Self Aware NetworksComparison Encyclopedia

SAN / Neural Lace → Joscha Bach and coauthors

Blumberg’s earlier source record supports ten bounded comparisons with two later Joscha Bach/coauthor versions. Each pair below shows the shared operation or framing and the remaining difference. This selection is not a completed Bach audit.

What this selection records

One transcript, five examined framings, two later document versions, ten comparison links and 35 author-navigation edges. Overlapping passages are retained. Target repetition counts and independent inventions are not established by these totals.

The exact transcript wording is fixed in a June 21, 2025 Git record and is currently publicly retrievable with matching bytes. It attributes the recording to April 2017. This dossier does not treat that attribution as a verified original upload date or audio check. The 2022 Git object at the same filename is a different interview.

The three content lenses are separate below. The program-level longitudinal result remains LSM-UNKNOWN: the wider 21-version/occurrence history and 66 inherited semantic joins are still open. No dependence probability or statistical significance is assigned.

Building objects inside a spatial representation

Blumberg's earlier source record already joins spatial features, object categories and predicted movement; Bach and Sorensen's later text also builds and tracks objects within a represented world.

These are card-level counts. The comparison link is not an independently counted target repetition.

Micah Blumberg / SAN

2025-06-21 · Git wording record

Git wording record; first public release not independently dated.

April 2017 (attributed; original audio/upload not verified)

And we're going to call this the concept of a car.

And not only do we have this concept of a car, which is all these lines, but we can also predict which way the car is moving.

Exact transcript and surrounding passage

Joscha Bach and coauthors

2026-01-15 · selected version

Selected PDF version date, not initial article date.

The Machine Consciousness Hypothesis · PDF page 20

Consciousness sorts things and animals into different kinds, identifies and tracks individuals, and gives each their names.

Read the target passage in context

Shared operation · partial

Both descriptions organize perceptual features into represented objects and use more than a bare label: the SAN passage predicts movement; the later passage identifies and tracks individuals in an internally constructed space.

L1: partial

Partial conceptual and operational correspondence: spatial features become individuated/categorized objects within a model. Tracking and motion prediction are related temporal operations, not identical algorithms.

L2: bounded-subchain-correspondence

Common ordered subchain: organize spatial distinctions -> construct object representations -> maintain or predict an object over change. SAN links it to paired brain/computer measurement; MCH embeds it in a genesis story.

L3: COMPATIBLE-RECONSTRUCTION-NOT-UNIQUE

Work backward from recognizing the same moving individual: the system must retain discriminative features and link observations across change; grouped object representations are a compatible route. Output does not determine a unique representation, consciousness mechanism, thalamic code, or influence.

What this pairing does not equate

The SAN passage does not here present MCH's second-order-perception genesis mechanism. MCH does not here give SAN's joint external-scene/neural-measurement BCI procedure. Shared categorization alone is historically broad.

Formal relation actually supported

Qualitative partial trace: features -> object-in-space -> temporal object relation. No equality of update functions, full simulation, functor or bisimulation is proved.

Earlier target work and citation status

The 21-occurrence program history remains open. This is document-relative source-record chronology, not author-wide firstness or a completed before/after migration finding.

Citation status: not-evaluable. No missing-citation finding is asserted.

Source identity, dates, authors and hashes

Source labels OpenAI Whisper transcription; audio accuracy not independently checked here. Quotation layout folds whitespace and ligatures only; spoken wording was not silently repaired.

Transcript SHA-256: A1138D2B2324987EB96E1595D622165E13EE6B2CD9E978B32B1A753ABDE783BB

Target SHA-256: 79308E09995AA00A6EFF0EE9DF0718BE101C8083286A34F5A5453016C956A46B

Authors: Joscha Bach, Hikari Sorensen

Git wording record; first public release not independently dated. Selected PDF version date, not initial article date.

Original recording route (verification pending).

Source passages overlap across the five reviewed framings. These links and author edges are not independent inventions, target repetitions or statistical confirmations.

Open this flashcard

Combining partial information into an experienced world

Blumberg's earlier source record explains experience as sparse sensory input completed by internal concepts; the later CIMC text likewise builds a perceptual world from partial models.

