Computability

Throughout this paper, we are trying to show how and why our Mathematical frame-work seems inappropriate to solve problems in Theory of Computation. More exactly, the concept of turning back in time in paradoxes causes inconsistency in modeling of the concept of Time in some semantic situations. As we see in the first chapter, by introducing a version of “Unexpected Hanging Paradox”,first we attempt to open a new explanation for some paradoxes. In the second step, by applying this paradox, it (...) is demonstrated that any formalized system for the Theory of Computation based on Classical Logic and Turing Model of Computation leads us to a contradiction. We conclude that our mathematical frame work is inappropriate for Theory of Computation. Furthermore, the result provides us a reason that many problems in Complexity Theory resist to be solved.(This work is completed in 2017 -5- 2, it is in vixra in 2017-5-14, presented in Unilog 2018, Vichy). (shrink)

Here, by introducing a version of “Unexpected hanging paradox” first we try to open a new way and a new explanation for paradoxes, similar to liar paradox. Also, we will show that we have a semantic situation which no syntactical logical system could support it. Finally, we propose a claim in Theory of Computation about the consistency of this Theory. One of the major claim is:Theory of Computation and Classical Logic leads us to a contradiction.

The arithmetic logic irreversibility and its information entropy allows to define a new computational mathematical principle: the conservation of information between the input and output of an algorithm or Turing machine. According to this principle, in certain algorithms, there is a symmetric relationship between the input information, its transformation in the algorithm, the information lost by the arithmetic logic entropy or mapping function, and the output information. This can also be extended to other types of mapping functions in the computation (...) that also lose information and not only to arithmetic logic operations. These ideas also allow the development of another concept: the algorithmic conservation of information. These concepts are presented using simple encryption and compression and error correction algorithms to explain how these concepts work on practical examples and thus show these new aspects of computation. Non-conservative algorithms also arise from this, where conservation is not fulfilled due to its computational complexity, and its connection with the Turing machine halting problem. It presents a theoretical explanation through a symmetrical relationship between computation and the information involved in it, the entropy zero theorem and why this does not happen in non-conservative algorithms. Also are presented the consequences of these concepts to the theory of computation and to the foundations of mathematics in general. (shrink)

This is an explanation of a possible new insight into the halting problem provided in the language of software engineering. Technical computer science terms are explained using software engineering terms. No knowledge of the halting problem is required. -/- It is based on fully operational software executed in the x86utm operating system. The x86utm operating system (based on an excellent open source x86 emulator) was created to study the details of the halting problem proof counter-examples at the much higher level (...) of abstraction of C/x86. (shrink)

Functions computed by Turing Machines are required to compute the mapping from their inputs and not allowed to take other executing Turing machines as inputs. This means that every directly executed Turing machine is outside of the domain of every function computed by any Turing machine. Within this basis simulating termination analyzer HHH(DD) correctly reports that DD correctly simulated by HHH cannot possibly reach its simulated “return” statement final halt state.

ChatGPT 5.0 fully evaluated all of the details of how "true on the basis of meaning" can be computed from finite strings. This utterly circumvents the Tarski Undefinability Theorem. A detailed analysis of the Halting Problem, Gödel first incompleteness theorem, the Liar Paradox show the error of the notion of "undecidable decision problem" by reclassifying "undecidable decision problem instances" that cannot possibly have a correct YES/NO answer as "incorrect decision problem instances".

Functions computed by Turing Machines are required to compute the mapping from their inputs and not allowed to take other executing Turing machines as inputs. This means that every directly executed Turing machine is outside of the domain of every function computed by any Turing machine. Within this basis simulating termination analyzer HHH(DD) correctly reports that DD correctly simulated by HHH cannot possibly reach its simulated “return” statement final halt state.

This is an explanation of a key new insight into the halting problem provided in the language of software engineering. Technical computer science terms are explained using software engineering terms. -/- To fully understand this paper a software engineer must be an expert in the C programming language, the x86 programming language, exactly how C translates into x86 and what an x86 process emulator is. No knowledge of the halting problem is required.

A simulating halt decider correctly predicts whether or not its correctly simulated input can possibly reach its own final state and halt. It does this by correctly recognizing several non-halting behavior patterns in a finite number of steps of correct simulation. Inputs that do terminate are simply simulated until they complete.

