Hey devs, the latest AI news is here! The holiday week didn’t slow down AI innovation. Several announcements delivered new models, developer tooling, and evaluation frameworks aimed squarely at engineers and creators. Here are the standout updates.
Anthropic introduces Bloom evaluation framework. It is an open‑source agentic framework that automates behavioral evaluations for frontier AI models. Bloom generates targeted scenarios to quantify how often specific behaviors occur, operating through four stages—understanding, ideation, rollout and judgment. Early benchmarks show Bloom’s automatic scores correlate strongly with human judgments. The tool aims to speed evaluation cycles from weeks to days and is available on GitHub.
Read more: https://www.anthropic.com/research/bloom
Google announced the general availability of Workspace Studio, a platform where anyone can design, manage, and share AI agents integrated with Gmail, Drive, and Chat. Powered by Gemini 3, Workspace Studio lets users build agents in minutes without coding, automating everything from email triage to multi‑step business workflows.
VS Code, Codex CLI, Copilot CLI add Agent Skills support. Agent Skills are folders containing instructions, scripts, and resources. The standard lets developers create reusable capabilities for tasks like testing or deployment automatically loading only the skills relevant to a given context.
Read more: https://code.visualstudio.com/docs/copilot/customization/agent-skills
If you are looking for the best AI for coding, check out my latest video:
Watch on YouTube: The best AI for coding
Windsurf Wave 13 adds parallel agents and Git worktrees. Cognition’s AI‑native editor Windsurf shipped Wave 13, enabling parallel Cascade agent sessions and a multi‑pane layout. Agents now work on separate Git worktrees, allowing multiple branches to be edited concurrently without conflicts. Wave 13 also introduces a dedicated terminal and context window indicator; SWE‑1.5 is free for three months as the default model.
Read more: https://windsurf.com/blog/windsurf-wave-13
Google introduces Conductor for context‑driven development in Gemini CLI. Conductor helps developers plan before coding and maintain a single source of truth for architecture, style guides, and goals. It creates specs and plans for new features, lets you review them before implementation, and allows you to pause and resume work without losing context.
Read more: https://developers.googleblog.com/conductor-introducing-context-driven-development-for-gemini-cli
That wraps up the final AI‑developer news roundup for 2025! Do you like this newsletter? Please subscribe if you’d like to see more, and feel free to share your own developer news or thoughts in the comments below.
Happy New Year! ;)