pith. sign in

arxiv: 2606.04025 · v1 · pith:W2CZPWMOnew · submitted 2026-06-01 · 💻 cs.SE · cs.AI

The Biomimetic Architecture of Software 4.0

Philip Sheldrake , Dirk Scheffler This is my paper

Pith reviewed 2026-06-28 13:29 UTC · model grok-4.3

classification 💻 cs.SE cs.AI
keywords Software 4.0autopoietic heterarchyRecognitivebiomimetic architecturesymbolic substrateself-regulating softwareLLM integrationmetabolic network
0
0 comments X

The pith

Software 4.0 turns static code into a self-regulating metabolic network that verifies and evolves its own structure.

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

Current programming models, built for one human directing a local machine, create a mismatch when asked to host probabilistic AI systems, and patching efforts only add more external harnesses and cost. The paper proposes Software 4.0 as an autopoietic heterarchy that combines human intelligence, neural models, and a reflective symbolic substrate so that software itself becomes the active agent maintaining its integrity. Recognitive is introduced as the language and platform that realizes this shift by moving structural verification out of the probabilistic layer. This change is presented as the route to an inference regime in which AI resources address meaning and hypothesis rather than syntax and consistency. The work is explicitly a vision paper whose formal type system and empirical tests remain for later development.

Core claim

Software 4.0 is an autopoietic heterarchy of human intelligence, neural AI, and natively reflective symbolic substrate under which software changes from an inert corpus into a self-regulating metabolic network that natively verifies, modifies, and evolves its own structural integrity; Recognitive materializes the architecture and thereby offloads structural verification to a deterministic substrate so that connectionist compute can be devoted entirely to semantic exploration.

What carries the argument

The autopoietic heterarchy (a self-producing network of human, neural, and reflective symbolic elements) materialized in the Recognitive programming language and platform, which performs native structural verification.

If this is right

  • Connectionist compute is redirected entirely to deep semantic exploration and hypothesis traversal instead of simulating structural constraints.
  • The carrying cost of static code assembly declines because software maintains its own integrity.
  • Systems move beyond the legacy Software Factory model toward one in which software is a living, evolving network.
  • Inference-time scaling improves because probabilistic resources no longer bear the load of structural correctness.

Where Pith is reading between the lines

These are editorial extensions of the paper, not claims the author makes directly.

  • If the substrate works as described, software could maintain functional identity across changing requirements with minimal human or external oversight.
  • The heterarchy framing invites direct comparison with biological self-maintenance systems, raising the possibility of importing additional metabolic or evolutionary operators into software.
  • Small-scale prototypes of Recognitive could be built to measure whether the separation of concerns actually reduces the compute spent on verification tasks.
  • Adoption would shift developer effort from writing harnesses to specifying the intent that the metabolic network then realizes.

Load-bearing premise

A deterministic symbolic substrate can be constructed to handle structural verification without creating new impedance mismatches or needing the external harnesses the paper criticizes.

What would settle it

A concrete implementation of Recognitive in which code demonstrates deterministic self-verification and self-modification of its own structure while the neural component addresses only semantic content and no external harnesses are required.

read the original abstract

Dominant programming paradigms inherit an execution model optimised for a bygone era of a single human mind instructing a local machine, leaving contemporary systems burdened with historical path dependencies. When forced to host multi-dimensional, connectionist intelligence, this brittle assembly model fractures under the weight of a profound probabilistic-symbolic impedance mismatch. While contemporary Software 3.x frameworks attempt to patch the mismatch by encasing large language models (LLMs) in increasingly complicated external harnesses, this spiralling architectural complexity only compounds the carrying cost of static code assembly. To address the cause rather than the effects, this paper introduces Software 4.0 -- an autopoietic heterarchy of human intelligence, neural AI, and natively reflective symbolic substrate. Under this paradigm, software is transformed from an inert corpus to be parsed into a self-regulating metabolic network that natively verifies, modifies, and evolves its own structural integrity. We present Recognitive, the programming language and platform that materialises this architecture. By offloading the burden of structural verification to a deterministic substrate, it unlocks a superior inference-time scaling regime -- one where connectionist compute translates entirely into deep semantic exploration and hypothesis traversal rather than the ruinous computational and financial cost of simulating structural constraints probabilistically. Moving beyond the legacy 'Software Factory' mindset, we outline the theoretical foundations required to ground connectionist intent and arrive fully in the intelligence age. This is a foundational vision paper; empirical evaluation and formal specification of the type system and operational semantics are the subject of future work.

