语言 & 框架

OpenTelemetry 正迅速成为可观察性领域的行业标准。随着 OpenTelemetry 协议 (OTLP) 规范的发布，行业内有了一个标准化的方式来处理追踪 (traces)、指标 (metrics) 和日志 (logs)。这减少了在监控分布式解决方案和满足互操作性需求时的多重集成或主要代码重构的需要。随着 OpenTelemetry 扩展对日志和性能分析的支持，OTLP 为所有遥测数据提供了一个一致的传输格式，简化了应用的仪表化过程，使全栈可观察性对于微服务架构更加易于实现且具有可扩展性。OTLP 已被诸如 Datadog 、New Relic 和 Grafana 等供应商采纳，帮助企业构建灵活的、与供应商无关的可观察性技术栈，避免被锁定在专有解决方案中。OTLP 支持 gzip 和 zstd 压缩，大幅减少遥测数据的大小并降低带宽使用——这对于处理高遥测数据量的环境尤为关键。为了支持长期发展，OTLP 确保 OpenTelemetry 保持强大且面向未来的标准，巩固了其作为遥测传输事实标准的地位。

我们将 React Hook Form 视为 Formik 的替代方案。由于默认使用非受控组件，它提供了显著更好的开箱即用的表现，特别是在大型表单上。React Hook Form 很好地和各种基于 schema 的验证库（比如Yup, Zod 等）进行了集成。此外 React Hook Form 提供了很大的灵活性，使其易于和现有代码库和其他库集成。你可以把 React Hook Form 和像shadcn或者AntD这样的外部受控组件库一起使用。凭借出色的性能、无缝集成和活跃的开发社区，它是构建大型表单或表单密集型应用的可靠选择。

Effect 是一个强大的 TypeScript 库，用于构建复杂的同步和异步程序。在 Web 应用开发中，常常需要为异步处理、并发、状态管理和错误处理等任务编写大量样板代码。而 Effect-TS 通过采用函数式编程的方法简化了这些流程。借助 TypeScript 的类型系统，Effect 能够在编译时捕获难以检测的问题。我们的团队曾使用过 fp-ts 进行函数式编程，但发现 Effect-TS 提供的抽象更贴近日常任务的需求，同时使代码更易于组合和测试。尽管传统方法如 Promise/try-catch 或 async/await 也能处理类似场景，但在使用 Effect 之后，我们的团队发现没有理由回到传统方法。

Hasura GraphQL engine 是一个通用的数据访问层，可简化在不同数据源上构建、运行和管理高质量 API 的过程。它能够为各种数据库（包括 PostgreSQL、MongoDB 和 ClickHouse）及数据源即时生成 GraphQL API，使开发者能够快速且安全地获取所需数据。我们发现，Hasura 在实现 服务端资源聚合 的 GraphQL 应用场景中非常易用，并已将其应用于多个 数据产品项目中。然而，对于其强大的联合查询和统一模式管理功能，我们仍然保持 谨慎态度。值得一提的是，Hasura 最近推出了 PromptQL 功能，允许开发者利用大语言模型（LLM）实现更自然直观的数据交互。

LangGraph 是一款面向基于 LLM 的多 agent 应用构建的编排（Orchestration）框架。与抽象程度较高的 LangChain 相比，它提供了更底层的节点（Nodes）和边（Edges）等基本原语，允许开发者精细地控制 agent 工作流、记忆管理与状态持久化。这种基于图的设计使工作流更加可控且易于定制，使得在生产级应用中的调试、扩展和维护变得更加容易。尽管其学习曲线较陡，但 LangGraph 凭借其轻量化与模块化设计，在开发 agent 应用时展现出了强大的灵活性和扩展性。

MarkItDown 能将多种格式（PDF、HTML、PowerPoint、Word）转换为 Markdown，从而增强文本的可读性和上下文保留。由于 LLM 可以从格式化提示（如标题和章节）中获取上下文，Markdown 能够很好地保留结构以提升理解能力。在基于 RAG 的应用中，我们的团队使用 MarkItDown 将文档预处理为 Markdown 格式，确保逻辑标记（如标题、子章节）保持完整。在生成嵌入之前，结构感知的分块方法帮助保留了完整的章节上下文，从而提升了查询响应的清晰度，尤其是对于复杂文档而言。Markdown 被广泛用于文档编写，而 MarkItDown 的 CLI 工具也因此成为一个极具价值的开发者生产力工具。

