Feb 21, 2025

Feb 21, 2025: Development tools that have recently added new AI capabilities Software companies are constantly trying to add more and more AI features to their platforms, and it can be hard to keep up with it all. We’ve written this roundup to share updates from 10 notable companies that have recently enhanced their products with AI. OpenAI expands Operator to more countries Operator is now available to Pro users across Australia, Brazil, Canada, India, Japan, Singapore, South Korea, and the UK. The company is still working on making it available in the EU, Switzerland, Norway, Liechtenstein, and Iceland, but has not released a timeline for when that will happen. Operator is an AI agent that can perform web-based tasks, such as filling out forms, ordering groceries, or creating memes. “The ability to use the same interfaces and tools that humans interact with on a daily basis broadens the utility of AI, helping people save time on everyday tasks while opening up new engagement opportunities for businesses,” OpenAI wrote in a blog post when Operator was first announced at the end of January. Visual Studio adds support for code referencing of GitHub Copilot completions Microsoft has announced that Visual Studio now supports code referencing for GitHub Copilot completions. Code referencing enables developers to verify if the suggestions coming from Copilot are based on public code, which could potentially lead to open-source licensing issues depending on what the developer is using the code for. When a developer accepts a suggestion that matches code found in a public GitHub repository, they will receive a notification that displays the match, including information about the license type and a link to the GitHub repository it was found in. Microsoft creates generative AI model for game developers The model, Muse , was trained on human gameplay, providing it with a “deep understanding of the environment, including its dynamics and how it evolves over time in response to actions.” Microsoft also believes Muse has potential for applications outside of gaming where understanding of space and environment is important, such as redesigning a retail space or building a digital twin of a factory flow to test different scenarios. Additionally, the company also announced Azure AI Foundry Labs , which is a hub for exploring the latest research findings from Microsoft. Google introduces PaliGemma 2 mix PaliGemma is an open vision-language model announced last May that can handle visual tasks. Version 2 was released in December with pretrained checkpoints in different sizes, and now PaliGemma 2 mix are models that have been tuned to a mixture of tasks and come in a variety of different sizes (3B, 10B, or 24B) and resolutions (224px or 448px). The specific tasks that PaliGemma 2 mix can help with include short and long captioning, OCR, image question answering, and object detection and segmentation. Anaconda launches AI tool for working with data Lumen AI is an open source tool that enables users to “explore, transform, and visualize data using natural language,” according to Anaconda. It can be used to generate SQL queries; transform data across local files, databases, and cloud data lakes; create charts, tables, and dashboards; and serialize and share workflows. “As AI costs decline and open-source models rival proprietary alternatives, now is the time for businesses to embrace truly transformative AI solutions,” said Peter Wang, co-founder and chief AI and innovation officer at Anaconda. “Lumen turns the potential of untapped data into action, helping teams accelerate innovation, optimize workflows, and uncover new opportunities—without needing deep technical expertise.” Sonar has acquired AutoCodeRover AutoCodeRover is an AI agent platform, and its technology will enhance “Sonar’s offering with a sophisticated foundation for automatically addressing real-world engineering issues such as debugging, issue remediation, and code refactoring.” AutoCodeRover features a LLM-agnostic design, making it compatible with a number of different models so that developers can choose the one that best suits their needs. “We are thrilled to join forces with Sonar and contribute to the mission of helping developers around the world accelerate their work,” said Ridwan Shariffdeen, CEO and co-founder of AutoCodeRover. “By combining our advanced agent technology with Sonar’s industry-leading code quality and code security solutions, we’ll drive even greater impact for developers and organizations worldwide.” DeepSeek to open source more of its models The company announced that next week it will begin open sourcing five more of its repositories. “As part of the open-source community, we believe that every line shared becomes collective momentum that accelerates the journey. Daily unlocks are coming soon. No ivory towers – just pure garage-energy and community-driven innovation,” the company wrote in a post on X . A few days earlier the company had also announced NSA , or Natively trainable Sparse Attention mechanism. NSA is designed to enable more efficient long-context training and inference. According to DeepSeek, NSA reduces training costs and maintains performance, meeting or outperforming Full Attention models on general benchmarks, long-context tasks, and instruction-based reasoning. Stytch introduces Connected Apps Connected Apps enables developers to connect AI agents, third-party apps, and multi-app ecosystems to their applications. According to Stytch, it works by making the application an OAuth 2.0 identity provider that can delegate access and permissions. “Whether you’re building AI-driven workflows or an enterprise integration, Connected Apps makes it easy for you to authenticate users, grant scoped permissions, and share data with other applications. With your app as the identity provider, Stytch brings all your connected applications into one place, making it easier to scale, control access, and manage permissions across all your integrations,” Stytch wrote in a blog post. Boomi launches new API management solution to help companies deal with API sprawl The integration company Boomi announced a new API management solution that empowers organizations to control API sprawl. Organizations that are utilizing generative AI have about five times as many APIs as those who aren’t, according to IDC’s API Sprawl and AI Enablement report from December 2024. The new solution provides automated discovery and centralized inventory management, helping to ensure that APIs are properly cataloged, monitored, and controlled. It enables organizations to transform data, applications, and AI assets into APIs, and centrally manage and govern those APIs to eliminate shadow APIs while allowing teams to invest in AI initiatives safely and securely. Securiti announces partnership with Databricks Databricks’ Mosaic AI and Delta tables will be integrated into Securiti’s Gencore AI platform. According to Securiti, the goal of the partnership is to make it easier and safer for developers to build generative AI systems and AI agents using their company’s data. “With organizations racing to adopt AI, the biggest challenge is safely accessing and governing their vast unstructured and structured data assets,” said Rehan Jalil, CEO of Securiti AI. “By combining Gencore AI’s data controls with Databricks Mosaic AI and Delta tables, we deliver the critical data governance, security controls, and contextual intelligence that enterprises need to innovate confidently with AI while meeting regulatory requirements.” Article Tags