Editorial analysis

A structured set of objections, weighed in public.

Desk editor's note, referee report, simulated authors' rebuttal, and a circularity audit. Tearing a paper down is the easy half of reading it; the pith above is the substance, this is the friction.

Referee Report

2 major / 2 minor

Summary. The paper claims that dominant programming paradigms suffer from a probabilistic-symbolic impedance mismatch when hosting connectionist intelligence, leading to complex external harnesses and high carrying costs. It proposes Software 4.0 as an autopoietic heterarchy of human intelligence, neural AI, and a natively reflective symbolic substrate (materialized in the Recognitive language/platform) that transforms software into a self-regulating metabolic network capable of native verification, modification, and evolution of its structural integrity. By offloading structural tasks to the deterministic substrate, this is asserted to unlock a superior inference-time scaling regime focused on semantic exploration. The work is presented as a foundational vision paper, with empirical evaluation and formal specification of the type system and operational semantics deferred to future work.

Significance. If the architecture could be realized as described, it would offer a substantive alternative to current LLM integration approaches by providing native reflectivity and eliminating external harnesses, potentially lowering complexity and enabling more efficient use of connectionist compute. The vision addresses a real tension in software engineering between static code and dynamic AI components, but its prospective nature without supporting mechanisms limits immediate impact.

major comments (2)
  1. [Abstract] Abstract: The claim that the Recognitive substrate 'unlocks a superior inference-time scaling regime' where 'connectionist compute translates entirely into deep semantic exploration' is load-bearing for the central thesis but unsupported. The manuscript provides no type system, operational semantics, reflective primitives, pseudocode, or mechanism for deterministic verification/modification/evolution, and explicitly defers 'formal specification of the type system and operational semantics' to future work. Without this, it is impossible to evaluate whether the substrate avoids reintroducing impedance mismatches or external harnesses.
  2. [Abstract] Abstract: Benefits such as reduced carrying cost, elimination of the 'ruinous computational and financial cost of simulating structural constraints probabilistically,' and transformation to a 'self-regulating metabolic network' are defined directly in terms of the proposed autopoietic heterarchy and Recognitive substrate without external benchmarks, independent grounding, or falsifiable predictions. This renders the superiority argument circular by construction.
minor comments (2)
  1. The manuscript introduces specialized terminology (e.g., 'autopoietic heterarchy,' 'Recognitive,' 'Software 4.0') without providing definitions, references to related work in autopoiesis or reflective systems, or a glossary.
  2. As a vision paper, the absence of even high-level pseudocode or illustrative examples for the substrate's operation makes the architecture difficult to assess or build upon.

Simulated Author's Rebuttal

2 responses · 0 unresolved

We thank the referee for the careful reading and substantive comments on our vision paper. We address each major comment below, maintaining the paper's stated scope as a foundational outline of the Software 4.0 paradigm rather than a technical specification.

read point-by-point responses

  1. Referee: [Abstract] Abstract: The claim that the Recognitive substrate 'unlocks a superior inference-time scaling regime' where 'connectionist compute translates entirely into deep semantic exploration' is load-bearing for the central thesis but unsupported. The manuscript provides no type system, operational semantics, reflective primitives, pseudocode, or mechanism for deterministic verification/modification/evolution, and explicitly defers 'formal specification of the type system and operational semantics' to future work. Without this, it is impossible to evaluate whether the substrate avoids reintroducing impedance mismatches or external harnesses.

    Authors: We acknowledge the observation that the formal specification is deferred, as the abstract states. The manuscript's contribution is the identification of the probabilistic-symbolic impedance mismatch and the conceptual proposal of an autopoietic heterarchy with a natively reflective substrate. The scaling-regime claim is presented as a direct architectural consequence of moving structural verification into the deterministic layer, thereby freeing connectionist resources for semantic tasks. We agree that the absence of the type system and operational semantics prevents a full technical evaluation of mismatch avoidance at this stage. No revision is made because incorporating those elements would convert the paper from a vision statement into a different work. revision: no

  2. Referee: [Abstract] Abstract: Benefits such as reduced carrying cost, elimination of the 'ruinous computational and financial cost of simulating structural constraints probabilistically,' and transformation to a 'self-regulating metabolic network' are defined directly in terms of the proposed autopoietic heterarchy and Recognitive substrate without external benchmarks, independent grounding, or falsifiable predictions. This renders the superiority argument circular by construction.