These are card-level counts. The comparison link is not an independently counted target repetition.

Micah Blumberg / SAN

2025-06-21 · Git wording record

Git wording record; first public release not independently dated.

April 2017 (attributed; original audio/upload not verified)

These are all, um, supporting concepts that are taking a little bit of information coming from your senses and bolstering that into your complete experience.

Exact transcript and surrounding passage

Joscha Bach and coauthors

2026-04-27 · selected version

Selected PDF version date, not initial article date.

CIMC Research Program for Machine Consciousness, Whitepaper v2 · PDF page 11

they minimize constraint violations between partial models of reality maintained in parallel.

Read the target passage in context

Shared operation · partial

Both make the experienced world depend on the combination of partial internal representations, not on simply copying the incoming sensory stream.

L1: partial

Partial representational assembly: limited sensory information and supporting concepts contribute to a complete experienced scene; concurrent partial models contribute to a consistent perceptual world.

L2: bounded-subchain-correspondence

Shared subchain: partial contributions -> integration in an internal representational system -> experienced world. Conflict detection and its resolution are explicit in CIMC and are not silently read into this SAN sentence.

L3: COMMON-INTEGRATION-REQUIREMENT-WITH-REPAIR-RESIDUAL

Work backward from a stable scene built from partial, potentially incompatible models: contributions must be represented and their conflicts must be handled by reconciliation, weighting, exclusion or suppression. The monologue's internally supplied context and converging senses support the integration portion, but do not specify that conflict-handling operation. Sensory completion and alternative model-selection schemes could yield the bounded perceptual output; neither the output nor the source fixes CIMC's attention/repair implementation.

What this pairing does not equate

Sensory completion using supporting concepts does not by itself specify CIMC's contradiction detection, attention allocation or suppression rule. Those differences limit this pair, without removing its shared integration operation.

Formal relation actually supported

Partial input/transform/output alignment of sensory-and-internal contributions -> integration -> scene; no proof that the two integration operators implement the same optimization.

Earlier target work and citation status

The 21-occurrence program history remains open. This is document-relative source-record chronology, not author-wide firstness or a completed before/after migration finding.

Citation status: not-evaluable. No missing-citation finding is asserted.

Source identity, dates, authors and hashes

Source labels OpenAI Whisper transcription; audio accuracy not independently checked here. Quotation layout folds whitespace and ligatures only; spoken wording was not silently repaired.

Transcript SHA-256: A1138D2B2324987EB96E1595D622165E13EE6B2CD9E978B32B1A753ABDE783BB

Target SHA-256: B472B40F89780EAB1BDA83C1935B58849D6F0C52D6078BB5319BE86B60F8CA9D

Authors: Joscha Bach, Hikari Sorensen, Franz Hildebrandt-Harangozó, Lou de Kerhuelvez, Jim Rutt

Git wording record; first public release not independently dated. Selected PDF version date, not initial article date.

Original recording route (verification pending).

Source passages overlap across the five reviewed framings. These links and author edges are not independent inventions, target repetitions or statistical confirmations.

Open this flashcard

Representations participate in changing mental states

Blumberg's earlier record proposes an operation that changes an ongoing neural experience; the later MCH text also treats mental states as causally transformable. The common state-transition principle is partial, not an identical controller.

These are card-level counts. The comparison link is not an independently counted target repetition.

Micah Blumberg / SAN

2025-06-21 · Git wording record

Git wording record; first public release not independently dated.

April 2017 (attributed; original audio/upload not verified)

But what you would do then is figure out the difference between the pattern that's there and the pattern that should be there if steak was there.

And then you send the difference between those two patterns so that the pattern that received just allows you to have steak into your reality where there was none before.

Exact transcript and surrounding passage

Joscha Bach and coauthors

2026-01-15 · selected version

Selected PDF version date, not initial article date.

The Machine Consciousness Hypothesis · PDF page 17

By characterizing consciousness as an operator on mental states, we can approach its functionality by observing how mental states change as a result of conscious operation.

Read the target passage in context

Shared operation · weak-partial

Both proposals give represented mental state a causal role in a transition to another mental state, rather than treating it as an inert report.

