The problem Sourfly solves

It solves a niche problem, actually to understand about the learning, connections, movement, neuron's channels and most important simulation.

The other main thing is that its easy to use and understand so most people who is just starting their rabbit hole in this can understand easily.

This model uses FlyWire FAFB v783 hardcoded neurons as per now until i connect it to a VPS.

const VISUAL = [
    "LTe42b",
    "LTe15",
    "LPT54",
    "LT87",
    "LPT48_vCal3",
    "VST2",
    "LPLC4"
];


const CENTRAL = [
    "CB0524",
    "SAD043",
    "LHAD1g1",
    "AVLP340",
    "CB0500",
    "AVLP435a",
    "Nod1",
    "CB0268",
    "CB0316",
    "PVLP020",
    "PS213",
    "LTe42a",
    "PLP213",
    "CB0492",
    "PS174",
    "PS098",
    "PLP248"
];


const DESCENDING = [
    "DNae005",
    "DNbe007",
    "DNge054",
    "DNp103",
    "DNp06",
    "DNp55",
    "DNb06",
    "DNp56",
    "DNp26",
    "DNg41",
    "DNp09",
    "DNa10",
    "DNp57",
    "DNg46"
];


const PATHWAYS = [
    ["LTe42b", "CB0524", "DNae005", 391, 130],
    ["LTe42b", "CB0524", "DNbe007", 391, 101],
    ["LTe15", "SAD043", "DNbe007", 360, 115],
    ["LTe15", "SAD043", "DNge054", 360, 40],
    ["LPT54", "SAD043", "DNge054", 266, 152],
    ["LPT54", "SAD043", "DNbe007", 266, 102],
    ["LT87", "LHAD1g1", "DNp103", 259, 121],
    ["LT87", "LHAD1g1", "DNp06", 259, 120],
    ["LT87", "AVLP340", "DNp55", 174, 175],
    ["LPT48_vCal3", "CB0500", "DNb06", 47, 542],
    ["VST2", "CB0500", "DNb06", 42, 542],
    ["LT1d", "AVLP435a", "DNp103", 330, 68],
    ["LPT22", "Nod1", "DNp26", 395, 55],
    ["LPT04_HST", "CB0268", "DNg41", 228, 87],
    ["LT86", "CB0316", "DNbe007", 153, 124],
    ["LPT51", "SAD043", "DNge054", 120, 152],
    ["LT82a", "PVLP020", "DNp09", 92, 191],
    ["VSm", "PS213", "DNb06", 104, 168],
    ["LTe17", "LTe42a", "DNp56", 132, 128],
    ["LTe07", "PLP213", "DNa10", 76, 207],
    ["LTe42a", "CB0492", "DNbe007", 195, 80],
    ["LPLC4", "PLP213", "DNa10", 75, 207],
    ["VST2", "PS174", "DNg46", 84, 176],
    ["aMe25", "PS098", "DNp57", 118, 125],
    ["vCal1", "PLP248", "DNa10", 189, 78]
];

Done

• Added neuron analyzer
• Lesion Mode
• bug fixes
• tick speed system

Roadmap

• Real time fly brain diagram in 3D showing currently processing neural networks.

Challenges we ran into

The main challenge was to manage all the neurons (around 18k) in millions. but we managed each one as by basic mapping slowly each cluster. we are using less neurons currently thats a reason in the current version fly feels dumb but will soon fix it.

there are these known bugs in the code

bug 1: The wall thrashing and thing (fixed)
bug 2: DUMB AS HELL
bug 3: Fly stopping and collision distance increasing