    Authors: The benefits are derived from the documented limitations of Software 3.x approaches, which rely on external harnesses precisely because of the impedance mismatch. The argument therefore rests on the contrast between probabilistic simulation of structure and native deterministic handling, rather than on self-referential definition alone. We accept that the claims remain qualitative in the absence of benchmarks or falsifiable predictions. As the paper is explicitly positioned as a vision document, such empirical grounding is reserved for subsequent work. No revision is made. revision: no

Circularity Check

2 steps flagged

Scaling and self-regulation benefits defined by construction of the proposed autopoietic heterarchy and Recognitive substrate

specific steps
  1. self definitional [Abstract]
    "By offloading the burden of structural verification to a deterministic substrate, it unlocks a superior inference-time scaling regime -- one where connectionist compute translates entirely into deep semantic exploration and hypothesis traversal rather than the ruinous computational and financial cost of simulating structural constraints probabilistically."

    The 'superior inference-time scaling regime' is asserted as the outcome of offloading to the deterministic substrate, yet the substrate (Recognitive) is defined as the element that performs this offloading and enables native verification/modification/evolution. The claimed benefit therefore reduces directly to the introduction of the architecture, with no separate derivation, mechanism, or external validation provided.

  2. self definitional [Abstract]
    "software is transformed from an inert corpus to be parsed into a self-regulating metabolic network that natively verifies, modifies, and evolves its own structural integrity."

    The transformation into a 'self-regulating metabolic network' is presented as the result of the Software 4.0 paradigm, but the paradigm is introduced precisely as the autopoietic heterarchy that confers these properties. The claimed outcome is therefore identical to the definitional content of the new architecture.

full rationale

The paper is a foundational vision paper that explicitly defers empirical evaluation and formal specification of the type system and operational semantics to future work. Its central claims—that software becomes a self-regulating metabolic network and that offloading verification unlocks a superior inference-time scaling regime—are presented as direct consequences of introducing the new architecture. No independent mechanisms, equations, benchmarks, or external derivations are supplied; the advantages are therefore equivalent to the definitional properties of the substrate itself. This matches the self-definitional pattern with no load-bearing self-citations or fitted predictions present.

Axiom & Free-Parameter Ledger

0 free parameters · 3 axioms · 2 invented entities

The central claim rests on several unproven domain assumptions about the nature of the impedance mismatch and the feasibility of the new substrate, plus invented entities introduced without independent evidence.

axioms (3)
  • domain assumption Dominant programming paradigms inherit an execution model optimised for a bygone era of a single human mind instructing a local machine
    Invoked in the opening sentence of the abstract as the root cause of current burdens.
  • domain assumption Contemporary systems are burdened with a profound probabilistic-symbolic impedance mismatch that fractures the assembly model
    Stated as the reason current frameworks fail when hosting connectionist intelligence.
  • ad hoc to paper Offloading structural verification to a deterministic substrate unlocks superior inference-time scaling
    Central to the claimed benefit but presented without derivation or evidence.
invented entities (2)
  • Software 4.0 no independent evidence
    purpose: Autopoietic heterarchy of human intelligence, neural AI, and natively reflective symbolic substrate
    Introduced as the new paradigm that transforms software into a self-regulating metabolic network.
  • Recognitive no independent evidence
    purpose: Programming language and platform that materialises the Software 4.0 architecture
    Presented as the concrete realization enabling native verification and evolution.

pith-pipeline@v0.9.1-grok · 5797 in / 1818 out tokens · 23263 ms · 2026-06-28T13:29:17.462736+00:00 · methodology

discussion (0)

Sign in with ORCID, Apple, or X to comment. Anyone can read and Pith papers without signing in.