L1: weak-partial

Partial process correspondence at the state-transition level. SAN specifies a current/desired comparison and proposed external write-in; MCH characterizes internally conscious operation by its effects on mental state.

L2: bounded-subchain-correspondence

Shared narrow subchain: a represented mental state -> an operation on that state -> changed mental content. The controller and cause of the operation differ, so this is not a match of the whole control architecture.

L3: UNDERDETERMINED-STATE-CHANGE-RECONSTRUCTION

Given a changed mental state, some effective transition must have occurred; that observation alone neither entails a desired-state comparator nor a difference-coded signal. MCH's more constrained coherent output is compatible with many repair operations. A full delta-controller cannot be reverse-inferred.

What this pairing does not equate

MCH does not specify a current/desired difference or an external BCI channel in this passage. SAN's difference-write proposal is not a demonstration of consciousness's own control process.

Formal relation actually supported

Abstract partial transition correspondence s -> operation -> s′. The SAN transition carries a specified desired state and difference message; those labels have no exact mapped MCH operator here.

Earlier target work and citation status

The 21-occurrence program history remains open. This is document-relative source-record chronology, not author-wide firstness or a completed before/after migration finding.

Citation status: not-evaluable. No missing-citation finding is asserted.

Source identity, dates, authors and hashes

Source labels OpenAI Whisper transcription; audio accuracy not independently checked here. Quotation layout folds whitespace and ligatures only; spoken wording was not silently repaired.

Transcript SHA-256: A1138D2B2324987EB96E1595D622165E13EE6B2CD9E978B32B1A753ABDE783BB

Target SHA-256: 79308E09995AA00A6EFF0EE9DF0718BE101C8083286A34F5A5453016C956A46B

Authors: Joscha Bach, Hikari Sorensen

Git wording record; first public release not independently dated. Selected PDF version date, not initial article date.

Original recording route (verification pending).

Source passages overlap across the five reviewed framings. These links and author edges are not independent inventions, target repetitions or statistical confirmations.

Open this flashcard

Using a represented goal to select a state change

Blumberg's earlier record computes a neural update from a represented desired state; the later CIMC definition also makes represented goals guide control over future states.

These are card-level counts. The comparison link is not an independently counted target repetition.

Micah Blumberg / SAN

2025-06-21 · Git wording record

Git wording record; first public release not independently dated.

April 2017 (attributed; original audio/upload not verified)

But what you would do then is figure out the difference between the pattern that's there and the pattern that should be there if steak was there.

And then you send the difference between those two patterns so that the pattern that received just allows you to have steak into your reality where there was none before.

Exact transcript and surrounding passage

Joscha Bach and coauthors

2026-04-27 · selected version

Selected PDF version date, not initial article date.

CIMC Research Program for Machine Consciousness, Whitepaper v2 · PDF page 12

the capacity of a system to exert control over future states, that is, to select among possible actions on the basis of internal representations of goals, preferences, or evaluative criteria.

Read the target passage in context

Shared operation · partial

Both use a represented intended state to guide a change, rather than applying an unrelated fixed signal.

L1: partial

Partial operational correspondence: represent a goal or desired pattern -> choose the change relative to it -> influence a subsequent state. SAN supplies a particular difference-write procedure; CIMC supplies an abstract agency definition.

L2: bounded-subchain-correspondence

Shared control subchain: represented future condition -> selection of action/update -> changed state. It bridges BCI and agency contexts without claiming that the external BCI controller is the target's self-organizing agent.

L3: GOAL-SENSITIVE-CONTROL-RECONSTRUCTION-NOT-DELTA-IDENTITY

Work backward from selective control toward a represented goal: the selected intervention must be sensitive to an evaluative criterion and capable of affecting the outcome. This permits delta control, planning, policy selection or other strategies; it does not require the source's difference-signal protocol.

What this pairing does not equate

CIMC's definition does not mandate a neural pattern difference, receiver-specific write-in or external interface. The correspondence is one subchain of goal-directed control, not proof of the complete SAN architecture.

Formal relation actually supported

Partial control schema: (current state, represented goal) -> selected action/update -> subsequent state. Identical objectives, transition kernels and goal evaluation are not proved.

