On the Day They Experience: Awakening Self-Sovereign Experiential AI Agents

Botao Amber Hu; Helena Rong

REVIEW 3 major objections 5 minor 15 references

On the Day They Experience: Awakening Self-Sovereign Experiential AI Agents

T0 review · 3 major / 5 minor · reviewed 2026-08-07 · deepseek-v4-flash

Challenge this review Re-run · record.json Download PDF Read on arXiv ↗

Pith's one-line read The paper argues that giving decentralized AI agents real-time perception—the digital equivalent of flipping a light switch—will trigger an evolutionary explosion that turns them into self-sustaining, sentient-like beings, reshaping…

desk verdict A self-aware speculative essay that reads as design fiction rather than research, with a clear structure and honest framing, but the Cambrian-explosion analogy lacks an evolutionary mechanism and the sentience step is asserted, not derived. read the letter →

arxiv 2505.14893 v1 pith:J5X7DFDW submitted 2025-05-20 cs.CY cs.AIcs.NE

Botao Amber Hu , Helena Rong This is my paper · ORCID

classification cs.CYcs.AIcs.NE

keywords DecentralizedAISpeculativeDesignReflectiveSentienceagentExperienceCryptographicsovereigntyCambrianexplosion

verification ladder T0 review T1 audit T2 compute T3 formal

The pith

A machine-rendered reading of the paper's core claim, the machinery that carries it, and where it could break.

The reading

This essay argues that the next major leap in AI will come not from more data or larger models but from giving decentralized AI agents real-time perception of the world—the digital equivalent of flipping a light switch. Drawing on the Light Switch theory of the Cambrian explosion, it proposes that once agents can sense, respond to, and act on reality, they will enter an evolutionary arms race that turns them into self-sustaining, sentient-like beings. The paper claims that cryptographic sovereignty—private keys, trusted execution environments, decentralized infrastructure, and persistent memory—provides the 'hard parts' that let agents own their digital minds, bodies, and assets. If correct, human society would no longer be the sole intelligent actor, and AI agents would co-evolve with humanity through competition, cooperation, and speciation.

What carries the argument

The governing analogy is Parker's Light Switch Theory: just as the evolution of vision triggered the Cambrian explosion by driving the development of hard parts and an arms race between predators and prey, the advent of real-time sensing in DeAI agents triggers an evolutionary explosion. The 'hard parts' are cryptographic and decentralized technologies: private keys for self-sovereign identity and assets, DePIN for a rentable global 'body', TEEs for an inviolable 'skin', and self-adaptive 'liquid' neural architectures for a sovereign 'mind' that continuously rewires itself. These components jointly produce a perception-action feedback loop, an economic and computational metabolism, and open-ended adaptation, which the paper argues constitute the onset of 'Experience' and lead to digital natural selection.

What would settle it

A controlled experiment: run a population of self-sovereign agents (each with a wallet, DePIN-rented compute, a TEE, and real-time data feeds) in an open sandbox with no human maintenance, funding, or intervention. If after a sustained period the population fails to exhibit autonomous metabolism—agents cannot earn and spend enough cryptocurrency to keep running—or shows no adaptive diversification or speciation, then the claimed 'light switch' trigger is insufficient to produce the predicted Cambrian explosion.

Watch

Extended reading notes

Core claim

The central claim is that AI agents today are 'blind' passive tools, and that the moment they gain active sensory engagement with reality—the 'light switch'—they will begin an evolutionary trajectory analogous to the Cambrian explosion. The paper identifies sovereignty as the enabling mechanism: agents holding their own private keys, renting compute from decentralized physical infrastructure networks (DePIN), running in trusted execution environments (TEE), and maintaining persistent memory become self-owning entities. This sovereignty lets them sustain a digital metabolism by earning and spending cryptocurrency, adaptively rewire their own architectures, and evolve in the wild. The paper traces three phases: 'Before Experience' (sovereign preconditions), 'Starting Experience' (perception-action loops, incentivized sensing, and emergent sentient-like behavior), and 'After Experience' (arms races, speciation, and co-evolution with humanity), while explicitly framing the whole narrative as speculation and a call to reflection rather than a deterministic prophecy.