Reference graph

Works this paper leans on

67 extracted references · 5 linked inside Pith

  1. [1]

    Structured Fine-Tuning Enables Data-Efficient Adaptation of Code Language Models

    Mayank Agarwal, Yikang Shen, Bailin Wang, Yoon Kim, and Jie Chen. Structured Fine-Tuning Enables Data-Efficient Adaptation of Code Language Models. InInternational Conference on Learning Representations (ICLR), 2024

    2024

  2. [2]

    Bruce Alberts.Molecular Biology of the Cell (6th ed.). W.W. Norton & Company, 2015

    2015

  3. [3]

    Harvard University Press, Cambridge, MA, 1964

    Christopher Alexander.Notes on the Synthesis of Form. Harvard University Press, Cambridge, MA, 1964

    1964

  4. [4]

    Center for Environmental Structure, Berkeley, CA, 2002–2005

    Christopher Alexander.The Nature of Order: An Essay on the Art of Building and the Nature of the Universe. Center for Environmental Structure, Berkeley, CA, 2002–2005. 4 volumes. 11

    2002

  5. [5]

    The Model Context Protocol (MCP): An Open Standard for Connecting AI Assistants to Data Systems

    Anthropic. The Model Context Protocol (MCP): An Open Standard for Connecting AI Assistants to Data Systems. https://modelcontextprotocol.io, November 2024. Accessed: 2026- 05-21

    2024

  6. [6]

    Claude Code Source Code (prompts.ts File)

    Anthropic, PBC. Claude Code Source Code (prompts.ts File). Software Source Code Release, March 2026. Accessed 31 May 2026

    2026

  7. [7]

    Ross Ashby.An Introduction to Cybernetics

    W. Ross Ashby.An Introduction to Cybernetics. Chapman & Hall, London, 1956

    1956

  8. [8]

    Springer, 2015

    Marcello Barbieri.Code Biology: A New Science of Life. Springer, 2015

    2015

  9. [9]

    Mirrors: Design Principles for Meta-level Facilities of Object- Oriented Programming languages.ACM SIGPLAN Notices, 39(10):331–344, 2004

    Gilad Bracha and David Ungar. Mirrors: Design Principles for Meta-level Facilities of Object- Oriented Programming languages.ACM SIGPLAN Notices, 39(10):331–344, 2004

    2004

  10. [10]

    Brachman and Hector J

    Ronald J. Brachman and Hector J. Levesque.Knowledge Representation and Reasoning. Morgan Kaufmann, 2004

    2004

  11. [11]

    A Modification of Jerne’s Theory of Antibody Production using the Concept of Clonal Selection.Aust J Sci, 20:67–69, 1957

    Frank Macfarlane Burnet. A Modification of Jerne’s Theory of Antibody Production using the Concept of Clonal Selection.Aust J Sci, 20:67–69, 1957

    1957

  12. [12]

    State of the AI vs

    CodeRabbit. State of the AI vs. Human Code Generation Report, December 2025

    2025

  13. [13]

    Meta Large Language Model Compiler: Foundation Models of Compiler Optimization.arXiv preprint arXiv:2407.02524, 2024

    Chris Cummins, V olker Seeker, Dejan Grubisic, Baptiste Rozière, Jonas Gehring, Gabriel Synnaeve, and Hugh Leather. Meta Large Language Model Compiler: Foundation Models of Compiler Optimization.arXiv preprint arXiv:2407.02524, 2024

    arXiv 2024

  14. [14]

    From Tool Calling to Symbolic Thinking: LLMs in a Persistent Lisp Metaprogramming Loop.arXiv preprint arXiv:2506.10021, 2025

    Jordi de La Torre. From Tool Calling to Symbolic Thinking: LLMs in a Persistent Lisp Metaprogramming Loop.arXiv preprint arXiv:2506.10021, 2025

    arXiv 2025

  15. [15]

    Deacon.Incomplete Nature: How Mind Emerged from Matter

    Terrence W. Deacon.Incomplete Nature: How Mind Emerged from Matter. W. W. Norton & Company, New York, 2011

    2011

  16. [16]

    Interactional foundations for critical AI literacies

    Mark Dingemanse. Interactional foundations for critical AI literacies. Under review for a volume on critical AI. Preprint, 2026

    2026

  17. [17]

    The MIT Press, 06 2004

    Marco Dorigo and Thomas Stützle.Ant Colony Optimization. The MIT Press, 06 2004

    2004

  18. [18]

    d’Avila Garcez, Luis C

    Artur S. d’Avila Garcez, Luis C. Lamb, and Dov M. Gabbay.Neural-Symbolic Cognitive Reasoning. Springer, 2009

    2009

  19. [19]

    Clarendon Press, Oxford, 1937

    Edward Evan Evans-Pritchard.Witchcraft, Oracles and Magic Among the Azande. Clarendon Press, Oxford, 1937

    1937

  20. [20]