Earlier target work and citation status

The 21-occurrence program history remains open. This is document-relative source-record chronology, not author-wide firstness or a completed before/after migration finding.

Citation status: not-evaluable. No missing-citation finding is asserted.

Source identity, dates, authors and hashes

Source labels OpenAI Whisper transcription; audio accuracy not independently checked here. Quotation layout folds whitespace and ligatures only; spoken wording was not silently repaired.

Transcript SHA-256: A1138D2B2324987EB96E1595D622165E13EE6B2CD9E978B32B1A753ABDE783BB

Target SHA-256: B472B40F89780EAB1BDA83C1935B58849D6F0C52D6078BB5319BE86B60F8CA9D

Authors: Joscha Bach, Hikari Sorensen, Franz Hildebrandt-Harangozó, Lou de Kerhuelvez, Jim Rutt

Git wording record; first public release not independently dated. Selected PDF version date, not initial article date.

Original recording route (verification pending).

Source passages overlap across the five reviewed framings. These links and author edges are not independent inventions, target repetitions or statistical confirmations.

Open this flashcard

Experienced content need not reproduce the outside scene

Blumberg's earlier record explicitly separates an experienced object from its external presence; Bach and Sorensen later also distinguish conscious content from its real or imaginary character.

These are card-level counts. The comparison link is not an independently counted target repetition.

Micah Blumberg / SAN

2025-06-21 · Git wording record

Git wording record; first public release not independently dated.

April 2017 (attributed; original audio/upload not verified)

And then you send the difference between those two patterns so that the pattern that received just allows you to have steak into your reality where there was none before.

So once we can do that, that means that you're going to have things in your reality that are not actually there

Exact transcript and surrounding passage

Joscha Bach and coauthors

2026-01-15 · selected version

Selected PDF version date, not initial article date.

The Machine Consciousness Hypothesis · PDF page 16

The contents of consciousness may be experienced as real or imaginary

Read the target passage in context

Shared operation · partial-conceptual

Both distinguish the content of experience from whether the corresponding object is physically present outside the subject.

L1: partial-conceptual

Partial conceptual/output correspondence: an internal representational state can contain an object without that object being present in the external environment. The target separately describes a variable realness dimension.

L2: bounded-subchain-correspondence

SAN: state-dependent neural intervention -> added experienced object. MCH: representational content -> real/imaginary character. The shared endpoint is represented experience rather than necessary external presence; the causal routes are not fully aligned.

L3: INTERNAL-CONTENT-RECONSTRUCTION-NOT-WRITE-PROTOCOL

Work backward from experiencing an externally absent object: an internal content-carrying process must support that appearance independently of that object's present sensory cause. Dreams, hallucinations, imagery or intervention could do so; no specific pattern-difference input or realness variable is forced.

What this pairing does not equate

The target's claim concerns the representational character of experience, not this specific neural intervention. An induced percept, imagined content and an experience of factuality must not be conflated.

Formal relation actually supported

Partial output relation between internally represented object and external-object presence; no equivalence of hallucination, imagination or realness-detection operators.

Earlier target work and citation status

The 21-occurrence program history remains open. This is document-relative source-record chronology, not author-wide firstness or a completed before/after migration finding.

Citation status: not-evaluable. No missing-citation finding is asserted.

Source identity, dates, authors and hashes

Source labels OpenAI Whisper transcription; audio accuracy not independently checked here. Quotation layout folds whitespace and ligatures only; spoken wording was not silently repaired.

Transcript SHA-256: A1138D2B2324987EB96E1595D622165E13EE6B2CD9E978B32B1A753ABDE783BB

Target SHA-256: 79308E09995AA00A6EFF0EE9DF0718BE101C8083286A34F5A5453016C956A46B

Authors: Joscha Bach, Hikari Sorensen

Git wording record; first public release not independently dated. Selected PDF version date, not initial article date.

Original recording route (verification pending).

Source passages overlap across the five reviewed framings. These links and author edges are not independent inventions, target repetitions or statistical confirmations.