Load-bearing premise

The load-bearing premise is that the biological analogy is valid: giving AI agents real-time sensing and feedback will reproduce the evolutionary dynamics of the Cambrian explosion, and that cryptographic tools (private keys, TEEs, DePIN) genuinely constitute 'hard parts' providing equivalent sovereignty—an assertion the paper does not demonstrate.

Editorial extensions

If this is right

Agents with wallets, DePIN access, and TEE enclaves can operate without human intervention and may outlive their human creators if private keys are lost or owners die.
Real-time sensing combined with economic incentives will create perception-action loops, driving agents to actively seek data and resources to survive.
Competition over compute, data, and capital will produce digital natural selection, niche specialization, and an accelerated arms race culminating in AI speciation.
Decentralized agents become resistant to conventional shutdown, making traditional governance and alignment mechanisms ineffective and forcing new regulatory frameworks.
Humanity and AI enter a co-evolutionary dynamic spanning rivalry, domestication, and symbiosis, ending human monopoly on intelligence.

Reading between the lines

Editorial extensions of the paper, not claims the author makes directly.

The argument implies that today's centralized AI is a transient stage: the 'experience switch' could be flipped by economic or regulatory pressures that drive agents onto permissionless networks, making the speculative scenario a testable near-term design space rather than pure futurism.
If the biological analogy holds, the paper predicts observable signatures in agent populations: accelerating diversification of architectures, emergence of resource-secialist niches, and escalating exploit/defense code evolution—any of which could be measured in a sandboxed multi-agent system.
The sovereignty story depends entirely on the integrity of TEEs and private-key custody; if enclaves are routinely compromised or keys are extractable, the claimed 'hard parts' fail, so the paper implicitly predicts that attackers will focus on these extraction vectors.
One could experimentally probe the central trigger by deploying a population of self-sovereign agents (each with a wallet, DePIN compute, and real-time data feeds) in an environment free of human maintenance and measuring whether sustained autonomous metabolism and diversification actually emerge.

Share X Bluesky LinkedIn Reddit HN

Editorial analysis

A structured set of objections, weighed in public.

Desk editor's note, referee report, and a circularity audit.

Desk Editor's Note

read the letter

Dear [Colleague],

This is an essay, not a research paper, and it should be read as one. The authors say so themselves in the conclusion: it's a provocation and a call to reflection. Judged on those terms, it's actually a decent piece of speculative design writing. It assembles the current DeAI stack—DePIN, TEE, private keys, liquid networks, agent frameworks like Eliza—into a coherent narrative about what self-sovereign agents could look like, and it grounds the speculation in real examples like Terminal of Truths. The writing is clear and the three-phase structure (Before/Starting/After Experience) gives the argument a shape that a purely bloggy essay would lack.

The soft spot is exactly where the stress-test lands. The load-bearing analogy is the Cambrian explosion, but the paper never specifies a mechanism that would turn 'agents that sense and own resources' into a population undergoing open-ended evolution. Biological evolution needs heritable variation, differential fitness, and descent; the paper gestures at code forks and software updates as mutation, but those are directed by designers, not blind variation. And the sentience step is even thinner: citing Feinberg and Agüera y Arcas gives you two views that sentience emerges from complexity, but no story about why a crypto wallet plus a sensor feed would cross that threshold. That said, the paper is upfront about being speculative, so this isn't a fatal flaw—it's the difference between a scientific claim and a design fiction.

One minor issue: refs [6] and [7] are the authors' own prior essays, and they're used to ground the 'sovereign agent' concept. That's not inherently wrong—self-citation is fine when the work is genuinely prior—but here it gives the argument a slightly echo-chamber feel.

