Google is a leader in The Forrester Wave™: Public Cloud Platforms, Q3 2026
cloud.google.com·5d ago
Cloud & Infrastructure
The New Stack·6h ago·5 min read
This week’s news from Zed, Anthropic, and OpenRouter shows why better harnesses matter more than better models
TL;DR
Zed, Anthropic, and OpenRouter highlight the importance of effective software harnesses around AI models, focusing on usability over model improvements. This week, Zed revamped code review processes, OpenRouter enhanced request management, and Anthropic simplified its architecture.
✦ Why It Matters
Evaluate your current AI model integrations and consider investing in better harness solutions for improved usability.
Key Takeaways
- Explore Zed's new code review features for better collaboration.
- Utilize OpenRouter for enhanced request management in your applications.
- Consider Anthropic's simplified stack for easier integration.
- Monitor token price trends to optimize your AI usage costs.
Full Summary
Recent developments from Zed, Anthropic, and OpenRouter emphasize that the software infrastructure, or harness, surrounding AI models is crucial for user engagement. Zed is transforming how users review code generated by AI agents, making the process more intuitive.
OpenRouter is providing businesses with greater control over the processing of requests, allowing for more efficient operations. Anthropic is streamlining its technology stack to reduce complexity, enhancing user experience.
Additionally, a benchmark revealed ongoing challenges faced by coding agents, while a tutorial offered insights on when to bypass model calls entirely. Notably, Vercel's AI Gateway reported a 23.2% drop in average token prices in August, indicating that inference costs are decreasing, prompting companies to invest in the surrounding software harness.