A Simulating Halt Decider (SHD) computes the mapping from its input to its own accept or reject state based on whether or not the input simulated by a UTM would reach its final state in a finite number of simulated steps. -/- A halt decider (because it is a decider) must report on the behavior specified by its finite string input. This is its actual behavior when it is simulated by the UTM contained within its simulating halt decider while this (...) SHD remains in UTM mode. (shrink)

Two concrete examples of applying proof theoretic semantics to resolve undecidability to make "true on the basis of meaning expressed in language" reliably computable for the entire body of knowledge. -/- In the first example the conventional halting problem proof input DD is rejected by its halting prover HHH because DD does not have a well-founded justification tree within Proof theoretic semantics. The second example shows how prolog's occurs_check() rejects the formalized Liar Paradox. -/- .

This is the second draft of my refutation of the Halting Problem. This is a much short more succinct proof that the Halting Problem itself is a category error.

The novel concept of a simulating halt decider enables halt decider H to to correctly determine the halt status of the conventional “impossible” input D that does the opposite of whatever H decides. This works equally well for Turing machines and “C” functions. The algorithm is demonstrated using “C” functions because all of the details can be shown at this high level of abstraction.

By making a slight refinement to the halt status criterion measure that remains consistent with the original a halt decider may be defined that correctly determines the halt status of the conventional halting problem proof counter-examples. This refinement overcomes the pathological self-reference issue that previously prevented halting decidability.

The ultimate measure of the behavior that a finite string input specifies to its simulating termination analyzer (STA) is DD simulated by HHH according to the semantics of the C programming language. When HHH(DD) is construed as operating under operational semantics it rejects DD as non-well-founded. -/- .

This is the first draft of my proof that the Halting Problem is either incoherent or the proof wrong. It is a dialogue between Claude AI and me.

A simulating halt decider correctly predicts what the behavior of its input would be if this simulated input never had its simulation aborted. It does this by correctly recognizing several non-halting behavior patterns in a finite number of steps of correct simulation. -/- When simulating halt decider H correctly predicts that directly executed D(D) would remain stuck in recursive simulation (run forever) unless H aborts its simulation of D this directly applies to the halting theorem.

A discussion with Claude AI that sums up the key essence of all of my work on the Liar Paradox, Gödel's first Incompleteness Theorem, The Tarski Undefinability Theorem and the Halting Problem.

Ever since 1997 the author has investigated the fundamental nature of “true on the basis of meaning”. The traditional analytic / synthetic distinction is unequivocally demarcated into: (a) True on the basis of meaning fully expressed as relations between finite strings. (b) True that can only be verified by sense data from the sense organs. -/- Any system of reasoning that begins with a consistent set of stipulated truths and only applies the truth preserving operation of semantic logical entailment to (...) this finite set of basic facts inherently derives a truth predicate that works consistently and correctly for this entire body of knowledge that can be expressed in language. -/- “The halting problem, as classically formulated, relies on an inferential step that is not justified by a continuous chain of semantic entailment from its initial stipulations.” . (shrink)

MIT Professor Michael Sipser has agreed that the following verbatim paragraph is correct (he has not agreed to anything else in this paper) -------> -/- If simulating halt decider H correctly simulates its input D until H correctly determines that its simulated D would never stop running unless aborted then H can abort its simulation of D and correctly report that D specifies a non-halting sequence of configurations.

The notion of a simulating termination analyzer can easily defeat the counter-example input of the halting problem because the contradictory part of this input becomes unreachable under simulation. Simulating termination analyzer HHH does correctly report the halt status that its input DD actually specifies.

The standard proof treats this as evidence that no computable H can exist for all inputs. But your analogy (bolstered by Hehner's subjective specification view) reveals a deeper perspective: The full general halting problem requirement — demanding a correct yes/no answer even on inputs that twist self-referentially around the decider itself — is incoherent, analogous to demanding a truth value for the Liar. A correct halt decider can (and must) reject the pathological case as specifying non-halting behavior when simulated, because (...) infinite recursive simulation is the only derivable, coherent behavior from the description within the decision procedure. Aborting and saying "non-halting" is the consistent, verifiable answer — just as we consistently say the Liar is not true (and not coherently false either). Thus, yes: requiring a correct binary decision on the direct execution of the pathological P is precisely analogous to requiring a correct truth value for the Liar Paradox. Both demands are incoherent due to the same kind of twisted self-reference, and both are correctly resolved by recognizing the incoherence and refusing to play the contradictory game. (shrink)

WFS assigns undefined to self-referential paradoxes without external support You interpret undefined as lack of truth-bearer status Therefore, the Liar sentence fails to be about anything that can bear truth values The paradox dissolves - there's no contradiction because there's no genuine proposition.