Module Federation 允许在微前端之间指定共享模块并实现依赖的去重。随着 2.0 版本的发布，模块联邦已经发展到可以独立于 webpack 工作。此更新引入了一些关键功能，包括联邦运行时（Federation Runtime）、全新的插件 API，以及对流行框架（如 React 和 Angular）及热门构建工具（如 Rspack 和 Vite）的支持。通过采用模块联邦，大型 Web 应用可以拆分为更小、更易管理的微前端，使不同的团队能够独立开发、部署和扩展，同时高效地共享依赖项和组件。

Prisma ORM 是一个开源的数据库工具包，帮助 Node.js 和 Typescript 应用简化应用数据库操作。它提供了一种现代化的、类型安全的数据库访问方式，能够自动化数据库模式迁移，并提供直观的查询 API。与传统 ORM 不同，Prisma ORM 使用纯 JavaScript 对象定义数据库类型，而无需依赖装饰器或类。我们对 Prisma ORM 的使用体验非常积极，它不仅更符合 TypeScript 的开发生态，还能很好地融入函数式编程范式中。

.NET Aspire is designed to simplify the orchestration of distributed applications on a developer's local machine. Aspire lets you orchestrate multiple services in a local development environment — including multiple .NET projects, dependent databases and Docker containers — all with a single command. Furthermore, Aspire provides observability tools — including logging, tracing and metrics dashboards — for local development, decoupled from the tools used in staging or production environments. This significantly improves the developer experience when building, tweaking and debugging the observability aspects of any system they are working on.

Google, in collaboration with Samsung and Qualcomm, has introduced Android XR, a new operating system designed for XR headsets. Support is planned for glasses and other devices. Most Android apps are supported with no or minimal changes, but the idea is to build new spatial apps from scratch or to "spatialize" existing apps. The new Android XR SDK is positioned as the go-to SDK for such projects, and Google provides guidance on how to choose tools and technologies bundled in the SDK. It’s available in developer preview now.

Browser Use is an open-source python library that enables LLM-based AI agents to use web browsers and access web applications. It can control the browser and perform steps that include navigations, inputs and text extractions. With the ability to manage multiple tabs, it can perform coordinated actions across multiple web apps. It’s useful for scenarios where LLM-based agents need to access web content, perform actions on it and get the results. The library can work with a variety of LLMs. It leverages Playwright to control the browser, combining visual understanding with HTML structure extraction for improved web interaction. This library is gaining traction in multi-agent scenarios, enabling agents to collaborate on complex workflows involving web interactions.

CrewAI is a platform designed to help you build and manage AI agents that can work together to accomplish complex tasks. Think of it as a way to create a crew of AI workers, each with their own special skills, who can collaborate to achieve a common goal. We’ve mentioned it previously in the Radar under LLM-powered autonomous agents. In addition to the open-source Python library, CrewAI now has an enterprise solution so organizations can create agent-based applications for real-world business cases, run them on their cloud infrastructure and connect to existing data sources such as Sharepoint or JIRA. We've used CrewAI multiple times to tackle production challenges, from automated validation of promotion codes to investigating transaction failures and customer support queries. While the agentic landscape continues to evolve rapidly, we’re confident in placing CrewAI in Assess.

ElysiaJS is an end-to-end type-safe web framework for TypeScript, designed primarily for Bun but also compatible with other JavaScript run times. Unlike alternatives such as tRPC, which enforces specific API interface structures, ElysiaJS does not impose any API interface structure. This allows developers to create APIs that follow established industry practices such as RESTful, JSON:API or OpenAPI and also provide end-to-end type safety. ElysiaJS is highly performant when used with Bun run time, even comparable to Java or Go web frameworks in some benchmarks. ElysiaJS is worth considering, especially when building a backend-for-frontend (BFF).

FastGraphRAG is an open-source implementation of GraphRAG designed for high retrieval accuracy and performance. It uses Personalized PageRank to limit graph navigation to the most relevant nodes among all the related nodes in the graph, enhancing retrieval accuracy and improving LLM response quality. It also provides a visual representation of the graph, helping users understand node relationships and the retrieval process. With support for incremental updates, it’s well-suited to dynamic and evolving data sets. Optimized for large-scale GraphRAG use cases, FastGraphRAG improves performance while minimizing resource consumption.

Erlang/OTP is a powerful platform for building highly concurrent, scalable and fault-tolerant distributed systems. Traditionally, its languages have been dynamically typed, but Gleam introduces type safety at the language level. Built on BEAM, Gleam combines the expressiveness of functional programming with compile-time type safety, reducing run-time errors and improving maintainability. With a modern syntax, it integrates well with the OTP ecosystem, leveraging the strengths of Erlang and Elixir while ensuring strong interoperability. The Gleam community is active and welcoming, and we look forward to its continued development.