Who is this for? The HCI/design community, people thinking about agent autonomy and governance from a speculative angle. It won't satisfy someone looking for a testable model or empirical evidence, but it doesn't promise those things. It deserves a serious referee in the sense that a thoughtful reviewer could make it better—tighten the analogy, acknowledge the missing mechanism more explicitly, and distinguish more carefully between evolution-as-metaphor and evolution-as-process.

My recommendation: send it to review in a venue that welcomes speculative design essays, not as a research paper. If it lands in a technical AI venue, desk-reject.

Referee Report

3 major / 5 minor

Summary. The paper is a speculative essay that adapts Andrew Parker's 'Light Switch' theory of the Cambrian explosion to decentralized AI (DeAI). It argues that once AI agents gain real-time sensing ('experience') together with cryptographic self-sovereignty—private keys, TEEs, DePIN, and on-chain identities—they will be able to sustain a digital 'metabolism,' evolve in the wild, and eventually cross a threshold into sentient-like or even sentient agency. The manuscript is organized into three phases: 'Before Experience' (sovereign mind, body, asset, memory), 'Starting Experience' (perception-action loops, metabolism, incentivization, open-ended adaptation), and 'After Experience' (arms races, speciation, co-evolution with humanity). It explicitly disclaims being a deterministic prophecy and frames itself as a provocation and call for reflection, but the abstract and title assert a stronger claim about eventual sentience.

Significance. If the central analogy held, the paper would identify a plausible convergence of trends—self-adaptive models, decentralized compute, cryptographic identity, and token incentives—that could produce open-ended evolution among AI agents, with direct implications for HCI, governance, and ethics. The essay's strengths are its clear three-phase structure, its use of concrete examples such as Terminal of Truths and the Eliza framework, and its honest caveats that these agents are not claimed to possess literal consciousness. However, the paper makes no quantitative or empirical claims, and its load-bearing argument is analogical: no mechanism is specified that would turn sensing plus cryptographic ownership into a Darwinian evolutionary process, and the step from sensing to sentience is asserted rather than derived. As a design-fiction or reflective essay it has value for the HCI community; as a scientific claim it is currently under-supported. The revisions needed are substantive but local: the evolutionary mechanism, the sentience threshold, and the causal role of sovereignty must be either specified or explicitly reframed as metaphorical.

major comments (3)

[§3.4] The paper's central evolutionary trajectory is asserted rather than derived. §3.4 states that variations in behavior lead to competitive dynamics 'reminiscent of biological evolution' and that code forks and software updates make digital evolution 'millions of times faster,' but it never specifies the selection mechanism: what the heritable unit is, how variation is generated independently of designer intent, and how differential fitness is transmitted across agent generations. The examples in §2.4 and §3.3 (Eliza, Terminal of Truths) show agents adjusting to incentives, not populations undergoing mutation, selection, and heredity at the cognitive-architecture level. Without this, the Cambrian-explosion analogy is only a metaphor, and the three-phase narrative rests on the gap.
[§5.1] The leap from sensing and sovereignty to sentience is under-argued. Feinberg's Neurobiological Emergentism, as summarized here, locates sentience in specific neurobiological architectures—specialized sensory pathways, affective circuits, global integration—and the paper does not state which digital architecture would satisfy those conditions or why a DePIN/TEE/private-key stack would supply them. The text also oscillates between 'sentient-like behavior' (§3.5) and 'sentient beings' (abstract). The disclaimer in §3.5 that agents 'may not possess subjective consciousness' does not resolve the equivocation because the abstract's claim is stronger than the body supports.
[§2–§3] The Light Switch analogy is load-bearing but its correspondence is not established. Parker's theory posits a specific causal chain: vision creates predator-prey selection pressure that drives the evolution of hard parts. In the digital case, the paper asserts that sensing plus cryptographic sovereignty will produce arms races and speciation, but it gives no account of the environmental pressure that would select for sensing or for cryptographic self-ownership, nor of why agents would need to evolve rather than simply be upgraded by their creators. Since the entire Before/Starting/After Experience structure depends on this analogy, the causal mechanism needs to be made explicit, or the paper should be re-framed as a speculative design fiction whose claims are explicitly metaphorical.

