A developer working with Astra created a font where each token has equal width to help visualize how language models process text. The tool allows users to generate custom token-space fonts by combining a base font with various tokenizers, and can be installed in Discord and Slack to view AI outputs as the model sees them.
Shipwithmuse.live is a curated catalog of 156 projects built by developers using Meta's Muse personal agent, showcasing real-world applications across 16 categories. Examples include AI-assisted flight check-ins, Beat Saber song pickers, Slack integrations, and autonomous financial tools that help users accomplish tasks from exercise planning to unclaimed property discovery.
A developer compares emoji search implementations using Jeval (an LLM-based classification tool) versus embedding-based approaches. Embedding models prove faster and significantly cheaper while delivering comparable or better search quality, suggesting embeddings are often better suited for classification tasks than LLMs.
A tool that compiles custom fonts where each LLM token occupies equal width, enabling visual representation of tokenization in Discord and Slack. Users select a base font and tokenizer, compile the font, and install it alongside a theme or userscript to display messages with uniform token widths.
A Hacker News discussion asking users which subscription services or software tools (personal or professional) they would switch away from if a better alternative became available, using Spotify and Slack as examples.
Kern.md is an agent runtime that maintains a single persistent session across multiple interfaces (Telegram, Slack, Matrix, Discord, TUI, web). The agent can read its configuration and secrets, remembers context across channels, and supports user pairing with role-based access control.
MCP server tool definitions consume significant tokens on every conversation turn—a typical multiserver setup costs ~55,000 tokens before processing user input. Beyond 30–50 tools, token costs degrade model accuracy in tool selection, making sprawling servers expensive both financially and in correctness.
MarkAI is an AI-powered platform that monitors public online communities to help startups find prospect conversations discussing their problem, even when customers don't use official product terminology. It analyzes discussions across seven platforms, scores relevance, and sends timely alerts to email or Slack when strong matches appear.
Praxos is a team messaging platform for people and AI agents that preserves conversation context for later reference, eliminating the need to repeatedly explain information or coordinate with others. Built for small teams and solo founders, it integrates with tools like Claude and Slack to centralize context across multiple conversation threads.