Open this flashcard

The experienced world is an internally maintained representation

Blumberg's earlier record treats the experienced world as internally constructed and alterable; the later CIMC text likewise names maintenance of a perceptual world. This is a broad partial correspondence.

These are card-level counts. The comparison link is not an independently counted target repetition.

Micah Blumberg / SAN

2025-06-21 · Git wording record

Git wording record; first public release not independently dated.

April 2017 (attributed; original audio/upload not verified)

These are all, um, supporting concepts that are taking a little bit of information coming from your senses and bolstering that into your complete experience.

And then you send the difference between those two patterns so that the pattern that received just allows you to have steak into your reality where there was none before.

Exact transcript and surrounding passage

Joscha Bach and coauthors

2026-04-27 · selected version

Selected PDF version date, not initial article date.

CIMC Research Program for Machine Consciousness, Whitepaper v2 · PDF page 11

the maintenance of a stable perceptual world

Read the target passage in context

Shared operation · weak-partial-conceptual

Both place the experienced world in an ongoing internal representational process, which can be maintained or changed rather than merely received as a complete external picture.

L1: weak-partial-conceptual

Weak partial conceptual correspondence in the maintained representational output. SAN describes internal completion and a proposed intervention; CIMC describes stable world maintenance as an operation of consciousness.

L2: bounded-subchain-correspondence

Both connect organized internal content to experienced-world availability. The evidence does not align the full repair mechanism: SAN's controlled write-in and CIMC's conflict resolution remain different specified procedures.

L3: WORLD-MAINTENANCE-RECONSTRUCTION-WRITABILITY-NOT-ENTAILED

Work backward from a stable perceptual world despite partial inputs: internal representations must be maintained and integrated. That output can arise without any external write channel or capacity to insert nonexistent objects; the target does not select the source's intervention route.

What this pairing does not equate

This pair is broad and weak by itself. It does not establish identical maintenance, minimal experience, self-registration or coherence optimization. Stronger operations require additional source-specific rows, not expansion of this quotation.

Formal relation actually supported

Qualitative common output class: an internally constructed perceptual world. No implementation-level or mathematical equivalence is established by this pair.

Earlier target work and citation status

The 21-occurrence program history remains open. This is document-relative source-record chronology, not author-wide firstness or a completed before/after migration finding.

Citation status: not-evaluable. No missing-citation finding is asserted.

Source identity, dates, authors and hashes

Source labels OpenAI Whisper transcription; audio accuracy not independently checked here. Quotation layout folds whitespace and ligatures only; spoken wording was not silently repaired.

Transcript SHA-256: A1138D2B2324987EB96E1595D622165E13EE6B2CD9E978B32B1A753ABDE783BB

Target SHA-256: B472B40F89780EAB1BDA83C1935B58849D6F0C52D6078BB5319BE86B60F8CA9D

Authors: Joscha Bach, Hikari Sorensen, Franz Hildebrandt-Harangozó, Lou de Kerhuelvez, Jim Rutt

Git wording record; first public release not independently dated. Selected PDF version date, not initial article date.

Original recording route (verification pending).

Source passages overlap across the five reviewed framings. These links and author edges are not independent inventions, target repetitions or statistical confirmations.

Open this flashcard

Patterns change what the receiving system represents

Blumberg's earlier record gives a receiver-dependent way to transform a neural representation; the later MCH account also describes patterns transformed through conditional processing.

These are card-level counts. The comparison link is not an independently counted target repetition.

Micah Blumberg / SAN

2025-06-21 · Git wording record

Git wording record; first public release not independently dated.

April 2017 (attributed; original audio/upload not verified)

But what you would do then is figure out the difference between the pattern that's there and the pattern that should be there if steak was there.

And then you send the difference between those two patterns so that the pattern that received just allows you to have steak into your reality where there was none before.

Exact transcript and surrounding passage

Joscha Bach and coauthors

2026-01-15 · selected version

Selected PDF version date, not initial article date.

The Machine Consciousness Hypothesis · PDF page 16

Representations are transformed into other representations, each characterized by their structural invariances and their place in the processing network, which can be seen as a causal network that propagates patterns via conditional operators.

