Tech »  Scaling data annotation using vision-language models to power physical AI systems