Thoughtworks recently published their Technology Radar Volume 31, providing an opinionated guide to the current technology landscape.
As per the Technology Radar, Generative AI and Large Language Models (LLMs) dominate, with a focus on their responsible use in software development. AI-powered coding tools are evolving, necessitating a balance between AI assistance and human expertise.
Rust is gaining prominence in systems programming, with many new tools being written in it. WebAssembly (WASM) 1.0’s support by major browsers is opening new possibilities for cross-platform development. The report also notes rapid growth in the ecosystem of tools supporting language models, including guardrails, evaluation frameworks, and vector databases.
In the Techniques quadrant, notable items in the Adopt ring include 1% canary releases, component testing, continuous deployment, and retrieval-augmented generation (RAG). The Radar stresses the need to balance AI innovation with proven engineering practices, maintaining crucial software development techniques like unit testing and architectural fitness functions.
For Platforms, the Radar highlights tools like Databricks Unity Catalog, FastChat, and GCP Vertex AI Agent Builder in the Trial ring. It also assesses emerging platforms such as Azure AI Search, large vision model platforms such as V7, Nvidia Deepstream SDK and Roboflow, along with SpinKube. This quadrant highlights the rapid growth in tools supporting language models, including those for guardrails, evaluations, agent building, and vector databases, indicating a significant shift towards AI-centric platform development.
The Tools section underscores the importance of having a robust toolkit that combines AI capabilities with reliable software development utilities. The Radar recommends adopting Bruno, K9s, and visual regression testing tools like BackstopJS. It suggests trialing AWS Control Tower, ClickHouse, and pgvector, among others, reflecting a focus on cloud management, data processing, and AI-related database technologies.
For Languages and Frameworks, dbt and Testcontainers are recommended for adoption. The Trial ring includes CAP, CARLA, and LlamaIndex, reflecting the growing interest in AI and machine learning frameworks.
The Technology Radar also highlighted the growing interest in small language models (SLMs) as an alternative to large language models (LLMs) for certain applications, noting their potential for better performance in specific contexts and their ability to run on edge devices. This edition drew a parallel between the current rapid growth of AI technologies and the explosive expansion of the JavaScript ecosystem around 2015.
Overall, the Technology Radar Vol 31 reflects a technology landscape heavily influenced by AI and machine learning advancements, while also emphasizing the continued importance of solid software engineering practices. Created by Thoughtworks’ Technology Advisory Board, the technology Radar provides valuable insights twice-yearly for developers, architects, and technology leaders navigating the rapidly evolving tech ecosystem, offering guidance on which technologies to adopt, trial, assess, or approach with caution.
The Thoughtworks Technology Radar is available in two formats for readers: an interactive online version accessible through the website, and a downloadable PDF document.