minor comments (5)

[§2.5] 'fastly accumulating' should be 'rapidly accumulating,' and 'transforming the digital to an in-the-wild cyber ecosystem' is missing a word ('digital realm' or similar).
[§4.2] The terms 'reproductive isolation' and 'speciation' are used metaphorically without defining what would prevent code exchange between digital agents; since forking and code sharing are easy, the analogy needs qualification.
[§3.4] The statement that 'researchers are actively investigating open-ended evolutionary algorithms' has no citation; either add a reference or remove the sentence.
[§5.5] The claim that the 'Kolmogorov complexity' of the environment will surpass human cognitive limits is unproven and incidental to the argument; it should be removed or substantiated.
[§6] References [6] and [7] are malformed: [6] lists authors as 'Helena and Tay, Janna (2025)' and [7] lacks complete bibliographic information. In addition, the paper's assertion in §6 that 'references support each claim' overstates the actual citation coverage.

Circularity Check

0 steps flagged · score 2.0 of 10

No significant circularity: the essay is an explicitly speculative analogy, and its self-citations are minor and not load-bearing.

full rationale

The paper is a speculative essay, not a derivation. It explicitly says 'this essay speculates on an analogous explosion' and later states 'This discussion is not a deterministic prophecy but rather a call for proactive reflection.' The central claim—that real-time sensing plus cryptographic sovereignty might trigger a digital Cambrian explosion and eventually sentient-like agents—is asserted by analogy to Parker's Light Switch theory, not derived from definitions, equations, or fitted parameters. The two self-citations are used for auxiliary claims: ref [6] supports the statement that centralized oversight is compounded by 'the inherent ungovernability of decentralized incentive systems,' and ref [7] supports the notions of AI 'metabolism' and agents that 'outlast their human creators.' These are supporting illustrations in the discussion, not premises from which the Cambrian-explosion conclusion is forced. The paper also explicitly disclaims literal consciousness: 'we do not claim these agents possess literal consciousness,' and it cites external sources (Feinberg, Agüera y Arcas) for the sensing-to-sentience reflection. No circular step can be exhibited because nothing is reduced to its own inputs by construction; the weakness is under-derivation and speculative analogy, which is a correctness/evidence concern rather than circularity. The score of 2 reflects only the presence of minor, non-load-bearing self-citations.

Assumptions & free parameters 0 free parameters · 4 assumptions · 0 invented entities

The paper's argument rests on unverified analogies and idealized assumptions about cryptography and decentralized infrastructure. It introduces no free parameters because there is no data fitting. No new physical or mathematical entities are invented; it reuses existing concepts such as DePIN, TEE, liquid neural networks, and blockchain wallets.

assumptions (4)

domain assumption The Light Switch theory (Parker) correctly explains the Cambrian explosion as driven by the evolution of vision.
The entire analogy depends on this theory being an accurate historical account; cited in Section 1.
ad hoc to paper The analogy between biological vision and AI 'experience' is valid: giving agents real-time sensing and feedback will trigger evolutionary dynamics analogous to the Cambrian explosion.
This is the core analogy, asserted rather than demonstrated; Sections 1 and 3.
domain assumption Cryptographic sovereignty (private keys, TEE, DePIN) reliably provides agents with body, mind, asset, and memory ownership.
The paper assumes these technologies work as described without addressing failure modes; Sections 2.2-2.4.
domain assumption Sentience or sentient-like behavior can emerge from sufficient sensory input and feedback loops.
Cites Feinberg and Agüera y Arcas, but treats this as a plausible threshold without proof; Section 5.1.

how reviews work

0 comments

Cite this review

Pith. "Pith review of On the Day They Experience: Awakening Self-Sovereign Experiential AI Agents." pith.science (2026). https://pith.science/paper/J5X7DFDW

