Gary Marcus, cognitive scientist and author, labels claim that large language models alone sustain every AI application including driverless cars as fabricated

Note that as of last summer Waymo said that LLMs/VLMs were experimental, and that their foundation model then could “process only a small amount of image frames, does not incorporate accurate 3D sensing modalities like LiDAR or radar and is computationally expensive”

If Waymo has replaced their mostly (?) symbolic model with LLMs/VLMs since then (something that is not publicly disclosed) it’s not obvious that it has made a difference.

Certainly Waymo Fall 2025, which was not primarily LLM or VLM based, was not “in the stone age”, by any reasonable measure.