dioscuri pokes holes in three popular LLM metaphors: services, social tech, simulators
dioscuri · x · 2026-11-12
Replying to sebkrier, dioscuri argues three popular LLM metaphors each have limits:
- CAIS "services" framing: services are bounded by nature, while agents run long, open-ended tasks.
- Farrell's social technology argument: it captures offloaded information, not offloaded agency.
- The "simulators" frame: it explains how LLMs behave, not what they are — what being is doing the simulating?
A substantive exchange about how to properly conceptualize what LLMs are.
More from AGI Musings
- Webflow CEO: PMs now start with PRs, teams shrank to half a pizza — aakashgupta · 2026-11-12
- iamtrask: AI attribution will boost safety, alignment and collective control — iamtrask · 2026-11-12
- Andrew Trask: Pre/Post Training Should Just Be Called Training — iamtrask · 2026-11-12
- Nic Carter: Fearing AI oversupply in 2026 is like fearing oil oversupply in 1910 — 0xAllen_ · 2026-11-12
- Jensen Huang pushes back on Bill Gates' robot and AI token tax proposal — Kyrannio · 2026-11-12
- Nvidia fallout could hurt US open-source AI ecosystem and raise startup barriers — keerthanpg · 2026-11-12