Eric Schmidt emphasizes that there is "no evidence" the scaling laws for AI are slowing down. He insists that large models are still "scaling with an ability that is unprecedented." His confidence lies in the potential for two or three more significant advancements in large language models within the next five years, despite the observed performance plateau among current models.
Schmidt acknowledges a growing concern in Silicon Valley regarding the future of AI advancement. He notes that AI scaling laws, which predict improvement through increased data and computing power, have not shown signs of halting yet. "They will eventually stop, but we're not there yet," he reassures, reflecting on the ongoing discussions about the sustainability of these advancements in AI.
Collection
[
|
...
]