Summary of Computational Dualism and Objective Superintelligence, by Michael Timothy Bennett
Computational Dualism and Objective Superintelligence
by Michael Timothy Bennett
First submitted to arxiv on: 2 Feb 2023
Categories
- Main: Artificial Intelligence (cs.AI)
- Secondary: Logic (math.LO)
GrooveSquid.com Paper Summaries
GrooveSquid.com’s goal is to make artificial intelligence research accessible by summarizing AI papers in simpler terms. Each summary below covers the same AI paper, written at different levels of difficulty. The medium difficulty and low difficulty versions are original summaries written by GrooveSquid.com, while the high difficulty version is the paper’s original abstract. Feel free to learn from the version that suits you best!
| Summary difficulty | Written by | Summary |
|---|---|---|
| High | Paper authors | High Difficulty Summary Read the original abstract here |
| Medium | GrooveSquid.com (original content) | Medium Difficulty Summary This paper challenges the concept of intelligent software by revealing that its behavior is determined by the hardware that interprets it. The authors argue that this “computational dualism” undermines claims about software superintelligence and propose a pancomputational alternative to avoid subjective claims. They formalize systems as behaviors and cognition as embodied, embedded, extended, and enactive, allowing for objective claims regarding intelligence. Specifically, they define intelligence as the ability to generalize, identify causes, and adapt, and establish upper bounds for intelligent behavior. This perspective suggests that AGI will be safer but more limited than previously theorized. |
| Low | GrooveSquid.com (original content) | Low Difficulty Summary This paper shows that our idea of super smart computer programs is wrong. Instead, how a program behaves depends on the machine it runs on. The authors say this means we can’t make fair claims about super intelligent computers because they’re tied to the machines that “understand” them. They have a new way of looking at things, where everything in the world is connected and there’s no difference between the computer and what it’s doing. This lets us make fair claims about how smart something is. The authors think intelligence means being able to figure out patterns, find causes, and adapt. They also set limits for how smart something can be. This changes our ideas about super intelligent computers and makes them seem safer but not as powerful as we thought. |
