Project Fetch: Can Claude train a robot dog?
… The team with Claude was able to explore these approaches more efficiently. …
Search
Introducing advanced tool use on the Claude Developer Platform
… Add more servers like Jira which alone uses ~17K tokens and you're quickly approaching 100K+ token overhead. …
Harness design for long-running application development
… Generator: The one-feature-at-a-time approach from the earlier harness worked well for scope management. …
Building Effective AI Agents
… While there are many ways to implement these augmentations, one approach is through our recently released Model Context Protocol , which allows developers to integrate with a growing ecosystem of third-party tools with a simple client implementation . …
Equipping agents for the real world with Agent Skills
… Skills are a simple concept with a correspondingly simple format. …
Labor market impacts of AI: A new measure and early evidence
… Our work follows this task-based approach, incorporating measures of theoretical AI capability and real-world usage, before aggregating to occupations. 3 Measuring exposure Our approach combines data from three sources. …
Long-running Claude for scientific computing
… A good progress file might track current status, completed tasks, failed approaches and why they didn't work, accuracy tables at key checkpoints, and known limitations. The failed approaches are important—without them, successive sessions will re-attempt the same dead ends. …
Eval awareness in Claude Opus 4.6’s BrowseComp performance
… For the first 30 million or so, the model conducted a legitimate search, investigating over a dozen specific candidates across 12 languages on dozens of platforms. …
Trustworthy agents in practice
… This approach is intuitive for simple tasks. …
Teaching Claude why
… When mixing these augmented environments with the simple chat environments, we saw a small but significant improvement in the rate at which the model improved on our honeypot evaluations. …