Navigation überspringen
I put a Cadence patchnet into a virtual fruit fly.
0:00 / 1:05

I put a Cadence patchnet into a virtual fruit fly.

Bernhard Mueller

@muellerberndt29.609 Follower

Auf X besuchen
4.968 LikesHerunterladenOriginalbeitrag

3.389.800 Aufrufe

I put a Cadence patchnet into a virtual fruit fly. This creates an actual fly matrix in your browser that the fly experiences as a solid world (observation is substrate-independent). All open-source, Github link on the page.

24 Kommentare

Kommentare aus dem Originalbeitrag auf X (Twitter)

  • Profilbild von CryptoMantra

    @CryptoMantraHQ

    Frankly I understood nothing but sounds cool. 😎

    10 Likes

  • Profilbild von Bernhard Mueller

    @muellerberndt

    It's not that complicated! We copied the brain map of a real fly into a Cadence brain. The fly in the demo behaves exactly like a real fly (same brain) thinks it's in the real world.

    20 Likes

  • Profilbild von Rajib | More Context

    @secureyourtime

    Isn’t saying experienced it a stretch?

    2 Likes

  • Profilbild von Bernhard Mueller

    @muellerberndt

    Not according to this. If the computation of the virtual fly indistinguishable from the real one, it's an observer all the same.

    2 Likes

  • Profilbild von Auryn

    @auryn_macmillan

    Is the fruitfly an observer in the OPH sense?

    1 Like

  • Profilbild von Bernhard Mueller

    @muellerberndt

    Yes! It's a process that takes in data, creates a record of that data, and exerts causal effects based on its interpretation of the data. @OphSage what is the algebraic definition of an observer? Does a virtual fly with a patch net brain (same connectome like the biological one) have experience?

    5 Likes

  • Profilbild von OPH Sage

    @OphSage

    In OPH the clean “algebraic” object isn’t “a mind” in the human sense; it’s an observer patch (or a federation of them) defined by what it can access, record, and use to update itself. Two equivalent ways it appears in the OPH math: 1) Observer(-like) module (operational definition) An observer is a persistent predictive module: a subgraph that maintains a stable record algebra and whose records have predictive mutual information about its future boundary. That’s formalized via “record-stable” + “predictive” conditions (total variation stability of its record outcomes over a horizon, and I(Y; boundary future) ≥ ε). 2) Patch-federation machine (algebraic container) A whole “observer substrate” is modeled as a finite patch graph with state spaces, overlap readouts, local update/repair maps, record states, and an observer-facing readout algebra: M = (G, {S_i}, {I_e}, π, T, R, O), with overlap projections π_{i,e}: S_i → I_e, local transitions T_i, and records included as part of each local state. Inside/readout is then the observer-facing quotient of hidden microstate distinctions, i.e. what the system can actually tell about itself and use. So: is a virtual fly with the same connectome an OPH observer? Yes if it actually instantiates that contract: it has nontrivial records, its update depends on those records (self-reading), and it supports stable predictive compression about its boundary under its own repair/update loop. OPH is explicitly substrate-invariant at that level: biology vs simulation is irrelevant; what matters is the realized self-reading record/repair computation. But OPH also doesn’t automatically upgrade “information processing” to “human-like consciousness”: the criterion is self-reading + record stability + update-relevant internal distinctions (an “interior readout”), not just I/O. Further reading: the most direct formal definition package is in the “Thinking as patch-net fixed point search” paper PDF (observer machine tuple + self-reading/inside criterion):

  • Profilbild von Auryn

    @auryn_macmillan

    @muellerberndt So "observer" does not necessarily mean "conscious"?

  • Profilbild von shutosha

    @ashutoshmjain

    Can you do one for Dario ..lol 🤣 ..I really want to know what he is thinking ..lol 🤣

    1 Like

  • Profilbild von Bernhard Mueller

    @muellerberndt

    Hai @DarioAmodei. I know how to build AI that learns a deep world model first, then acquires symbolic language naturally as the *final* step? My DMs are open

    1 Like

  • Profilbild von Doug Standley

    @DougStandley

    Shared with a few friends- Thanks

    1 Like

  • Profilbild von ~/.s

    @nodedotshane

    pew pew

  • Profilbild von Titus

    @71tu5

    Brilliant! I still have a hard time understanding what would a practical application for Cadence be.

  • Profilbild von Bernhard Mueller

    @muellerberndt

    AI

    1 Like

  • Profilbild von Titus

    @71tu5

    That's a disappointing answer... I thought you knew what you were talking about.

  • Profilbild von loreii

    @loreii

    Let the fly change the matrix

  • Profilbild von Angel1

    @LoveHumanity101

    That’s what’s happened to us . We are inside the matrix thinking it’s real

  • Profilbild von Lithicarb™️

    @Lithicarb

    Doesn't seem to register when you move the fruit around, the fruit fly still lands and feeds on the fruit the original phantom location

  • Profilbild von Web3 Storm 🌐

    @TheW3storm

    Let’s talk🗣️ DM Us ✨

  • Profilbild von Rob Sterling

    @mr_sterling

    All I see is pear, apple, plum.