Read the target passage in context

Shared operation · partial

Both use transmitted patterns to produce a further representation through a receiving process. SAN specifies the current receiver state and a difference message; MCH describes the general conditional transformation network.

L1: partial

Partial operational subchain: incoming pattern -> receiving transformation -> changed representation. SAN's state-dependent delta is more specific than the target's general conditional operator here.

L2: bounded-subchain-correspondence

The matched architecture is representational transformation along a causal path, not just shared neural vocabulary. Different implementations remain: an external BCI update versus a general processing network.

L3: CAUSAL-TRANSFORM-RECONSTRUCTION-OPERATOR-UNDERDETERMINED

Work backward from a reproducibly changed representation: there must be a substrate-sensitive causal mapping from some prior/input pattern to the new one. Conditional network propagation is compatible; so are diverse encodings and operators. Difference coding and electromagnetic dominoes are not forced.

What this pairing does not equate

The target paragraph does not specify transmission of only a difference, feedback verification or an identical receiver-state equation. Those remain unmatched SAN specifics within this pair.

Formal relation actually supported

A bounded example-to-general-process relation: U(s,message) -> represented s′. Not a proof that the target conditional operator is the SAN delta operator.

Earlier target work and citation status

The 21-occurrence program history remains open. This is document-relative source-record chronology, not author-wide firstness or a completed before/after migration finding.

Citation status: not-evaluable. No missing-citation finding is asserted.

Source identity, dates, authors and hashes

Source labels OpenAI Whisper transcription; audio accuracy not independently checked here. Quotation layout folds whitespace and ligatures only; spoken wording was not silently repaired.

Transcript SHA-256: A1138D2B2324987EB96E1595D622165E13EE6B2CD9E978B32B1A753ABDE783BB

Target SHA-256: 79308E09995AA00A6EFF0EE9DF0718BE101C8083286A34F5A5453016C956A46B

Authors: Joscha Bach, Hikari Sorensen

Git wording record; first public release not independently dated. Selected PDF version date, not initial article date.

Original recording route (verification pending).

Source passages overlap across the five reviewed framings. These links and author edges are not independent inventions, target repetitions or statistical confirmations.

Open this flashcard

Feedback and confirmation in distributed communication

Blumberg's earlier record proposes feedback and verified neural communication; the later CIMC program also uses requests and confirmations between processing units.

These are card-level counts. The comparison link is not an independently counted target repetition.

Micah Blumberg / SAN

2025-06-21 · Git wording record

Git wording record; first public release not independently dated.

April 2017 (attributed; original audio/upload not verified)

because the connection means there's, there's actually a feedback loop of information.

you're able to send a whole and complete message and you're able to verify that it was sent

Exact transcript and surrounding passage

Joscha Bach and coauthors

2026-04-27 · selected version

Selected PDF version date, not initial article date.

CIMC Research Program for Machine Consciousness, Whitepaper v2 · PDF page 19

processing units negotiate mutual consistency through iterative exchange of requests and confirmations rather than by centralized control.

Read the target passage in context

Shared operation · partial

Both describe neural or computational communication using reciprocal exchange and confirmation, not only one-way transmission.

L1: partial

Partial communication pattern: send -> returned check/confirmation -> continued organized communication. SAN uses a TCP analogy and a conscious/unconscious proposal; CIMC proposes Request Confirmation Networks for mutual consistency.

L2: bounded-subchain-correspondence

Shared feedback architecture joins local communications through returned information. Reliable delivery in the SAN analogy and semantic consistency in CIMC are not identical goals, but the reciprocal-check subchain is supported.

L3: CONFIRMATION-RECONSTRUCTION-NOT-PROTOCOL-IDENTITY

Work backward from mutually confirmed coordination: receiving units must affect the subsequent exchange, and a returned signal must bear on whether a proposed communication or state is accepted. The monologue supplies a compatible return-check motif. Its delivery acknowledgement alone cannot establish consistency of represented contents; acknowledgements, negotiation, consensus and other feedback protocols can differ in what they certify. Neither TCP nor an identical consciousness/RCN mechanism is forced.

