Reproduce “On the Impossibility of Supersized Machines” on OpenResearch
Turn your favorite coding agents into research agents that review literature, develop hypotheses, and run experiments.
Universal app for Apple silicon and Intel. Requires macOS 11 or later.
About this paper
Ben Garfinkel, Miles Brundage, Daniel Filan, Carrick Flynn, Jelena Luketina, Michael Page, Anders Sandberg, Andrew Snyder-Beattie and 1 more
This paper offers a satirical critique of specific anxieties within AI safety discourse concerning machines becoming physically "supersized" or uncontrollably large. It employs humor and deliberately absurd arguments to highlight perceived logical inconsistencies and advocate for more precise, nuanced thinking in the field, challenging overly speculative narratives about advanced AI.
Key insights
- AI safety research would benefit from more precise definitions for concepts such as "superintelligence" or "human-level AI" to clarify genuine concerns.
- Assessments of AI risks can be influenced by cognitive biases and emotional responses, suggesting a need for critical examination of underlying assumptions.
- The future of human-machine relationships likely involves complex interdependencies rather than a simple narrative of machines surpassing or replacing humans.
Summary by alphaXiv.
Tip: change arxiv.org to autoarxiv.org in any arXiv abstract link to open the paper here.