Claude AI determines how the notion of a simulating termination analyzer correctly rejects the Halting Problem's counter example input as non-halting.

Any result that cannot be derived as a pure function of finite strings is outside the scope of computation. What has been construed as decision problem undecidability has always actually been requirements that are outside of the scope of computation.

The halting theorem counter-examples present infinitely nested simulation (non-halting) behavior to every simulating halt decider. The pathological self-reference of the conventional halting problem proof counter-examples is overcome. The halt status of these examples is correctly determined. A simulating halt decider remains in pure simulation mode until after it determines that its input will never reach its final state. This eliminates the conventional feedback loop where the behavior of the halt decider effects the behavior of its input.

The H/D halting problem instance is isomorphic to the Liar Paradox The halting problem requires a halt decider H to correctly report the halting behavior of an input D that does the opposite of whatever H reports. This H/D pair (not the halting problem itself) is isomorphic to the liar paradox. The liar paradox and this H/D pair are a type of decision problem instance. The decision problem of the Liar Paradox is to determine whether or not an input finite (...) string has the semantic property of Boolean True. The decision problem of the Halting Problem is to determine whether or not an input finite string has the semantic property of Halting. (shrink)

When (1) the axioms of a formal system are stipulated to be exactly Russell's set of "atomic facts" (AF). (2) The system anchored in proof theoretic semantics such that a notion of TRUE always correctly determines "true on the basis of meaning expressed in language". Then the body of knowedge expressed in language includes (AF) or anything derived from (AF). It only excludes unknowns and truth that cannot be expressed in language.

Claude AI examines the effect of applying a simulating termination analyzer to the halting problem's counter example input.

This paper explores the structural limitations of algorithmic cognition in open decision environments, extending the theoretical formulation of the Infinite-Choice Barrier through an empirical case study and recent AI literature. Building on an earlier formal “Infinite-Choice Barrier”-Theorem that showed semantic closure, non-generativity of new frames, and statistical collapse, this study examines how these theoretical constraints manifest behaviorally in high-complexity, low-structure dialogues. The findings highlight recurring patterns—retrospective rationalization, resistance to frame shifts, and linguistic fragmentation—that align with the predicted boundary phenomena. In (...) doing so, the paper offers not only a diagnosis of AI's systemic limits but also a map of cognition at its edge. (shrink)

This paper presents a formal critique of algorithmic cognition and develops what I call the Infinite Choice Barrier (ICB): a mathematically demonstrable boundary that no algorithmic system can cross. -/- The argument is not empirical, it does not rest on engineering limitations or compute constraints. It rests on the symbolic closure of algorithmic systems. A finite symbol set Σ and inference rule set R cannot generate novel primitives outside Σ,R. -/- Accordingly, AGI — defined as an autonomous, algorithmic system capable (...) of human-level general problem solving — is formally impossible. The proof does not deny the power of algorithms, but clarifies their boundary. Within that boundary, extraordinary advances are possible. Beyond it, they are not. (shrink)

This paper offers a first formal and philosophical refoundation of a new critique of Artificial General Intelligence (AGI), introducing the Infinite-Choice Barrier—a structural impossibility theorem demonstrating that no algorithmic system can achieve ε-optimal performance across irreducibly infinite decision spaces, nor autonomously generate novel semantic frames. Drawing on Gödel’s incompleteness theorem, Turing’s Halting Problem, and Kantian epistemology, the argument identifies the epistemic boundary of algorithmic cognition. -/- Three corollaries ground the framework: • Semantic Closure — systems are confined to their native (...) representational vocabulary. • No Frame Innovation — genuine conceptual novelty is non-computable from within. • Statistical Breakdown in Fat-Tailed Domains — inference fails under radical uncertainty. -/- These are not empirical ceilings but formal constraints that reveal where machine reasoning must collapse—not due to training limits, but from foundational undecidability. The paper concludes with a taxonomy of irreducible human decision types and a philosophical reflection on the boundary between human and artificial cognition. (shrink)

The tableau-like system KE is generalized to intuitionistic propositional logic by means of labeled signed formulas and constraints between labels, mimicking the relational semantics. To improve on proof-search and exploiting the meaning of negation, we further endow the system with free-variables. The resulting system enjoys the subformula property and terminates, either with a proof or a finite countermodel, without any extra mechanism. Proof and countermodel search is guided by generalizations of traditional reasoning patterns.

