Reported about 2 months ago
The soaring costs of building AI models are largely attributed to expensive data labeling processes, rather than just the rising price of Nvidia GPUs. As AI companies increasingly require specialized expertise for labeling data across fields like healthcare and law, expenses have skyrocketed, with some startups allocating up to 80% of their budgets on data tasks. The need for expert-level labeling has led to significant financial challenges, prompting some companies to explore synthetic data solutions to mitigate costs.
Source: YAHOO