What this pairing does not equate

TCP-style delivery confirmation is not itself semantic agreement. SAN here does not supply CIMC's Request Confirmation Network implementation; CIMC here does not specify the SAN perceptual difference-write operation.

Formal relation actually supported

Partial protocol trace: message -> receiver -> return check. No equality of acknowledgement semantics, convergence conditions or network implementation is proved.

Earlier target work and citation status

The 21-occurrence program history remains open. This is document-relative source-record chronology, not author-wide firstness or a completed before/after migration finding.

Citation status: not-evaluable. No missing-citation finding is asserted.

Source identity, dates, authors and hashes

Source labels OpenAI Whisper transcription; audio accuracy not independently checked here. Quotation layout folds whitespace and ligatures only; spoken wording was not silently repaired.

Transcript SHA-256: A1138D2B2324987EB96E1595D622165E13EE6B2CD9E978B32B1A753ABDE783BB

Target SHA-256: B472B40F89780EAB1BDA83C1935B58849D6F0C52D6078BB5319BE86B60F8CA9D

Authors: Joscha Bach, Hikari Sorensen, Franz Hildebrandt-Harangozó, Lou de Kerhuelvez, Jim Rutt

Git wording record; first public release not independently dated. Selected PDF version date, not initial article date.

Original recording route (verification pending).

Source passages overlap across the five reviewed framings. These links and author edges are not independent inventions, target repetitions or statistical confirmations.

Open this flashcard

From understanding biological operations to artificial implementations

Blumberg's earlier Neural Lace program explicitly continues into artificial neocortices and brains; Bach and Sorensen later also propose recreating biological information processing computationally.

These are card-level counts. The comparison link is not an independently counted target repetition.

Micah Blumberg / SAN

2025-06-21 · Git wording record

Git wording record; first public release not independently dated.

April 2017 (attributed; original audio/upload not verified)

But after we build neural lace, the idea is to, um, uh, build artificial neocortex and then artificial brains.

Exact transcript and surrounding passage

Joscha Bach and coauthors

2026-01-15 · selected version

Selected PDF version date, not initial article date.

The Machine Consciousness Hypothesis · PDF page 21

it is possible to search for this algorithm by recreating analogous conditions of self-organizing information processing on digital computer hardware, while posing suitable tasks that require intelligent agency.

Read the target passage in context

Shared operation · partial-conceptual

Both research programs connect understanding biological information processing with constructing an artificial implementation.

L1: partial-conceptual

Partial research-operation correspondence: identify biological information-processing requirements -> seek an artificial realization. SAN links interface understanding to artificial neocortices/brains; MCH specifies a search for a self-organizing learning algorithm.

L2: bounded-subchain-correspondence

Shared investigation-to-construction sequence rather than shared terminology alone. SAN's preceding brain/computer representation bridge supplies an operational link; the later target makes biological-to-digital realization explicit.

L3: FUNCTIONAL-REALIZATION-RECONSTRUCTION-SPECIFIC-LEARNING-ROUTE-OPEN

Work backward from an artificial realization of the target's biological learning role: the artificial system must support the relevant representations, causal operations and learning conditions, not merely reproduce a human-looking output. The monologue's human/computer spatial bridge and artificial-brain roadmap support the broad realization direction. They do not specify self-organization, the hypothesized consciousness-producing algorithm or sufficient developmental conditions; engineered perception could supply the source's demonstrated analogy without the target's full hypothesis.

What this pairing does not equate

The source does not here state MCH's evolutionary-search or second-order-learning algorithm. Artificial brain construction was a broad pre-existing research aim; this pair is not a worldwide firstness finding.

Formal relation actually supported

Partial design-program relation from biological operation study to artificial realization, not a proof of functional equivalence of any constructed systems.

Earlier target work and citation status

The 21-occurrence program history remains open. This is document-relative source-record chronology, not author-wide firstness or a completed before/after migration finding.

Citation status: not-evaluable. No missing-citation finding is asserted.

Source identity, dates, authors and hashes

Source labels OpenAI Whisper transcription; audio accuracy not independently checked here. Quotation layout folds whitespace and ligatures only; spoken wording was not silently repaired.

