Are LTMs the next LLMs? This new type of AI can do what large-language models can't

"A major difference between LLMs and LTMs is the type of data they're able to synthesize and use. LLMs use unstructured data-think text, social media posts, emails, etc. LTMs, on the other hand, can extract information or insights from structured data, which could be contained in tables, for instance. Since many enterprises rely on structured data, often contained in spreadsheets, to run their operations, LTMs could have an immediate use case for many organizations."

"San Francisco-based Fundamental, founded roughly 18 months ago by CEO Jeremy Fraenkel, has made a public LTM model, NEXUS, which will allow organizations to tap into that data to make predictions and forecasts. The data types in the mix could include customer behavior, information from various sensors, or myriad other things-but again, it's all locked up in rows and columns."

Large-language models (LLMs) operate on unstructured data such as text, social media posts, and emails, while large tabular models (LTMs) work with structured, table-formatted data. LTMs can extract information, insights, predictions, and forecasts from rows-and-columns datasets like spreadsheets, sensor outputs, and customer records. San Francisco-based Fundamental has released a public LTM called NEXUS to let organizations tap structured data for forecasting and prediction tasks. Many enterprises rely heavily on structured data for operations, making LTMs immediately applicable across industries. LLMs currently cover roughly 20% of overall data, leaving structured data as a major opportunity for LTMs.

Read at Fast Company