LLM will be replaced by Multiverse: humans will be employed in due daytoday job in those verses

Summary

The speaker argues that AI may move beyond language models toward “world-to-world” systems: simulated environments that model events and interactions, rather than primarily linking words, images, or sounds.

From language models to simulated worlds

The talk contrasts text-generating large language models (LLMs) and image-generation systems, such as diffusion models, with systems that create evolving environments. These environments would operate according to defined rules and potentially fundamental physics.

Potential applications

The speaker describes virtual laboratories where researchers could simulate chemical experiments, molecule growth, or models of the early universe. Researchers could test hypotheses and observe outcomes within these environments.

A different kind of intelligence

The speaker suggests that spatial and temporal dynamics, physical interactions, and experience within a simulated environment could support forms of intelligence that current LLMs do not provide.

Human participation and employment

Rather than simply replacing people, virtual worlds could create work for those who contribute to or help develop evolving virtual societies and environments. The speaker emphasizes human intuition and experience as valuable inputs.

Broader implications

The presentation frames simulated “multiverses” as a possible shift in how people understand reality, experience, and AI. It also argues that AI currently lacks the biological and evolutionary experience humans possess.

The speaker mentions Fei-Fei Li, known for her work on ImageNet, in connection with efforts to build world models. The talk also refers to SOMU (“self-operating mathematical universe”) as a related direction. The speaker predicts that AGI may arrive around February 2028; this is a forecast, not an established fact.

Source and Scope

Main speaker: Anirban Bandyopadhyay, presenting his own commentary and referring to research directions associated with Fei-Fei Li and SOMU.

Reviews, guides, or tutorials: None. The video is an overview and speculative analysis rather than a product review or how-to.