ISSUE 511
OpenCode, Amazon's vibe-coding woes, and Nvidia's Nemoclaw

"A core principle of the VS Code team has always been that we use VS Code to build VS Code. The same is true now with AI."

— Pierce Boggan, the Product Lead for VS Code and GitHub Copilot at Microsoft, in an exclusive interview with The New Stack. Read more.

Inception says its diffusion LLM is 10x faster than Claude, ChatGPT, Gemini

Move over, "fancy autocomplete." In this episode of The New Stack Agents, Inception CEO Stefano Ermon joins The New Stack to explain how Mercury 2 uses diffusion to deliver LLM speeds 10x faster.

Week in review: OpenCode, Amazon's vibe-coding woes, and Nvidia's Nemoclaw

Here's what people flocked to last week on The New Stack: Open source coding agents stole the show, with our look at OpenCode's new subscription plan (and on how agents like it are solving a huge headache for developers) generating massive reader interest.

Also: We explored the Markdown-vs-MCP debate that is remolding AI agent configuration, Anthropic's new multi-agent code review architecture for Claude Code, and OpenAI's expansion of Codex to Windows. Finally, there was an exploration of AerynOS, a new Linux distribution built from the ground up.

In the past few days, reporting on vibe-coding problems at Amazon and Nvidia’s reported plans for NemoClaw generated interest, as did our interview with the Product Lead for VS Code and GitHub Copilot at Microsoft, who has revealed that his team moved to weekly releases after 10 years of monthly — and he credits AI for making it possible.

Thanks for reading. See you next Friday.

— Nick Lucchesi, editor-in-chief, TNS

JetBrains names the debt AI agents leave behind
JetBrains launches Junie CLI and Air to bring context-aware, vendor-neutral infrastructure to the agentic coding era.

How to deploy an AI server on your Debian/Ubuntu server
Running AI locally keeps your data private and your queries off the grid — and it's easier to set up than you think.

With GridGain acquisition, MariaDB bets on in-memory computing and Apache Ignite

MariaDB agrees to acquire GridGain to combine relational database with in-memory computing, targeting the AI latency gap in agentic applications.

Cursor builds always-on agents to tackle developer task tedium

Future workflows will see boosted productivity with Cursor Automations: AI agents that handle code reviews, bug fixes & triage inside a developer's codebase.

Scaling real-time AI & ML workloads for performance and efficiency
March 18

In this live session, we will dive into why real-time context is now the core data primitive for intelligent systems. Join us to learn where traditional infrastructure falls short, what modern AI/ML workloads demand, and how purpose-built systems enable predictable, efficient scale.

AI-powered K8s observability best practices in 2026
March 19

As AI workloads move into production, the ephemeral nature of Kubernetes is becoming even more difficult to manage, making end-to-end visibility a requirement rather than a luxury. Join us to explore how to leverage AI-powered insights to ensure the health, performance, and security of your increasingly complex K8s environments.

Building production-ready agentic AI
March 24

In this webinar, we will explain why real-time context is now the core data primitive for intelligent systems. Learn where traditional infrastructure falls short, what modern AI/ML workloads demand, and how purpose-built systems enable predictable, efficient scale.

Building a self-service developer platform the thoughtful way
On-Demand

Developers want a public cloud-like experience — even on-prem. In this session hosted by Charles Humble, Broadcom’s Jad El-Zein uncover how to automate full-stack Kubernetes deployments with a single Git commit.

Write for TDS: Reach a global audience and get paid

We’re proud to highlight our sister site, Towards Data Science, which recently updated its author program to better reward the community. If you’re working on unique AI projects, deep-dive tutorials, or industry analysis, TDS offers a platform to reach a massive audience and get paid for your work. Take advantage of updated earning tiers and a higher monthly article cap today!

Connect with The New Stack