What “game over” for the latest paradigm in AI might look like

Paywalled, care to summarize for us?

> This week I saw not one but three striking premonitions for how the scaling maximalism hypothesis might end.

> 1. There might not be enough data in the world to make scaling maximalism work. https://arxiv.org/pdf/2209.12407.pdf

> 2. There might not be enough available compute in the world to make scaling maximalism feasible.

> 3. Some important tasks might simply not scale. https://arxiv.org/abs/2002.06177

Note: scaling maximalism hypothesis -> "that AGI might arise from larger scale without fundamental new innovation"

> Some important tasks simply may not scale

Reminds me of that concept Steven Wolfram introduced me to called computational irreducibility

A low hanging fruit for A.I. to remove is to debunk the received wisdom of economic sophistry holding back the feelers to the next step change.

I’ll be honest, I have no idea what you’re trying to say here. Care to elaborate?

What “game over” for the latest paradigm in AI might look like (garymarcus.substack.com)