This week’s IT Professional highlights the evolving role of AI in software development, emphasizing the shift of costs and challenges from coding to review and governance. As organizations adapt, reliance on agent-native systems raises concerns about trust and security. Effective management and oversight are crucial to leverage AI’s potential while ensuring safety.
This week’s roundup highlights the evolving role of AI in software engineering, emphasizing cost shifts from coding to verification, governance, and reliability challenges. Key developments include Goldman Sachs’ use of agentic AI, GitHub’s new functionalities, and rising costs due to inadequate verification practices, pointing to the need for stronger oversight and documentation.
Tabular Foundation Models (TFMs) are advanced AI designed specifically for handling tabular data, unlike conventional large language models (LLMs) that struggle with such data. TFMs treat numeric values accurately and seek to discern essential patterns within rows and columns, enhancing insight discovery in spreadsheets and structured datasets as they evolve.
Goldman Sachs integrates AI software engineers alongside 12,000 human developers to modernize legacy systems and boost efficiency. Deploying agentic AI like Devin enables independent project execution, drastically reducing bug fix times. However, this automation raises concerns about job redundancy, particularly for junior developers, highlighting a potential skills gap in the industry.
Cloudflare Inc. has launched Cloudflare OS, an open-source AI workspace designed for enterprise employees. It enables users to create custom micro-apps for day-to-day tasks, enhancing collaboration while retaining organizational knowledge. The platform offers complete ownership of developed content, integrates securely with existing systems, and allows use of various AI models without vendor lock-in.
At Kilo Code, engineers now spend only 1% of their time coding, relying heavily on AI agents. This shift raises questions about system safety and cost management. Companies like Replit and Symbotic adapt by implementing human oversight and managing AI expenditures carefully as multi-model support evolves to enhance efficiency and reduce wasted budgets.
Well-written documentation is essential for software development, yet often viewed as a chore. By documenting for your future self, you can capture decisions, manual tasks, and production fixes effectively. This approach reduces mental burden, facilitates task automation, and ensures continuity for future developers, fostering a more efficient workflow.
The U.S. CISA has added three actively exploited vulnerabilities to its Known Exploited Vulnerabilities catalog, including critical issues in N-able N-central. These flaws allow attackers to access servers and managed endpoints of organizations using managed service providers. Urgent patching is required, especially for those relying on third-party IT services, as significant risks are present.
Nvidia Corp. has open-sourced its cuFile API, enhancing data access speed for AI applications. At the Future of Memory and Storage conference, Nvidia unveiled its Storage-Next initiative, collaborating with 40 technology leaders to optimize GPU-driven storage solutions. This approach reduces bottlenecks, facilitating rapid data retrieval and improving AI inference performance.
GitHub has introduced stacked pull requests, a feature that allows developers to divide a large feature into smaller, ordered pull requests for easier review. This technique enhances review quality and speed, addressing issues arising from larger pull requests. As AI tools generate more code, efficient review processes become essential for maintaining productivity and quality.
In October 2025, Hurricane Melissa struck Jamaica, significantly impacting communities. The AI model WeatherNext, developed by Google’s DeepMind, accurately predicted the storm’s trajectory with 80% confidence, offering forecasters an extra day of warning. This breakthrough improves preparedness for extreme weather events, showcasing advancements in hurricane prediction accuracy.
The shift to an “Agent-Native SDLC” signifies a change from human coding to autonomous agents managing delivery. While productivity surges, verification becomes the main bottleneck, costing ~$900,000 annually for a 50-developer team. With AI-generated code presenting increased risks, organizations must evolve CI practices and focus on effective verification and governance measures.
On July 25, Jeff Dean, a prominent Google engineer, addressed aspiring founders at Y Combinator’s Startup School before revealing his new venture, Discovery Loop. This startup aims to automate scientific methods, driven by Dean and fellow AI experts. Their goal is to revolutionize multiple fields through advanced machine learning and experimentation.
Companies are experiencing reliability challenges as AI accelerates code deployment, leading to increased risks of outages. To address this, implementing automated reliability guardrails is crucial. These guardrails validate resilience through testing for common failure modes, provide feedback for fixes, and enhance overall reliability while maintaining speed in development processes.
