InfoQ

3 days ago | infoq.com | Anthony Alford

The Agentica Project and Together AI have released DeepCoder-14B-Preview, an open source AI coding model based on Deepseek-R1-Distilled-Qwen-14B. The model achieves a 60.6% pass rate on LiveCodeBench, outperforming OpenAI's o1 model and matching the performance of o3-mini. DeepCoder-14B-Preview is fine-tuned from the Deepseek model on a dataset of 24K coding problems using reinforcement learning (RL).

3 days ago | infoq.com | Renato Losio

InfoQ ホームページ ニュース Redis 8はベクトル類似性の新しいデータ型でAIアプリケーションをターゲットにする 原文リンク(2025-04-20) Redisは最近、ベクトル類似性のために設計されたデータ型であり、AIアプリケーションのための新しいオプションであるVector Setの追加を発表した。この新しいデータ型は、Redisの生みの親であるSalvatore Sanfilippo氏（別名「antirez」）がRedisに復帰して以来、初めての大きな貢献となる。 Vector SetsはSorted Setsに似た新しいデータ型だが、スコアの代わりに文字列要素がベクトルに関連付けられており、アイテムを追加し、追加されたアイテムの中から指定されたベクトルにもっとも似ているサブセットを取り出すことができる。フィルター検索機能もサポートされており、ベクトルの類似性とスカラーのフィルターを同時に行うことができる。Sanfilippo氏はブログでこう説明している。...

3 days ago | infoq.com | Daniel Curtis

Biome, the all-in-one JavaScript toolchain, has released v2.0 Beta. Biome 2.0 Beta introduces a number of new features in this beta which bring it closer to ESLint and Prettier, such as plugins, to write custom lint rules, domains to group your lint rules by technology and improved sorting capabilities. Biome has support for JavaScript, TypeScript, JSX, TSX JSON and GraphQL and claims to be 35x faster when compared to Prettier.

4 days ago | infoq.com | Sergio De Simone

Over the past few years, Pinterest engineers have been working to adopt Docs-as-Code across the organization. By using the same tools and processes for both code and documentation, they’ve improved overall documentation quality and team satisfaction. The shift also helped transform Pinterest’s documentation culture, fostering collaboration, improving quality control, discoverability, and more.

5 days ago | infoq.com | Renato Losio

The recent article by Zhou Sun, "HTAP is Dead," sparked a debate in the data community about the future of hybrid transaction/analytical processing. HTAP was meant to help integrate historical and online data at scale, supporting more flexible query methods and reducing business complexity. In the article, Sun, co-founder and CEO at Mooncake Labs, argues that the long-promised vision of unifying transactional and analytical workloads in a single system has failed to materialize.