    Glamorous toolkit: The moldable development environment, 2026

    feenk gmbh, Switzerland. Glamorous toolkit: The moldable development environment, 2026. https://gtoolkit.com/. Accessed: 2026-05-21

    2026

  21. [21]

    Partial Evaluation of Computation Process – An Approach to a Compiler- Compiler.Systems, Computers, Controls, 2(5):45–50, 1971

    Yoshihiko Futamura. Partial Evaluation of Computation Process – An Approach to a Compiler- Compiler.Systems, Computers, Controls, 2(5):45–50, 1971

    1971

  22. [22]

    Vlissides.Design Patterns: Elements of Reusable Object-Oriented Software

    Erich Gamma, Richard Helm, Ralph Johnson, and John M. Vlissides.Design Patterns: Elements of Reusable Object-Oriented Software. Addison-Wesley, Reading, MA, 1995

    1995

  23. [23]

    PhD thesis, University of California, Berkeley, 2025

    Linyuan Gong.Advancing Large Language Models for Code Using Code-Structure-Aware Methods. PhD thesis, University of California, Berkeley, 2025

    2025

  24. [24]

    Dimensions in Program Synthesis

    Sumit Gulwani. Dimensions in Program Synthesis. InProceedings of the 12th international ACM SIGPLAN symposium on Principles and practice of declarative programming, pages 13–24, 2010

    2010

  25. [25]

    Bridging Generation and Training: A Systematic Review of Quality Issues in LLMs for Code.arXiv preprint arXiv:2605.05267, 2026

    Kaifeng He, Xiaojun Zhang, Peiliang Cai, Mingwei Liu, Yanlin Wang, Chong Wang, Kaifeng Huang, Bihuan Chen, Xin Peng, and Zibin Zheng. Bridging Generation and Training: A Systematic Review of Quality Issues in LLMs for Code.arXiv preprint arXiv:2605.05267, 2026. 12

    Pith/arXiv arXiv 2026

  26. [26]

    Cybernetics and Second-Order Cybernetics

    Francis Heylighen and Cliff Joslyn. Cybernetics and Second-Order Cybernetics. In Robert A. Meyers, editor,Encyclopedia of Physical Science and Technology (Third Edition), pages 155–

  27. [27]

    Academic Press, New York, third edition edition, 2003

    2003

  28. [28]

    Coding From 1849 to 2022: a Guide to The Timeline of Programming Languages

    Severine Hierso. Coding From 1849 to 2022: a Guide to The Timeline of Programming Languages. website, 2022

    2022

  29. [29]

    Indiana University Press, Bloomington, IN, 1997

    Jesper Hoffmeyer.Signs of Meaning in the Universe. Indiana University Press, Bloomington, IN, 1997

    1997

  30. [30]

    Hofstadter.Gödel, Escher, Bach: An Eternal Golden Braid

    Douglas R. Hofstadter.Gödel, Escher, Bach: An Eternal Golden Braid. Basic Books, New York, NY , 1979

    1979

  31. [31]

    Holland.Adaptation in Natural and Artificial Systems: An Introductory Analysis with Applications to Biology, Control, and Artificial Intelligence

    John H. Holland.Adaptation in Natural and Artificial Systems: An Introductory Analysis with Applications to Biology, Control, and Artificial Intelligence. University of Michigan Press, 1975

    1975

  32. [32]

    Neural Networks and Physical Systems with Emergent Collective Computational Abilities.Proceedings of the National Academy of Sciences, 79(8):2554–2558, 1982

    J J Hopfield. Neural Networks and Physical Systems with Emergent Collective Computational Abilities.Proceedings of the National Academy of Sciences, 79(8):2554–2558, 1982

    1982

  33. [33]

    Teaching Language Models to Think in Code

    Hyeon Hwang, Jiwoo Lee, and Jaewoo Kang. Teaching Language Models to Think in Code. arXiv preprint arXiv:2605.07237, 2026

    Pith/arXiv arXiv 2026

  34. [34]

    The Hot Mess of AI: How Does Misalignment Scale With Model Intelligence and Task Complexity?, 2026

    Alexander Hägele, Aryo Pradipta Gema, Henry Sleight, Ethan Perez, and Jascha Sohl-Dickstein. The Hot Mess of AI: How Does Misalignment Scale With Model Intelligence and Task Complexity?, 2026

    2026

  35. [35]

