Episode Details

When Daniel needed to find an industrial trolley for his apartment in Israel, he discovered that searching visually with Google Lens and the Hebrew word for "price" — mechir — was the only reliable way to cut through the country's fragmented e-commerce landscape. This episode unpacks why Israel's small retailers use inconsistent terminology across languages, how Google Lens actually works under the hood (it's not just image search), and why the mechir hack exploits a gap in the visual indexing pipeline. We also explore what a proper MCP-based solution would look like: an AI agent that could take product images and return structured results from local vendors. The technical challenges are real — crawling hundreds of small sites with no APIs — but the vision points toward a future where visual procurement is programmable, not manual.