Playwright CLI revolutionizes AI-driven browser automation by shifting state storage from an AI model to local disk files, enhancing efficiency. This new command-line interface offers faster execution, reduced token usage, and streamlined workflows. It complements existing Playwright tools, making it ideal for AI-powered test generation and large-scale automation, while suggesting when to use its predecessor, Playwright MCP.
Web automation is evolving with Browserbase and Stagehand, introducing an AI-native framework that eliminates selector reliance and flakiness. This new ecosystem allows QA engineers to run autonomous tests using natural language, improving efficiency and stability while automating complex workflows in dynamic web applications without extensive maintenance.
Artificial Intelligence is transforming software development, necessitating new testing methods for AI models. Traditional testing fails due to non-deterministic outputs, making Evals crucial for assessing model accuracy and performance. Evals provide structured evaluation frameworks to measure AI model consistency and regression, integrating seamlessly into CI/CD pipelines for continuous quality assurance.
On August 28-29, Kongsberg Digital organized an AI Hackathon for all employees, focusing on meaningful AI applications and skill enhancement. Over 200 participants formed 21 teams. The author shares their personal learning journey, exploring AI model development, specifically Retrieval Augmented Generation (RAG), utilizing tools like Ollama and LangChain for practical implementation.
OpenAI’s ChatGPT, launched in November 2022, quickly gained over 100 million users, elevating the company’s valuation to $29 billion and fueling AI interest. In early 2024, Bhavish Aggarwal introduced Krutrim AI, which supports 22 Indian languages. While still in Beta phase, it shows promise but requires refinement to fully compete.