision-making still depends on human experience, clinical interpretation, and empathy toward patients. Important challenges also remain, including data privacy, algorithm reliability, and legal responsibility in the event of diagnostic errors. For this reason, AI should be considered a powerful support tool for healthcare professionals rather than a substitute for human care.

AI as cognitive support for developers

A recent study by GitHub Next explores how artificial intelligence is reshaping modern software development workflows. The interviews revealed a clear distinction between repetitive, standardized tasks — such as boilerplate code generation, simple refactoring, or test generation — and high cognitive-load activities, including complex debugging, architectural design, contextual system analysis, and the evaluation of technical trade-offs. “We want AI to eliminate repetitive tasks by suggesting improvements, writing documentation or tests, and identifying issues… not to interrupt your creative flow or your autonomy.” Developers particularly value AI when it acts as a reasoning support layer: summarizing context, analyzing large codebases, suggesting implementation strategies, and generating actionable plans. At the same time, there remains a strong demand for human oversight in critical decisions and structural code changes, both to ensure reliability and to preserve understanding and control of the system. According to the study, the paradigm is evolving from a simple “AI assistant” toward a true “cognitive partner,” with tools increasingly designed to support problem solving, systems thinking, and complexity management. One particularly interesting finding is that most developers are not seeking full automation, but rather transparent, verifiable tools that can be integrated into code review, testing, and validation workflows. In the medium term, this shift could move the developer role toward skills more focused on software architecture, AI agent orchestration, and supervision of generative workflows.

What AI is actually good for, according to developers