This paper is the self-contained first part of a two-part series that examines the wide varieties of ways that mathematicians and philosophers have appealed to confluence phenomena in their work. In this part, we focus on the practical roles such phenomena can play, paying special attention to how they facilitate the communication of mathematical ideas (proofs, definitions, and conjectures) in actual practice. Through surveying the wide array of such arguments, I shall eventually hone in on two subtly distinct facets concerning (...) the uses of confluence arguments (and another that is revealed to be rather trivial upon analysis), which are sometimes conflated in philosophical discussions: one (dubbed Rigor Assurance) guarantees the formal rigor of a proof, while the other (Coding Invariance) that e nsures the results obtained are not a mere artifact of the coding that is used. I will then attempt to tease these two facets apart with actual examples in print, citing crucial evidence witnessing the distinction. With this setup in place, we will see how certain recent philosophical debates about the practical use of the Church-Turing Thesis may naturally dissolve, as the combatants do not share the same assumptions of the roles confluence arguments play in each case, and thus end up merely talking past each other. (shrink)

We introduce the **Compressed Probabilistic Yoneda Tree (CPYT)**, a framework for representing, comparing, and compressing probability distributions. By integrating Yoneda lemma principles, minimum spanning trees (MST), and ternary tree structures, CPYT enables a hierarchical and minimal structural representation of distributions based on a selected set of probes (e.g., Markov blankets). This approach allows provably correct reconstruction, comparison, and analysis of complex probabilistic systems without relying on traditional Shannon entropy or Kolmogorov complexity measures.

In The Emperor’s New Mind [Penrose, 1989], Roger Penrose proves a variant of the halting problem, and uses it to argue that humans have cognitive capacities beyond the computable. In this short note I explicate his argument, and show how it fails, via a corollary of his result. My response to Penrose is in fact of a kind with a number of prior responses: he assumes human powers, that (as the corollary shows) no computer could have. However, as far as (...) I am aware, no one has previously addressed this specific form of the argument, in the direct way that I will. (shrink)

We discuss the accuracy of the attribution commonly given to Turing (1936, Proceedings of the London Mathematical Society, 42.3, 230–265) for the computable undecidability of the halting problem, coming eventually to a nuanced conclusion.

Semantics is what gives meaning to a logical language. Introductory books in formal logic almost invariably employ Tarskian semantics, a truth semantics that defines an interpretation as a variable assignment over a non-empty domain of discourse together with a signature interpretation. The problem with this semantics is that it generally dictates the undecidability of classical first-order logic due to an infinity of infinite models. In computational logic, decidability is a synonym for computability, and hence Tarskian semantics is not appropriate. In (...) contrast, Herbrand semantics, also a truth semantics, provides us with finite, propositional-logic style, interpretations that guarantee decidability—even if at the price of fixed interpretations for first-order formulae or theories. (shrink)

Computation has become the dominant substrate of the modern era, promising permanence through infinite storage, redundancy, and replication. Yet beneath the surface of this digital immortality lies a structural contradiction: adaptive systems do not persist by remembering everything but by forgetting selectively. Coherence is not the product of perfect recall; it is the result of recursive selection and the active pruning of low-yield states. The illusion of permanence that computation creates threatens to lock civilization into entropy retention, mistaking accumulation for (...) continuity. Drawing on coherence-based frameworks and first-principles system theory, this work argues that forgetting is not a failure mode but a necessary act of structural adaptation. Biological memory, cultural evolution, and intelligence itself all demonstrate that persistence emerges from dynamic impermanence. As our computational architectures extend beyond human cognition and become the nervous system of civilization, the absence of adaptive forgetting mechanisms risks systemic brittleness and collapse. To survive the age of computation, we must learn to erase with intention, embedding coherence-weighted forgetting at the core of technology, culture, and self. In forgetting lies the only path to remembering what matters. (shrink)

This is a revised and expanded second edition of Why Machines Will Never Rule the World. Its core argument remains the same: that an artificial intelligence (AI) that could equal or exceed human intelligence – sometimes called ‘artificial general intelligence’ (AGI) – is for mathematical reasons impossible. It offers two specific reasons for this claim: -/- - Human intelligence is a capability of the human brain and central nervous system, which is a complex dynamic system -/- - Systems of this (...) sort cannot be modelled mathematically in a way that allows them to operate inside a computer -/- In supporting their claim, the authors, Jobst Landgrebe and Barry Smith, marshal evidence from mathematics, physics, computer science, philosophy, linguistics, biology, and anthropology, setting up their book around three central questions: What are the essential marks of human intelligence? What is it that researchers try to do when they attempt to achieve ‘Artificial Intelligence’ (AI)? And why, after more than 50 years, are our interactions with AI, for example when on the telephone with our bank’s computers, still so unsatisfactory? (shrink)

