Ask HN: Will progress on LLMs carry over to self driving?

With LLMs, huge training runs and multi modality both seem capable of producing a step change in the quality of models.

Are we likely to see a step change in self driving capability using similar techniques, or do the constraints around e.g. latency and on-board compute mean there’s not much carry over?

Comments URL: https://news.ycombinator.com/item?id=40186662

Points: 1

# Comments: 0