AI¶

2026¶

Searchable City: A semantic map of NYC built by applying Vision Language Models to millions of Google Street View images, enabling search by descriptive terms like "Gothic" or "Chinese" to reveal urban patterns invisible to traditional mapping. Learn more at GitHub [Keywords: AI VLM semantic search street view urban NYC geospatial]
Earth Copilot: An AI-powered geospatial application enabling exploration and visualization of Earth science satellite data through natural language queries, built on Azure AI services and Semantic Kernel agents. [Keywords: AI Microsoft NASA satellite data natural language geospatial]

PocketFlow Tutorial Codebase Knowledge: An AI-powered tool that analyzes GitHub repositories and automatically generates beginner-friendly tutorials explaining how codebases work, built on the 100-line Pocket Flow LLM framework. [Keywords: AI LLM code documentation GitHub tutorial generation]
AI and knowledge work: Twitter thread discussing the impact of AI on knowledge work and productivity. [Keywords: AI knowledge work productivity]

Vision to Text, Geospatial Foundation Model ChatGPT vision: Discussion on using ChatGPT vision API for geospatial tasks and satellite image interpretation. [Keywords: AI ChatGPT vision model geospatial satellite imagery foundation model]
Aurora: A Foundation Model for the Earth System: Microsoft's AI foundation model for atmospheric forecasting — weather prediction, air quality, and ocean wave modeling — pre-trained on large datasets and adaptable to specialized tasks. [Keywords: AI foundation model weather climate ERA5 atmospheric forecasting Python]
geospy.ai: AI tool that matches photos to their geographic location using image recognition. [Keywords: AI geolocation image recognition photo matching]

Voice Cloning: Voice cloning implementation using Bark text-to-audio model with custom voice cloning support. [Keywords: AI voice cloning Bark text-to-speech Python]