AI is polytheistic, not monotheistic
AI summary of the linked articleChris Dixon's essay argues that AI is best understood as polytheistic rather than monotheistic, with many strong models from different factions converging on similar capabilities, so no single dominant AGI is likely. The text presents AI as constrained rather than omnipotent: it is economically constrained by costly API calls, mathematically constrained by its inability to solve chaotic, turbulent or cryptographic equations, practically constrained by the need for prompting and verification, and physically constrained by its reliance on humans to supply context. The article also claims that AI is better suited to visuals than text because interfaces and images are easier for humans to verify, and that AI is empirically decentralizing. |
Recommended by 1 curator |
1 comment My main problem is with "AI is better for visuals than verbals. That is, AI is better for the frontend than the backend, and better for images/video than for text." It would be nice to have the "flappy bird app coded in an instance" experience, but mine is often more like this one https://youtu.be/EL7Au1tzNxE?si=2HnFl_Ym_vJUy_kL (Lots of hinting and nudging the model and then the usage limits are reached.) |
More from open.substack.com on Kiwi News |
