New research from our team at @AnthropicAI shows how giving Claude a simple 'think' tool dramatically improves instruction adherence and multi-step problem solving for agents. We've documented our findings in a blog post:
@alexalbert__
-
Industry needs more challenging public evaluations
By
–
This is a call for more evals btw, the whole industry is in desperate need of more challenging public evals!
-
Influencing LLM Development Through High-Quality Evaluations
By
–
Most people don't realize they can significantly influence what frontier LLMs improve at, it just requires some work. Publish a high-quality eval on a task where models currently struggle, and I guarantee future models will show substantial improvement on it.
-
Claude Web Search Rolling Out to Paid Plans in US
By
–
We’re rolling out web search to paid Claude plans in the US in feature preview to start. Toggle it on here: http://
claude.ai/settings/profi
le#feature-preview
… Coming to the free plan and more countries very soon! -
Claude AI Now Features Web Search Capability
By
–
Web search is now available in claude dot ai. Claude can finally search the internet!
-
Manual ESPN bracket entry for Claude’s tournament tracking
By
–
I also manually inputted it into espn so you can follow along and see how claude's bracket does
-
Code and Claude’s Bracket Analysis Results Shared
By
–
Here's the code and claude's bracket analysis/results if you're interested. This was done at midnight last night so there may be bugs.
-
Creating a Game Analysis Harness with Claude Code
By
–
The harness was easy to make with claude code, didn't touch the actual code once Definitely could be optimized much more but at a high level here's how it works rn For each game, we dispatch search subagents across a few queries and then bring all their analysis back to claude
-
Claude’s Conservative Upset Predictions in Full Bracket
By
–
Here's the full bracket, Claude's a bit more on the conservative side when it comes to upsets
-
Claude Code Creates March Madness Bracket Picker in 20 Minutes
By
–
I used Claude Code to make an agentic march madness bracket picker in <20 minutes. Here's the final four that Claude picked: