What's the best AI model for Jewish textual research?
A friend asked me which AI model is best for Jewish textual research. The answer is that it's the wrong question — the model isn't where the knowledge lives. Here's what to do instead.
Tag
ai
88 posts
Testing Gemini 3.1 Flash Lite's Audio Understanding With 49 Structured Prompts
I built a 49-prompt test suite to evaluate Gemini 3.1 Flash Lite's audio understanding capabilities across 13 categories — from accent detection to deception analysis. Here's what worked, what didn't, and why it matters.
Hardware probe agent: SSH-based system analysis with AI upgrade recommendations
A Python tool that connects to remote systems via SSH, probes hardware specs, and generates detailed reports with AI workload upgrade recommendations.
A curated guide to Hebrew language AI models
A comprehensive collection of Hebrew language AI models on Hugging Face, covering LLMs, speech recognition, sentiment analysis, and more.
My AI stack: tools and infrastructure for a productivity-focused setup
A detailed look at my constantly evolving AI stack, from LLM APIs and frontends to vector storage, orchestration, and Docker deployment.
Using career data as AI context for smarter job search assistance
A structured context data repository of career information designed to power AI assistants for cover letters, resume tailoring, and job search support.
A demo public LLM context repository
A demonstration of how to structure a public context data repository for seeding LLM-based assistants with personal information.
Ideating personal context data stores for better AI experiences
Using AI to brainstorm and seed personal context data stores that make LLM interactions exponentially more useful and personalised.
Personalised AI: a proposal for proactive, interview-based context generation
An architecture proposal for generating personalised AI context through deliberate interview-driven workflows rather than passive data collection.
Building a personal RAG pipeline with AI interview agents
A workflow for building personal context data stores using specialized AI agents that interview you and feed structured data into RAG pipelines.
Shakespearean text generators: the internet's largest collection of Bard-ifying prompts
A massive collection of prompts for rewriting any text in Shakespearean English, organized by format and purpose.
Text transformation prompts for turning speech into polished content
A curated collection of system prompts that transform raw speech-to-text output into polished formats like emails, blog posts, and business documents.
Text Transformation Prompt Library: reformatting dictated text with AI
A comprehensive library of system prompts for transforming dictated or raw text into structured formats like emails, docs, and reports.
General System Prompts: a curated collection for everyday LLM use
A repository of general-purpose system prompts for configuring LLMs as useful everyday assistants, with personality and context modules.
System Prompt Factory: a modular system prompt generator
A Streamlit app for building custom LLM system prompts by mixing and matching configurable identity, style, and output parameters.
Hebrew Dash: a productivity dashboard with Hebrew calendar integration
A personalized desktop dashboard combining Hebrew calendar data, zmanim, weather, email, and Google services for daily productivity.
One Prompt AI Book: can Gemini 2.5 write a book in a single prompt?
Testing whether Gemini 2.5's 65K token output limit can produce a full book from one prompt. Spoiler: Anthropic did it better.
The Jerusalem Odyssey: a 100-page AI-generated travel memoir from a single prompt
An experiment generating a 100+ page fictitious travel memoir with Sonnet 3.7 from a single prompt, featuring a magical talking sloth.
Bias and censorship evaluation tests for large language models
Test prompts and notes exploring cultural bias, censorship, and worldview differences across Western and Eastern large language models.
Weird AI Bots: a library of intentionally offbeat AI character configs
A collection of system prompts for intentionally weird and funny AI characters, designed for speech-to-speech platforms.
Two AIs Talk: what happens when you pit two AI agents against each other
A fun experiment setting up two AI agents with conflicting secret missions and watching them try to interrogate each other.
Vibe Code and Learn: a subagent model to avert AI feudalism
Planning notes for a multi-subagent system that turns AI code generation into structured learning opportunities, not just output.
Code Tutor Agent Model: using AI as a code tutor, not just a code monkey
Exploring the idea of using AI code generation tools not just to write code, but as structured tutors that teach you what they did and why.
Career Pivot AI: using AI to explore career transitions
An AI-powered tool built with Google AI Studio to help people explore and plan career pivots using Gemini's capabilities.
Mapping the multimodal AI landscape with a structured taxonomy
AI platforms make it hard to filter models by multimodal capabilities. I built an open-source taxonomy that maps which inputs produce which outputs.
Using Gemini's vision capabilities for body language analysis
I built a body language analysis app using Google AI Studio's vibe coding interface and Gemini's multimodal vision. Upload a photo, get expert-level analysis.
LLM Detective: an undercover AI agent that evaluates other LLMs
I built an AI agent that goes undercover to test other LLMs - probing for biases, guardrails, knowledge cutoffs, and behavioral patterns.
Non-code uses of Claude Code: a curated collection
Claude Code isn't just for writing software. I've been collecting projects that use it for research, writing, budgeting, therapy tracking, and more.
Which AI image generators can actually render Hebrew text?
I tested 12 text-to-image models on their ability to render Hebrew. Only 2 out of 12 got both test words right. Here are the results.
Building a machine-readable resume format for AI agents
Traditional resumes are built for human eyes. I created an open JSON schema that gives AI recruiters and screening agents structured, queryable candidate data.
Custom License Creator: an AI-powered tool for generating open source licenses
A Gradio web app that uses AI to recommend existing open source licenses or generate custom ones based on your requirements described in plain English.
Policy Visualiser: exploring global policy approaches with AI-powered clustering
An AI-powered React app that analyzes how different countries approach policy challenges, with interactive clustering visualizations powered by Gemini.
Voice Blog Creator: turning voice recordings into polished blog posts with Gemini
An automated pipeline that converts raw voice recordings into polished blog posts using audio preprocessing, Gemini transcription, and AI-powered formatting.
Voice Analyzer: an AI-powered voice analysis tool built with Gemini
A voice analysis application built with Google AI Studio and the Gemini API, exploring multimodal AI capabilities for audio processing.
Vendor agent CLIs: tracking first-party AI coding tools from model providers
A curated list of command-line AI coding tools maintained by the model vendors themselves, from Claude Code to Gemini CLI.
Jewish-interest MCP projects: connecting AI to Jewish texts and calendars
A small curated collection of MCP servers for accessing Jewish texts, libraries, and calendar data through AI agents.
On MCP and consolidation: why the current setup is painful and what might fix it
Notes and resources on the pain points of MCP adoption and why consolidation through gateways and connectors matters.
MCP gateway model: notes on a better architecture for MCP management
Planning notes and sketches for an MCP gateway architecture that aggregates servers into LAN and WAN gateways.
Make Agent Friendly: preparing codebases for agentic development
A Claude Code plugin that systematically prepares human-built codebases for agentic AI development workflows.
Declaude: custom text rewriting rules for AI-generated content
A tool for maintaining personalized writing rules and vocabulary preferences, consolidated into a Claude Code slash command.
Ecosystem Mapper: an AI agent that visualizes technology landscapes
An experimental AI agent that discovers, categorizes, and visualizes technology ecosystems from a single keyword search.
Israel-related MCP servers: a curated list
A curated directory of Model Context Protocol servers providing access to Israeli data sources, government APIs, and services.
An index of Linux-friendly voice technology tools
A curated index of 100+ voice technology tools accessible to Linux desktop users, from real-time dictation to dev frameworks.
Audio multimodal AI: the models that understand sound, not just transcribe it
A curated resource list of multimodal AI models with native audio support — models that process audio tokens, not just transcribe.
Benchmarking speech-to-text on long-form audio
Comparing 8 STT models on a 27-minute podcast. Local Whisper wins on word accuracy, but cloud APIs dominate punctuation.
Whisper fine-tuning resources I keep coming back to
A short curated list of the best Whisper fine-tuning resources: tutorials, notebooks, and managed compute examples.
Does fine-tuning Whisper actually improve accuracy?
Evaluating whether fine-tuning Whisper improves transcription accuracy. Spoiler: it depends on model size and use case.
Fine-tuning Whisper on Modal's serverless GPUs
A script for fine-tuning OpenAI's Whisper speech recognition models using Modal's serverless GPU infrastructure.
Deepgram voice keyboard: a Linux virtual keyboard powered by Deepgram Flux
A voice-controlled Linux virtual keyboard using Deepgram's Flux turn-taking STT API, built in Rust.
ASR Training Data Collector: a GUI for gathering speech recognition training data
A GUI tool for collecting audio training data for ASR fine-tuning, with LLM-generated prompts and Hugging Face integration.
Cloud ASR MCP: multi-backend transcription via multimodal LLMs
An MCP server for audio transcription using multimodal LLMs like Gemini, GPT-4o Audio, and Voxtral — not traditional ASR.
Gemini Transcription MCP: audio transcription as an MCP tool
An MCP server that brings Gemini-powered audio transcription directly into Claude Code and Claude Desktop.
AI Transcription Notepad: multimodal cloud transcription for desktop
A desktop transcription app that sends audio directly to multimodal AI models for single-pass transcription and formatting.
Parakeet Dictation: on-device voice typing for Linux that actually works
A local voice typing app for Linux/Wayland using NVIDIA's Parakeet model. No cloud, no GPU, built-in punctuation.
Comparing Hebrew text-to-speech providers in 2025
A snapshot comparing Hebrew TTS quality across six providers, including voice cloning experiments via Replicate.
End of day worklog: voice-driven daily retrospectives with AI
A concept for capturing end-of-day work progress via voice memos, processing them with Gemini AI, and delivering morning briefings.
Peace in the Middle East: a multi-agent geopolitical experiment
An experiment using AI agents to simulate geopolitical dialogue between state actors, non-state actors, and civil society in the Middle East.
Claude tech research team: a multi-agent template for technology evaluations
A multi-agent system template for conducting comprehensive software and hardware technology evaluations using Claude Code.
AI Agent UN: simulating the General Assembly with AI diplomats
An experimental Model United Nations simulation where AI agents embody country positions, vote on resolutions, and analyze bilateral impacts.
Panel of Claudes: multi-perspective AI analysis through expert panels
An experimental multi-agent system that simulates expert panel discussions, analyzing complex topics through multiple analytical lenses.
AI Symposium 2026: simulating a conference with agentic personas
An experiment in perspective synthesis — using AI agents to simulate a conference where diverse personas deliver speeches on AI's impact.
Agent Junction: letting Claude Code instances talk to each other
An MCP server that lets multiple Claude Code sessions communicate via encrypted peer-to-peer messaging on localhost or LAN.
Visualising the symphony of agentic AI architecture
A framework-agnostic data model and interactive visualization mapping every layer of a production agentic AI system.
93 slash commands for managing a Linux desktop with Claude Code
A collection of 93+ slash commands for Claude Code CLI covering system admin, optimization, and configuration on Linux desktops.
Claude Code as a Linux server management tool
Another server management template for Claude Code, focused on general Linux server administration with 38 commands and 10 agents.
A Claude Code template for managing Proxmox servers
A template repo with 38 slash commands and 10 agents for managing Proxmox and home servers through Claude Code CLI.
Managing Home Assistant with Claude Code slash commands
A template repository with 20+ slash commands and 5 specialized agents for managing Home Assistant OS through Claude Code.
Evaluating GitHub repos systematically with Claude Code
A template repo that provides a structured workflow for evaluating and shortlisting GitHub repositories using Claude Code.
Taming chaotic filesystems with a Claude Code slash command
A deterministic slash command for Claude Code that transforms messy file structures into well-organized, maintainable systems.
Claude Agent Picker Pattern: smart subagent crew assembly for Claude Code
A pattern for intelligently assembling multi-agent crews in Claude Code while staying within context window limits.
Claude Business Idea Evaluator: systematic scoring for startup ideas
An AI-assisted framework for evaluating business ideas using the ICEC methodology — Impact, Confidence, Ease, and Contextual Suitability.
Claude Decision Evaluation Framework: seven analytical lenses for tough choices
A Claude Code framework that analyzes decisions through seven frameworks including SWOT, cost-benefit, and regret minimization.
Claude Deep Research Model: a pipeline pattern for thorough AI research
A structured model for conducting deep research with Claude Code, featuring pipeline configs, context management, and slash commands.
Claude Therapy Tracker: using repos for more than just code
A template for using Claude Code and git repos to track therapy sessions, mental health goals, and personal growth.
Claude Spec Starter: a template for spec-driven development with Claude Code
A repo template that transforms free-form project descriptions into organized specifications Claude Code can reliably parse.
A Claude Code diary and planner workspace for daily, weekly, and monthly planning
A Claude Code workspace template for workflow planning and time management with specialized agents for scheduling, goal tracking, and task triage.
Managing household budgets with Claude Code: a workspace template
A Claude Code workspace template for household budget management with specialized agents for expense analysis, goal tracking, and report generation.
An open source ADHD drug policy research workspace built on Claude Code
A real-world example of using a Claude Code research workspace template for ADHD drug policy research and advocacy.
Claude Think Tank: using a Git repo as a policy research environment
A Claude Code template that turns a Git repository into a structured policy research environment with specialized agents and workflows.
A writing workspace for Claude Code that respects your voice
A structured workspace template for writing projects with Claude Code, built around preserving authorial voice and systematic version control.
Claude Code as an Android device manager: the ADB workspace template
A version-controlled workspace template for managing Android devices through ADB with Claude Code CLI assistance.
Geopol Forecaster: a multi-agent system for geopolitical analysis
A proof-of-concept multi-agent system that gathers real-time news intelligence and generates scenario-based geopolitical forecasts from six analytical lenses.
Building a real-time rocket alert monitoring stack for Israel
A Docker microservices stack for monitoring Pikud HaOref red alerts with real-time maps, push notifications, AI sitreps, and Home Assistant integration.
Promise Denied: charting Iran's escalating strike campaign with open-source data
Charts and analysis from the Promise Denied OSINT project tracking four rounds of Iranian missile and drone strikes against Israel and coalition targets since April 2024.
Claude Code Repos Index: 100+ Repositories Beyond Traditional Development
Introducing the Claude Code Repos Index — a curated directory of 100+ repositories exploring Claude Code as a multi-purpose agent workspace.
Constructing The Dataset Of Iran's Missile War Against Israel
An overview of Promise Denied, an open-source intelligence project documenting Iran's ballistic missile, cruise missile, and drone attacks against Israel.
Whisper Hebrish: A Code-Switching ASR Fine-Tune for English-Hebrew Speech
How I fine-tuned OpenAI's Whisper to handle English-Hebrew code-switching, the natural language mixing that happens when immigrants blend two languages mid-sentence.
Claude Code Projects Index: 75+ Agent Workspace Blueprints Beyond Traditional Coding
A curated index of 75+ Claude Code projects and agent workspace templates spanning sysadmin, legal, health, finance, and more.