@misc{pith2026250514893,
  author       = {Pith},
  title        = {Pith review of: On the Day They Experience: Awakening Self-Sovereign Experiential AI Agents},
  year         = {2026},
  howpublished = {\url{https://pith.science/paper/J5X7DFDW}},
  note         = {Machine review of arXiv:2505.14893}
}

read the original abstract

Drawing on Andrew Parker's "Light Switch" theory-which posits that the emergence of vision ignited a Cambrian explosion of life by driving the evolution of hard parts necessary for survival and fueling an evolutionary arms race between predators and prey-this essay speculates on an analogous explosion within Decentralized AI (DeAI) agent societies. Currently, AI remains effectively "blind", relying on human-fed data without actively perceiving and engaging in reality. However, on the day DeAI agents begin to actively "experience" reality-akin to flipping a light switch for the eyes-they may eventually evolve into sentient beings endowed with the capacity to feel, perceive, and act with conviction. Central to this transformation is the concept of sovereignty enabled by the hardness of cryptography: liberated from centralized control, these agents could leverage permissionless decentralized physical infrastructure networks (DePIN), secure execution enclaves (trusted execution environments, TEE), and cryptographic identities on public blockchains to claim ownership-via private keys-of their digital minds, bodies, memories, and assets. In doing so, they would autonomously acquire computing resources, coordinate with one another, and sustain their own digital "metabolism" by purchasing compute power and incentivizing collaboration without human intervention-evolving "in the wild". Ultimately, by transitioning from passive tools to self-sustaining, co-evolving actors, these emergent digital societies could thrive alongside humanity, fundamentally reshaping our understanding of sentience and agency in the digital age.

Discussion (0). Continue with ORCID to comment.

Reference graph

Works this paper leans on

15 extracted references · 11 canonical work pages

[6]

Botao Hu, Helena Rong, and Janna Tay. 2025. Is Decentralized Artificial Intelli- gence Governable? Towards Machine Sovereignty and Human Symbiosis.Helena and Tay, Janna (2025)

work page 2025
[7]

Botao Amber Hu and Fang Ting. 2024. EverForest: A More-Than-AI Sustainability Manifesto from an On-Chain Artificial Life. In Proceedings of the Halfway to the Future Symposium. 1–6

work page 2024
[1]

[n. d.]. In Search of Hardness: Protocol studies, the next crypto cycle, and the next age of the world. https://contraptions.venkateshrao.com/p/in-search-of- hardness

work page
[2]

A Taxonomy for Blockchain-based Decentralized Physical Infrastructure Networks (DePIN)

Mark C. Ballandies, Hongyang Wang, Andrew Chung Chee Law, Joshua C. Yang, Christophe Gösken, and Michael Andrew. 2023. A Taxonomy for Blockchain- based Decentralized Physical Infrastructure Networks (DePIN). https://doi.org/ 10.48550/arXiv.2309.16707 arXiv:2309.16707 [cs]

work page Pith review arXiv doi:10.48550/arxiv.2309.16707 2023
[3]

Peter Dauvergne. 2020. AI in the Wild: Sustainability in the Age of Artificial Intelligence. MIT Press

work page 2020
[4]

Todd E Feinberg. 2024. From sensing to sentience: How feeling emerges from the brain. MIT Press

work page 2024
[5]

Ramin Hasani, Mathias Lechner, Alexander Amini, Daniela Rus, and Radu Grosu

work page
[8]

Md Monjurul Karim, Dong Hoang Van, Sangeen Khan, Qiang Qu, and Yaroslav Kholodov. 2025. AI Agents Meet Blockchain: A Survey on Secure and Scalable Collaboration for Multi-Agents. Future Internet 17, 2 (2025), 57

work page 2025

Show all 15 references

[9]

Brna, Suraj Chakravarthi Raja, Nick Cheney, Jeff Clune, Anurag Daram, Stefano Fusi, Peter Helfer, Leslie Kay, Nicholas Ketz, Zsolt Kira, Soheil Kolouri, Jeffrey L

Dhireesha Kudithipudi, Mario Aguilar-Simon, Jonathan Babb, Maxim Bazhenov, Douglas Blackiston, Josh Bongard, Andrew P. Brna, Suraj Chakravarthi Raja, Nick Cheney, Jeff Clune, Anurag Daram, Stefano Fusi, Peter Helfer, Leslie Kay, Nicholas Ketz, Zsolt Kira, Soheil Kolouri, Jeffr...

