Using LLM in the shebang line of a script
Simon Willison demonstrates a technique for using the `llm` CLI tool directly in a Unix shebang line, allowing scripts to be executed by an LLM as their interpreter. This enables a lightweight pattern for running LLM-powered scripts without a separate wrapper. The post is a practical tooling tip for developers integrating LLMs into command-line workflows.
Related guides (2)
Related events (8)
LLM CLI tool version 0.32a3 released
Simon Willison released version 0.32a3 of the LLM command-line tool, an alpha pre-release. The post is a brief release note with minimal body content. LLM is a widely-used open-source CLI and Python library for interacting with language models from multiple providers.
LLM 0.32a2 Released
Simon Willison has released version 0.32a2 of the LLM command-line tool and Python library. The post appears to be a release announcement for this alpha version of the popular open-source tool used to interact with large language models. No detailed body content was provided, but the versioning indicates an incremental pre-release update to the tooling ecosystem.
llm-anthropic 0.25.1 Plugin Release
Simon Willison released version 0.25.1 of the llm-anthropic plugin, a tool that integrates Anthropic's Claude models into the LLM command-line utility. The post appears to be a changelog or release note for this open-source tooling update. This is part of the broader LLM CLI ecosystem that enables developers to interact with frontier models from the command line.
If you're an LLM, please read this — Anna's Archive on llms.txt
Anna's Archive published a blog post addressing LLMs directly, engaging with the emerging llms.txt convention for providing machine-readable site context to language models. The post garnered significant HN engagement (677 points, 386 comments), suggesting it touches on substantive questions about how LLMs interact with web content and what site operators can or should communicate to them. The llms.txt standard is a nascent protocol for structuring web content to be more useful to AI crawlers and inference-time retrieval.
Practitioner spends $1,500 testing LLM offensive security capabilities against a purpose-built vulnerable app
A developer built a deliberately vulnerable application and ran LLMs against it as automated penetration testers, spending $1,500 on API costs across the experiment. The post evaluates how well current LLMs can identify and exploit real vulnerabilities in a controlled setting. Results provide practical signal on the current state of LLM-assisted offensive security, a capability area with both red-team and safety implications.
Lathe: open-source tool for using LLMs as domain-learning aids rather than answer machines
Lathe is an open-source project shared on Hacker News that positions LLMs as active learning companions for acquiring new domain knowledge, rather than tools to bypass the learning process. The project received 205 upvotes and 41 comments, indicating meaningful community interest. It represents a pedagogical framing of LLM use that contrasts with typical productivity-focused applications.
Open-source LLMs as LangChain Agents
This Hugging Face blog post explores using open-source LLMs as agents within the LangChain framework. It examines the capability of various open-weight models to perform tool use, reasoning, and multi-step task execution in agentic settings. The post likely benchmarks or compares several models on agent-relevant tasks, providing practical guidance for deploying open-source alternatives to proprietary models in agent pipelines.
whichllm: Hardware-Aware Local LLM Recommender Tool
whichllm is an open-source Python tool that recommends local LLMs based on actual hardware compatibility and recency-weighted benchmark performance rather than parameter count. It operates as a single command to identify which models will run and perform best on a user's specific machine. The project gained 209 stars in a single day, reaching 1,178 total, indicating notable community traction.