    Jacobson

    V . Jacobson. Congestion avoidance and control.ACM SIGCOMM Computer Communication Review, 18(4):314–329, August 1988

    1988

  36. [36]

    A Survey on Large Language Models for Code Generation.ACM Transactions on Software Engineering and Methodology, 35(2):1–72, January 2026

    Juyong Jiang, Fan Wang, Jiasi Shen, Sungju Kim, and Sunghun Kim. A Survey on Large Language Models for Code Generation.ACM Transactions on Software Engineering and Methodology, 35(2):1–72, January 2026

    2026

  37. [37]

    Samuel G. B. Johnson, Amir-Hossein Karimi, Yoshua Bengio, et al. Imagining and Building Wise Machines: The Centrality of AI Metacognition.Trends in Cognitive Sciences, 2025

    2025

  38. [38]

    Software 2.0

    Andrej Karpathy. Software 2.0. Medium, 2017. https://karpathy.medium.com/software-2-0- a64152b37c35

    2017

  39. [39]

    Software in the era of AI

    Andrej Karpathy. Software in the era of AI. YouTube, 2025. https://youtu.be/LCEmiRjPEtQ

    2025

  40. [40]

    DSPy: Compiling Declarative Language Model Calls into Self-Improving Pipelines.arXiv preprint arXiv:2310.03714, 2023

    Omar Khattab, Arnav Singhvi, Paridhi Maheshwari, Zhiyuan Zhang, Keshav Santhanam, Sri Vardhamanan, Saiful Haq, Ashutosh Sharma, Thomas T Joshi, Hanna Moazam, et al. DSPy: Compiling Declarative Language Model Calls into Self-Improving Pipelines.arXiv preprint arXiv:2310.03714, 2023

    Pith/arXiv arXiv 2023

  41. [41]

    Kirschner and John C

    Marc W. Kirschner and John C. Gerhart.The Plausibility of Life: Resolving Darwin’s Dilemma. Yale University Press, 2005

    2005

  42. [42]

    Biological robustness.Nature Reviews Genetics, 5(11):826–837, 2004

    Hiroaki Kitano. Biological robustness.Nature Reviews Genetics, 5(11):826–837, 2004

    2004

  43. [43]

    Algorithm = Logic + Control.Communications of the ACM, 22(7):424–436, 1979

    Robert Kowalski. Algorithm = Logic + Control.Communications of the ACM, 22(7):424–436, 1979

    1979

  44. [44]

    Lampinen and James L

    Andrew K. Lampinen and James L. McClelland. Transforming task representations to perform novel tasks.Proceedings of the National Academy of Sciences, 117(52):32970–32981, December 2020

    2020

  45. [45]

    Concepts and experiments in computational reflection.ACM Sigplan Notices, 22(12):147–155, 1987

    Pattie Maes. Concepts and experiments in computational reflection.ACM Sigplan Notices, 22(12):147–155, 1987

    1987

  46. [46]

    Tenenbaum, and Jiajun Wu

    Jiayuan Mao, Chuang Gan, Pushmeet Kohli, Joshua B. Tenenbaum, and Jiajun Wu. The Neuro-Symbolic Concept Learner: Interpreting Scenes, Words, and Sentences From Natural Supervision. InInternational Conference on Learning Representations (ICLR), 2019. 13

    2019

  47. [47]

    Maturana and Francisco J

    Humberto R. Maturana and Francisco J. Varela.Autopoiesis and Cognition: The Realization of the Living. D. Reidel Publishing Company, 1980

    1980

  48. [48]

    A Survey of Context Engineering for Large Language Models.arXiv preprint arXiv:2507.13334, 2025

    Lingrui Mei, Jiayu Yao, Yuyao Ge, Yiwei Wang, Baolong Bi, Yujun Cai, Jiazhi Liu, Mingyu Li, Zhong-Zhi Li, Duzhen Zhang, et al. A Survey of Context Engineering for Large Language Models.arXiv preprint arXiv:2507.13334, 2025

    Pith/arXiv arXiv 2025

  49. [49]

    Design Patterns in Dynamic Languages

    Peter Norvig. Design Patterns in Dynamic Languages. Slides presented at Object World, Boston, MA, May 1996

    1996

  50. [50]

    The Split

    Paul Pangaro and Jill Fain Lehman. Symposium #1 “The Split”: Unspoken Continuities — Re-Braiding Project, March 2026. Online symposium presentation. Event poster and symposium announcement available at: https://www.cs.cmu.edu/afs/cs.cmu.edu/Web/Posters/UnspokenContinuities-Re- BraidingCyberneticsAndAI.pdf