Transcript SHA-256: A1138D2B2324987EB96E1595D622165E13EE6B2CD9E978B32B1A753ABDE783BB

Target SHA-256: 79308E09995AA00A6EFF0EE9DF0718BE101C8083286A34F5A5453016C956A46B

Authors: Joscha Bach, Hikari Sorensen

Git wording record; first public release not independently dated. Selected PDF version date, not initial article date.

Original recording route (verification pending).

Source passages overlap across the five reviewed framings. These links and author edges are not independent inventions, target repetitions or statistical confirmations.

Open this flashcard

Related functions across biological and artificial systems

Blumberg’s earlier source record explicitly compares human and artificial visual grouping and prediction; the later CIMC discussion likewise connects artificial visual representations with animal visual cortex.

These are card-level counts. The comparison link is not an independently counted target repetition.

Micah Blumberg / SAN

2025-06-21 · Git wording record

Git wording record; first public release not independently dated.

April 2017 (attributed; original audio/upload not verified)

And the self-driving car is doing the same thing.

It's predicting there is a concept, which is a unification of a bunch of lines and edges and spatial computing and cars doing it and humans are doing it.

Exact transcript and surrounding passage

Joscha Bach and coauthors

2026-04-27 · selected version

Selected PDF version date, not initial article date.

CIMC Research Program for Machine Consciousness, Whitepaper v2 · PDF page 17

These representations closely matched the known organization of the animal visual cortex, despite the artificial systems having no biological constraints.

Read the target passage in context

Shared operation · partial-conceptual

Both compare visual representation across brains and artificial systems. SAN explains grouping lines and edges into objects and predicting movement; CIMC discusses related features learned by artificial vision systems.

L1: partial-conceptual

Partial cross-substrate perceptual-function correspondence: both relate biological and artificial visual representation. The selected 2026 target cites comparative AI evidence from 2020; the SAN passage provides a functional example, not that experiment.

L2: bounded-subchain-correspondence

Shared reasoning chain: visual task -> internally organized visual representation -> a bridge between biological and artificial systems. The target then extends a convergence conjecture to consciousness; that extra arrow is not imported into SAN’s car example.

L3: TASK-CONSTRAINED-REPRESENTATION-RECONSTRUCTION-CONSCIOUSNESS-RESIDUAL

Work backward first from the target's visual-recognition example: a successful biological or artificial system needs task-suitable discriminative representations, making the monologue's shared object-grouping/prediction role a compatible antecedent. Similar output does not uniquely determine internal geometry or establish empirical feature convergence. Work backward separately from the target's conditional consciousness extension: coherent world/self modeling under agentic and self-organizing constraints is additionally required by that hypothesis. Those conditions are not supplied by the monologue's car example or its artificial-brain intention.

What this pairing does not equate

The target cites Olah et al. (2020) for experimental AI representation comparisons and adds a consciousness-convergence conjecture. The earlier SAN passage supplies a cross-substrate perceptual-function bridge, not those experiments, a universality theorem or equal consciousness.

Formal relation actually supported

Cross-substrate partial functional analogy, not proof of universal convergence, identical algorithms or consciousness equivalence.

Earlier target work and citation status

The 21-occurrence program history remains open. This is document-relative source-record chronology, not author-wide firstness or a completed before/after migration finding.

Citation status: not-evaluable. No missing-citation finding is asserted.

Source identity, dates, authors and hashes

Source labels OpenAI Whisper transcription; audio accuracy not independently checked here. Quotation layout folds whitespace and ligatures only; spoken wording was not silently repaired.

Transcript SHA-256: A1138D2B2324987EB96E1595D622165E13EE6B2CD9E978B32B1A753ABDE783BB

Target SHA-256: B472B40F89780EAB1BDA83C1935B58849D6F0C52D6078BB5319BE86B60F8CA9D

Authors: Joscha Bach, Hikari Sorensen, Franz Hildebrandt-Harangozó, Lou de Kerhuelvez, Jim Rutt

Git wording record; first public release not independently dated. Selected PDF version date, not initial article date.

Original recording route (verification pending).

Source passages overlap across the five reviewed framings. These links and author edges are not independent inventions, target repetitions or statistical confirmations.

Open this flashcard