HumanLayer
PulseAugur coverage of HumanLayer — every cluster mentioning HumanLayer across labs, papers, and developer communities, ranked by signal.
7 day(s) with sentiment data
-
Key AI Personnel Shifts: Anthropic, OpenAI, AMD, Meta Announce Moves
Cat Wu, formerly Head of Product at Anthropic, is moving to OpenAI to lead Core Products PM. Meanwhile, Sharon Zhou, CEO of Lamini, is departing AMD to join Anthropic as VP of AI. In other news, Factory is acquiring Hum…
-
8 Habits to Maximize Claude AI's Performance and Workflow
This article provides practical advice for users looking to maximize their effectiveness when interacting with Claude, an AI model. It highlights eight key habits, including managing Claude's context window to avoid per…
-
Software engineers get a deep dive into how AI coding agents work
This article provides a technical deep dive for software engineers into the inner workings of AI coding agents. It explains the fundamental concept of an agent as an autonomous computer program driven by machine learnin…
-
Opus 5.5 and GPT-6 SOL show diminishing returns on coding benchmarks
A recent analysis of the SlopCodeBench benchmark suggests that increasing the effort level for models like Opus 5.5 and GPT-6 SOL may not yield significant improvements in strict pass rates, and can even lead to a decre…
-
Claude Code best practices: Keep CLAUDE.md files concise
Developers using Claude Code should keep their CLAUDE.md instruction files concise, ideally under 200 lines, to ensure Claude adheres to the directives. Overly long files can cause Claude to ignore important rules, dimi…
-
HumanLayer skill optimizes CLAUDE.md files for AI models
The `improve-claude-md` skill, developed by HumanLayer and documented on dev.to, aims to optimize CLAUDE.md files for AI models like Claude Code. It achieves this by restructuring the file, preserving essential project …
-
Cost analysis reveals CLAUDE.md lines are cheap but impact model attention
An analysis of the cost associated with using CLAUDE.md files with Anthropic's Claude models, specifically Opus 5.5, suggests that each line of CLAUDE.md costs approximately one-tenth of a cent per session when caching …
-
Claude Code CLAUDE.md Best Practices Guide Released
This article provides best practices for configuring CLAUDE.md, a file that automatically loads user preferences into Claude Code sessions. It analyzes five real-world CLAUDE.md files from sources like Karpathy and Anth…
-
Opus 5 benchmarked on SlopCodeBench for coding agent context engineering
A benchmark test was conducted on Opus 5, evaluating its performance on the SlopCodeBench dataset. The results of this evaluation, which focused on advanced context engineering for coding agents, were shared via a GitHu…
-
Context engineering: AI code generation needs human oversight
Dex Horthy, CEO of HumanLayer, has coined the term "context engineering" to describe the practice of understanding and working around the limitations of Large Language Model (LLM) contexts. Horthy's research, based on c…
-
AI agent performance hinges on harness, not just model
AI agents often underperform not due to the underlying model, but because of the 'harness' that surrounds it. This harness includes system prompts, tool descriptions, execution environments, and orchestration logic, ess…
-
AI infrastructure startups launch tools for agents, DevOps, security, and healthcare
Several startups are launching AI-powered tools aimed at improving infrastructure and developer productivity. Trigger.dev offers an open-source platform for building reliable AI agents and workflows, utilizing snapshott…