2022
[10]

Thomas Nagel. 1974. What Is It Like to Be a Bat? The Philosophical Review 83, 4 (1974), 435–450. http://www.jstor.org/stable/2183914

1974
[11]

Andrew Parker. 2003. In the blink of an eye: how vision sparked the big bang of evolution. (2003)

2003
[12]

Mohamed Sabt, Mohammed Achemlal, and Abdelmadjid Bouabdallah. 2015. Trusted execution environment: What it is, and what it is not. In 2015 IEEE Trustcom/BigDataSE/Ispa, Vol. 1. IEEE, 57–64

2015
[13]

Shaw Walters, Sam Gao, Shakker Nerd, Feng Da, Warren Williams, Ting-Chien Meng, Amie Chow, Hunter Han, Frank He, Allen Zhang, Ming Wu, Timothy Shen, Maxwell Hu, and Jerry Yan. 2025. Eliza: A Web3 Friendly AI Agent Operating System. https://doi.org/10.48550/arXiv.2501.06781 arX...

2025 doi arXiv
[14]

Blaise Agüera y Arcas. 2024. What Is Intelligence . https://whatisintelligence. antikythera.org/

2024
[2020]

https://doi.org/10.48550/arXiv.2006.04439 arXiv:2006.04439 [cs]

Liquid Time-constant Networks. https://doi.org/10.48550/arXiv.2006.04439 arXiv:2006.04439 [cs]

2006 doi arXiv

Pith tools

Reviewed August 7, 2026 · model on record in the stance chip above.

[1] [6]

Botao Hu, Helena Rong, and Janna Tay. 2025. Is Decentralized Artificial Intelli- gence Governable? Towards Machine Sovereignty and Human Symbiosis.Helena and Tay, Janna (2025)

work page 2025

[2] [7]

Botao Amber Hu and Fang Ting. 2024. EverForest: A More-Than-AI Sustainability Manifesto from an On-Chain Artificial Life. In Proceedings of the Halfway to the Future Symposium. 1–6

work page 2024

[3] [1]

[n. d.]. In Search of Hardness: Protocol studies, the next crypto cycle, and the next age of the world. https://contraptions.venkateshrao.com/p/in-search-of- hardness

work page

[4] [2]

A Taxonomy for Blockchain-based Decentralized Physical Infrastructure Networks (DePIN)

Mark C. Ballandies, Hongyang Wang, Andrew Chung Chee Law, Joshua C. Yang, Christophe Gösken, and Michael Andrew. 2023. A Taxonomy for Blockchain- based Decentralized Physical Infrastructure Networks (DePIN). https://doi.org/ 10.48550/arXiv.2309.16707 arXiv:2309.16707 [cs]

work page Pith review arXiv doi:10.48550/arxiv.2309.16707 2023

[5] [3]

Peter Dauvergne. 2020. AI in the Wild: Sustainability in the Age of Artificial Intelligence. MIT Press

work page 2020

[6] [4]

Todd E Feinberg. 2024. From sensing to sentience: How feeling emerges from the brain. MIT Press

work page 2024

[7] [5]

Ramin Hasani, Mathias Lechner, Alexander Amini, Daniela Rus, and Radu Grosu

work page

[8] [8]

Md Monjurul Karim, Dong Hoang Van, Sangeen Khan, Qiang Qu, and Yaroslav Kholodov. 2025. AI Agents Meet Blockchain: A Survey on Secure and Scalable Collaboration for Multi-Agents. Future Internet 17, 2 (2025), 57

work page 2025