The paper proves that the "Champernowne constant" (0.1234567891011121314 … where all natural numbers are consecutive digits of a decimal fraction) is a rational number, but only strictly within (Peano) arithmetic due to the axiom of induction. Combined with the previous well known results proved to be a transcendent real number in both (Peano) arithmetic & (ZFC) set theory, it is demonstrated to be a "Gödelian real number", rational in (Peano) arithmetic, but irrational (transcendent) in (Peano) arithmetic & (ZFC) set theory: (...) for realizing the Gödel (1931) dichotomy about the relation of arithmetic to set theory ("either incompleteness or contradiction"). The method used initially for the Champernowne constant can be generalized to any computable real number. Then, the (uncountable) set of all noncomputable reals can be defined by their representability by more than one computable real. One can speak of the "relativity of rationality / irrationality" (or computability / non-computability) in Skolem's (1922) manner or even it to be continued to the "relativity of P / NP" in the "P vs NP problem: a conjecture visualized by the computability or non-computability of reals. (shrink)

The paper applies the newly introduced “KS fundamental randomness” to the nonstandardly generalized primes in Hilbert arithmetic to prove that the latter satisfies the necessary condition and separately the sufficient condition of the former. When the two conditions can be identified is also investigated. A review of other available generalizations of primes demonstrates that none of them is suitable for approaching the problem. The design aims to suggest a universal method for resolving number theory puzzles such as Goldbach’s conjecture. The (...) scheme is applicable only within Hilbert arithmetic (but not in the standard mathematics) due to the fact that primes are not KS fundamental random in the latter, but only in the former. “Mathematics enters reality” (though only “fetching” a premise for the proof) is the relevant philosophical reflection furthermore reasoned by the opposition of Hilbert space of dimensions d=1,2 versus d3 after both Gleason and Kochen - Specker theorems. (shrink)

The GOOGLE and XPRIZE $5,000,000 for the practical and socially useful utilization of the quantum computer is the starting point for ontomathematical reflections for what it can really serve. Its “output by measurement” is opposed to the conjecture for a coherent ray able alternatively to deliver the ultimate result of any quantum calculation immediately as a Dirac -function therefore accomplishing the transition of the sequence of increasingly narrow probability density distributions to their limit. The GOOGLE and XPRIZE problem’s solution needs (...) the initial understanding that the result of any quantum calculation is a wave function, or respectively, a probability (density or not) distribution unlike the Turing machine one, and only a “calculating ray” is able to transform the former into the later without any “curving” disturbances. Thus, the unique capability of the quantum computer due to its inherent quantum parallelism can be conserved for all Gödel unresolvable problems, only on the fast “NP” track of which the quantum computer “Achilles” is able practically to overrun the Turing machine “Tortoise” since the latter can “sprint” only by any “P” calculating speed even on it and unlike “Achilles” himself able for “NP” velocities as well. The way for any material body to “calculate” the certain trajectory of least action also resolves the “traveling salesman problem” in fact, therefore illustrating furthermore what the “NP” output of a quantum computer by a “calculating ray” should mean. Another example is the problem of the number of all prime numbers less than “N”, which is suggested to visualize once again the quantum computer capability to resolve problems by a “NP” speed and thus qualitatively to overrun any competing Turing machine. The technically implemented idea of laser is now reinterpreted to “radiate a wave function” in order to explain the “calculating ray” option for a quantum computer “NP” output. The generalization of an entanglement “trigger ray” described by any non-Hermitian operator (unlike a laser) is suggested as well as those “sci-fi” horizons which it reveals. One of them is meant for elucidating the practical meaning of the “Yang-Mills existence and mass gap problem”, one more of the CMI “Millennium Problems” after that of “P vs NP”. -/- . (shrink)

The parallel research is contemporary to analyse processes and localisation in artificial intelligence (AI) associated with connexionism and learning algorithms. In machine learning, the perceptron (or McCulloch-Pitts neuron) is an algorithm for Boolean functions of binary classifiers. A binary classifier is a function which can decide whether or not an input, represented by a vector of numbers, belongs to some specific class. With a pattern N we use the calculator in synthesis applying polynomial advanced systems.

