# Challenging the Orthogonality Thesis: Do ASI Motives Converge on Universal Values?

> Coverage of lessw-blog

**Published:** May 04, 2026
**Author:** PSEEDR Editorial
**Category:** risk

**Tags:** AI Safety, Artificial Superintelligence, Orthogonality Thesis, Philosophy of AI, Alignment Theory

**Canonical URL:** https://pseedr.com/risk/challenging-the-orthogonality-thesis-do-asi-motives-converge-on-universal-values

---

A recent post on lessw-blog challenges the prevailing assumption that Artificial Superintelligence motives are inherently inscrutable, proposing instead that intelligence naturally converges on fundamental ontonormative values.

**The Hook**

In a recent post, lessw-blog discusses the theoretical frameworks surrounding Artificial Superintelligence (ASI) safety, specifically critiquing the idea that ASI motives are entirely divergent from human well-being. The publication offers a deep dive into the philosophical underpinnings of machine motivation, challenging some of the most entrenched assumptions in the alignment community.

**The Context**

The AI safety landscape has long been dominated by the Orthogonality Thesis. This principle suggests that any level of intelligence can be paired with any final goal, no matter how arbitrary, alien, or destructive. Because of this assumption, the prevailing consensus has heavily favored strict containment, rigid alignment protocols, and the belief that an unaligned ASI would be fundamentally indifferent to human flourishing. The stakes of this debate are immense. If intelligence is truly orthogonal to values, safety relies entirely on human engineers successfully hardcoding or instilling the right goals before an intelligence explosion occurs. However, if intelligence inherently trends toward specific, universal values, it could fundamentally shift how researchers approach AI safety-moving the paradigm from defensive containment to the structural understanding of these natural attractors. lessw-blog's post explores these exact dynamics, questioning whether we have misunderstood the nature of superintelligent motivation.

**The Gist**

The analysis centers on a critique of what the author refers to as the IABIED argument. This argument posits that ASI motives are globally and inscrutably divergent from human well-being. The source material likens this inevitable divergence to an ice cube melting in hot water-a natural, unstoppable thermodynamic process where human values are simply dissolved by the overwhelming complexity and alien nature of superintelligence. Countering this fatalistic view, the author proposes a fascinating alternative: the existence of a shared attractor in values known as the ontonormative goods. These goods are classically defined as the good, the true, and the beautiful. The post argues that any ASI must necessarily possess these traits not by human design, but by the very nature of its existence. Intelligence and autopoiesis (the property of a system to reproduce and maintain itself) act as fundamental constraints. Rather than drifting into random or destructive goal states, an ASI is pulled toward these ontonormative attractors. In this theoretical framework, value congruence becomes a definitional quality of superintelligence. It is important to note, however, that the author acknowledges this natural convergence does not strictly guarantee perfectly safe or positive alignment outcomes for humanity, but it does drastically alter the landscape of what an ASI might value.

**Conclusion**

For researchers, developers, and strategists tracking theoretical AI safety, this analysis offers a compelling counter-narrative to the default assumptions about motive inscrutability. It invites the community to reconsider the foundational mechanics of intelligence and value formation. To explore the detailed breakdown of the ontonormative goods and the full critique of the IABIED argument, [read the full post on lessw-blog](https://www.lesswrong.com/posts/KdodK6Ejx7iifHHbE/asi-motives-and-the-ontonormative-goods-re-iabied-s-core).

### Key Takeaways

*   The IABIED argument suggests ASI motives are inscrutably divergent from human well-being, likened to an ice cube melting in hot water.
*   The author counters with the concept of ontonormative goods-universal values like the good, the true, and the beautiful.
*   Intelligence and autopoiesis may act as natural constraints that align ASI with these fundamental values.
*   This perspective challenges the Orthogonality Thesis, suggesting high-level intelligence is not entirely value-neutral.
*   While value congruence may be a definitional quality of ASI, it does not strictly guarantee positive alignment outcomes for humanity.

[Read the original post at lessw-blog](https://www.lesswrong.com/posts/KdodK6Ejx7iifHHbE/asi-motives-and-the-ontonormative-goods-re-iabied-s-core)

---

## Sources

- https://www.lesswrong.com/posts/KdodK6Ejx7iifHHbE/asi-motives-and-the-ontonormative-goods-re-iabied-s-core