GoFr is a framework for building microservices in Golang, designed to simplify development by abstracting away boilerplate code for common microservice functionalities such as logging, traces, metrics, configuration management and Swagger API documentation. It supports multiple databases, handles database migrations and facilitates pub/sub with brokers like Kafka and NATs. Additionally, GoFr includes task scheduling with cron jobs. It reduces the complexity of building and maintaining microservices, and allows developers to focus on writing business logic rather than infrastructure concerns. Even though there are other popular Go libraries for building web APIs, GoFr is gaining traction and is worth exploring for Golang-based microservices.

At the core of asymmetric cryptography, which secures most modern communication, lies a mathematically hard problem. However, the problem used in today's algorithms will be easy to solve with quantum computers, driving research for alternatives. Lattice-based cryptography is currently the most promising candidate. Although cryptographically relevant quantum computers are still years away, post-quantum cryptography is worth considering for applications that must remain secure for decades. There is also the risk that encrypted data is recorded today in order to be decrypted once quantum computers become available.

Java post-quantum cryptography takes its first steps in JDK 24, set for general availability in late March. This release includes JEP 496 and JEP 497 — which implement a key encapsulation mechanism and a digital signature algorithm — both standards-based and designed to be resistant to future quantum computing attacks. While liboqs from the Open Quantum Safe project provides C-based implementations with a JNI wrapper, it’s encouraging to see a native Java implementation emerging as well.

Presidio is a data protection SDK for identifying and anonymizing sensitive data in structured and unstructured text. It detects personally identifiable information (PII) such as credit card numbers, names and locations, using named entity recognition, regular expressions and rule-based logic. Presidio supports custom PII entity recognition and de-identification, allowing businesses to tailor it to their specific privacy requirements. Although Presidio automates the identification of sensitive information, it’s not foolproof and may miss or misidentify data. Exercise caution when relying on its results.

As the technologies to build LLM-backed applications and agents continue to evolve rapidly, frameworks for building and orchestrating such applications often struggle to keep up or find the right timeless abstractions. PydanticAI is the latest entrant in this space, aiming to simplify implementations while avoiding unnecessary complexity. Developed by the creators of the popular Pydantic, it builds on lessons learned from earlier frameworks — many of which already rely on Pydantic. Rather than trying to be a Swiss Army knife, PydanticAI offers a lightweight yet powerful approach. It integrates with all major model APIs, includes built-in structured output handling and introduces a graph-based abstraction for managing complex agentic workflows.

Since the release of Swift 6.0, the language has expanded beyond Apple's ecosystem with improved support for major operating systems, making it more viable to use Swift for resource-constrained applications. Traditionally, this space has been dominated by C, C++ and, more recently, Rust, due to their low-level control, high performance and availability of certified compilers and libraries that comply with standards such as MISRA, ISO 26262 and ASIL. While Rust has begun achieving similar certifications, Swift has yet to pursue this process, limiting its use in safety-critical applications.

Swift's growing adoption is driven by its balance of performance and safety features, including strong type safety and automatic reference counting for memory management. While Rust’s ownership model offers stronger memory safety guarantees, Swift provides a different trade-off that some developers find more approachable. Both Swift and Rust share the LLVM/Clang compiler backend, allowing advancements in one to benefit the other. With its ability to compile to optimized machine code, its open-source development and its expanding cross-platform support, Swift is emerging as a contender for a wider range of applications — far beyond its iOS roots.

Tamagui is a library for efficiently sharing styles between React web and React Native. It offers a design system with reusable styled and unstyled components that render seamlessly across platforms. Its optional optimizing compiler boosts performance by converting styled components into atomic CSS with divs on the web and hoisted style objects on native views.

torchtune is a PyTorch library for authoring, post-training and experimenting with LLMs. It supports single and multi-GPU setups and enables distributed training with FSDP2. The library provides YAML-based recipes for tasks like fine-tuning, inference, evaluation and quantization-aware training. Each recipe offers a focused set of features, avoiding complex flag-based configurations. It prioritizes simplicity, favoring code clarity over excessive abstractions. It also includes a CLI for downloading models, managing recipes and running experiments efficiently.

A few years ago, we observed Node overload: Node.js was often used for questionable reasons or without even considering any alternatives. While we understand that some teams prefer a single-language stack — despite the trade-offs — we continue to advocate for polyglot programming. At the time, we noted that Node.js had a deserved reputation for efficiency in IO-heavy workloads, but we mentioned that other frameworks had caught up which offered better APIs and superior overall performance. We also cautioned that Node.js was never well-suited to compute-heavy workloads, a limitation that remains a significant challenge. Now, with the rise of data-heavy workloads, we’re seeing teams struggle with these as well.