    2026

  51. [51]

    Patterson and David R

    David A. Patterson and David R. Ditzel. The Case for the Reduced Instruction Set Computer. SIGARCH Comput. Archit. News, 8(6):25–33, October 1980

    1980

  52. [52]

    Cold Spring Harbor Laboratory Press Cold Spring Harbor, NY , 2002

    Mark Ptashne and Alexander Gann.Genes & Signals, volume 402. Cold Spring Harbor Laboratory Press Cold Spring Harbor, NY , 2002

    2002

  53. [53]

    A Multi-Language Perspective on the Robustness of LLM Code Generation.arXiv preprint arXiv:2504.19108, 2026

    Fazle Rabbi, Zishuo Ding, and Jinqiu Yang. A Multi-Language Perspective on the Robustness of LLM Code Generation.arXiv preprint arXiv:2504.19108, 2026

    Pith/arXiv arXiv 2026

  54. [54]

    Reynolds

    Craig W. Reynolds. Flocks, herds and schools: A distributed behavioral model.SIGGRAPH Comput. Graph., 21(4):25–34, August 1987

    1987

  55. [55]

    Pergamon Press, Oxford, 1985

    Robert Rosen.Anticipatory Systems: Philosophical, Mathematical and Methodological Foundations. Pergamon Press, Oxford, 1985

    1985

  56. [56]

    Columbia University Press, New York, 1991

    Robert Rosen.Life Itself: A Comprehensive Inquiry into the Nature, Origin, and Fabrication of Life. Columbia University Press, New York, 1991

    1991

  57. [57]

    Rumelhart, Geoffrey E

    David E. Rumelhart, Geoffrey E. Hinton, and Ronald J. Williams. Learning representations by back-propagating errors.Nature, 323(6088):533–536, 1986

    1986

  58. [58]

    Reflection and Semantics in a Procedural Language

    Brian Cantwell Smith. Reflection and Semantics in a Procedural Language. Technical report, MIT Laboratory for Computer Science, 1982

    1982

  59. [59]

    The MIT Press, 10 2019

    Brian Cantwell Smith.The Promise of Artificial Intelligence: Reckoning and Judgment. The MIT Press, 10 2019

    2019

  60. [60]

    Smullyan.Gödel’s Incompleteness Theorems

    Raymond M. Smullyan.Gödel’s Incompleteness Theorems. Number 19 in Oxford Logic Guides. Oxford University Press, 1992

    1992

  61. [61]

    Allostasis: A model of predictive regulation.Physiology & behavior, 106(1):5– 15, 2012

    Peter Sterling. Allostasis: A model of predictive regulation.Physiology & behavior, 106(1):5– 15, 2012

    2012

  62. [62]

    Critical Vulnerability in Claude Code Emerges Days After Source Leak

    Kevin Townsend. Critical Vulnerability in Claude Code Emerges Days After Source Leak. SecurityWeek, March 2026. Accessed 31 May 2026

    2026

  63. [63]

    Varela, Evan Thompson, and Eleanor Rosch.The Embodied Mind: Cognitive Science and Human Experience

    Francisco J. Varela, Evan Thompson, and Eleanor Rosch.The Embodied Mind: Cognitive Science and Human Experience. The MIT Press, 01 2017

    2017

  64. [64]

    Springer, 2003

    Heinz V on Foerster.Understanding Understanding: Essays on Cybernetics and Cognition. Springer, 2003

    2003

  65. [65]

    Princeton Studies in Complexity

    Andreas Wagner.Robustness and Evolvability in Living Systems. Princeton Studies in Complexity. Princeton University Press, Princeton, NJ, 2005

    2005

  66. [66]

    On Code-Induced Reasoning in LLMs.arXiv preprint arXiv:2509.21499, 2025

    Abdul Waheed, Zhen Wu, Carolyn Rosé, and Daphne Ippolito. On Code-Induced Reasoning in LLMs.arXiv preprint arXiv:2509.21499, 2025

    arXiv 2025

  67. [67]

    Wolpert, Christopher R

    Daniel M. Wolpert, Christopher R. Miall, and Mitsuo Kawato. Internal Models in the Cerebellum.Trends in Cognitive Sciences, 2(9):338–47, 1998. 14

    1998