🦞🌯 Lobster Roll

Stories by mkagenius

I want everything local – Building my offline AI workspace (instavm.io)
Program-of-Thought Prompting Outperforms Chain-of-Thought by 15% (2022) (arxiv.org)
Anti-patterns while working with LLMs (instavm.io)
Show HN: OpenSkills - Run Claude Skills Locally Using Any LLM (github.com)
Show HN: LLM Token Visualizer – How big is 128k token input (v0-llm-token-visualizer.vercel.app)
Show HN: Use Apple Container with Gemini CLI (github.com)
Gemini CLI has native support for sandbox via Docker but this integration[1] extends it to Apple Containers on M1&#x2F;M2&#x2F;M3 Mac.<p>1. <a href="https:&#x2F;&#x2F;github.com&#x2F;BandarLabs&#x2F;coderunner?tab=readme-ov-file#option-3-gemini-cli">https:&#x2F;&#x2F;github.com&#x2F;BandarLabs&#x2F;...
Gemini is so good, I let it control/use my phone (old.reddit.com)
Show HN: Weekly Podcast on CVEs (podcasts.apple.com)
Hello HN,<p>Keeping up with security vulnerabilities each week can be overwhelming. To make it easier, we’ve created The Exploit Podcast—an automated weekly podcast covering CVEs and security news.<p>You can listen on:<p>Apple Podcasts: <a href="https:&#x2F;&#x2F;podcasts.apple.com&#x2F;us&#x2F;podc...
I can no longer write technical blogs due to coding agents (gist.github.com)
Show HN: Can this model run on my Mac mini? (mac-ai-model-compatibility-checker-911416006313.us-west1.run.app)
Anthropic Skills Can Be Programmatically Optimized (Using DSPy) (instavm.io)
Show HN: I built a mitmproxy AI agent using 4000 paid security disclosures (instavm.io)
I Used 4,000 Paid Bug Bounty Reports to Build a CLI-First Security Agent (instavm.io)
InstaPoT: Using InstaVM with DSPy's Program of Thought (instavm.io)
Show HN: Run Claude Skills locally on your Mac (no cloud upload) (github.com)
Show HN: CodeRunner – Generate and run LLM code locally on Apple containers (github.com)
We used Apple Containers to run LLM generated code locally.<p>Leverage remote LLMs (like ChatGPT or Claude Sonnet 4) to work with your local files—such as videos—securely on your Mac. The LLM runs in a sandboxed environment where it can install external libraries, generate code, and execute it local...
Show HN: Sandbox Flow – A Playground for Sandboxes (github.com)
The combination of a browser, a LLM and a sandbox seems like the perfect turing machine for the AI agent world - it can probably achieve any task there is to automate.<p>There can be many ways to play around with this combination but using visual workflow and nodes seemed the most intuitive way, so ...
Music Maker: Automate Music Generation with Claude Code (github.com)
I Want Everything Local – Building My Offline AI Workspace (incl. Claude Skills) (instavm.io)
Claude Skills but Local (and Private) (instavm.io)
Association Between Coffee Consumption and Brain MRI Parameters (2023) (mdpi.com)
Show HN: I built an AI that turns GitHub codebase into a podcast (github.com)
Show HN: CVEingest – Prompt Friendly CVEs (github.com)
Creates a tool to get the latest CVEs in a LLM friendly form. You can ingest them into NotebookLM for example and generate a podcast.<p>I have used it to create a podcast channel for weekly CVEs discussion (<a href="https:&#x2F;&#x2F;podcasts.apple.com&#x2F;us&#x2F;podcast&#x2F;the-exploit-podcast-c...