tech
Companies with huge datasets can save up to 80% in AI costs by using open-weight models: Him & Hers CEO
He said big patient datasets are Hims & Hers' "real asset," as they can be used to train open-weight models.
TL;DR
- Hims & Hers CEO Andrew Dudum suggests companies with large datasets use open-weight AI models to reduce costs.
- Dudum estimates savings can reach 70-80% compared to traditional AI models.
- Open-weight models allow access and customization of trained parameters, offering cost-effectiveness and performance benefits.
- Models trained on a company's specific data, like Hims & Hers' patient dataset, can surpass general market models.
- The trend reflects a shift in the tech industry from high AI token usage to cost optimization and efficient model routing.
- Companies are actively seeking data, with examples of Google and Handshake AI acquiring large datasets.