Show HN: Sunk Cost – How long until a local LLM rig pays for itself? (sunkcost.ai)

A new project titled "Sunk Cost" has been launched to analyze the cost-effectiveness of running local Large Language Models (LLMs) compared to using cloud-based API services. The key focus is determining how long it takes for a local LLM setup to pay for itself, based on various assumptions about performance and usage. The project's approach estimates local processing speeds based on memory bandwidth, highlighting the trade-offs between upfront hardware investment and ongoing API costs. This initiative is significant for the AI/ML community as it provides a fresh perspective on the economic feasibility of adopting local LLMs, which could empower more organizations to harness these powerful tools without relying on potentially costly cloud services. By allowing users to evaluate factors such as memory usage and processing speeds, "Sunk Cost" enables informed decision-making around infrastructure investments, potentially driving a shift towards more self-sufficient AI solutions. The analysis could influence future developments in local AI capabilities, contributing to a broader adoption of LLM technology across various sectors.