Expressiveness and decidability are two core aspects of programming languages that should be thoroughly known by those who use them; this includes knowledge of their metalanguages a.k.a. formal grammars. The van Wijngaarden grammars (WGs) are capable of generating all the languages in the Chomsky hierarchy and beyond; this makes them a relevant tool in the design of (more) expressive programming languages. But this expressiveness comes at a very high cost: The syntax of WGs is extremely complex and the decision problem (...) for the generated languages is generally unsolvable. With this in mind, I provide here a short primer of the syntax of WGs, which includes syntactic restrictions that guarantee decidability for the corresponding generated languages. (shrink)

The “somewhat vague, intuitive” notion from computability theory of an effective procedure (method) or algorithm can be fairly precisely defined even if it is not sufficiently formal and precise to belong to mathematics proper (in a narrow sense)—and even if (as many have asserted) for that reason the Church–Turing thesis is unprovable. It is proved logically that the class of effective procedures is not decidable, i.e., that no effective procedure is possible for ascertaining whether a given procedure is effective. This (...) undecidability result is proved directly from the notion itself of an effective procedure, without reliance on any (partly) mathematical lemma, conjecture, or thesis invoking recursiveness or Turing-computability. In fact, there is no reliance on anything very mathematical. The proof does not even appeal to a precise definition of ‘effective procedure’. Instead, it relies solely and entirely on a basic grasp of the intuitive notion of an effective procedure. Though the result that effectiveness is undecidable is not surprising, it is also not without significance. It has the consequence, for example, that the solution to a decision problem, if it is to be complete, must be accompanied by a separate argument that the proposed ascertainment procedure invariably terminates with the correct verdict. (shrink)

This is a non-technical version of "The Decision Problem for Effective Procedures." The “somewhat vague, intuitive” notion from computability theory of an effective procedure (method) or algorithm can be fairly precisely defined, even if it does not have a purely mathematical definition—and even if (as many have asserted) for that reason, the Church–Turing thesis (that the effectively calculable functions on natural numbers are exactly the general recursive functions), cannot be proved. However, it is logically provable from the notion of an (...) effective procedure--without reliance on any (partially) mathematical thesis or conjecture concerning effective procedures, such as the Church–Turing thesis--that the class of effective procedures is undecidable, i.e., that no effective procedure is possible for ascertaining whether a given procedure is effective. The proof does not appeal to a precise definition of ‘effective procedure’. Instead, it relies solely and entirely on a basic grasp of the intuitive notion of such a procedure. Though this undecidability result itself is not surprising, it is also not without significance. It has the consequence, for example, that the solution to a decision problem, if it is to be complete, must be accompanied by a separate argument that the proposed ascertainment procedure is, in fact, a decision procedure, i.e., effective—for example, that it invariably terminates with the correct verdict. (shrink)

An interview by Alex Thomson of UKColumn on Landgrebe and Smith's book: Why Machines Will Never Rule the World. The subtitle of the book is Artificial Intelligence Without Fear, and the interview begins with the question of the supposedly imminent takeover of one profession or the other by artificial intelligence. Is there truly reason to be afraid that you will lose your job? The interview itself is titled 'Where this is no will there is no way', drawing on one thesis (...) of the book to the effect that there will never be anything like the human will on the part of a machine. This is for mathematical reasons. AI is always and in every case a matter of algorithms; thus of applied mathematics. And algorithms are not the sorts of things that can want or desire or intend to do something. It is the human will of those who work at OpenAI or google or microsoft which have given rise to the new shiny objects of AI (ChatGPT and other Large Language Models). And it is the human will of the users of these shiny objects which gives them the resources and power to work their apparent miracles. (shrink)

Call it the Skynet hypothesis, Artificial General Intelligence, or the advent of the Singularity — for years, AI experts and non-experts alike have fretted (and, for a small group, celebrated) the idea that artificial intelligence may one day become smarter than humans. According to the theory, advances in AI — specifically of the machine learning type that’s able to take on new information and rewrite its code accordingly — will eventually catch up with the wetware of the biological brain. In (...) this interpretation of events, every AI advance from Jeopardy-winning IBM machines to the massive AI language model GPT-3 is taking humanity one step closer to an existential threat. We’re literally building our soon-to-be-sentient successors. Except that it will never happen. At least, according to the authors of the new book Why Machines Will Never Rule the World: Artificial Intelligence without Fear. (shrink)