Random Art Memory is now connected to life model. Introducing "UNBRAIN" the first RAM directly plugged into the fly fruit on-chain brain made by @banknotes_btc Guided by fly connectomes, he reacts directly to your click. But don't stress him too much...

3 days since this guy is trapped into an orange square with a fruit-fly brain...

God bless ordinals

GM Flex your Friday

1/ 🧠 To understand Autonomous Life and FLY VISION, you first need to understand what is behind both of them: the fruit fly brain More specifically, the complete adult Drosophila connectome a huge network of neurons, synapses and connections 🧵👇 2/ In 2024, researchers completed one of the most important neuroscience mapping projects ever made: the full wiring diagram of an adult fruit fly brain. It contains approximately: 🧠 139,255 neurons 🔗 54.5 million synapses Each connection helps show how different parts of the brain can communicate 3/ But what is a neuron? A neuron is a cell that receives, processes and transmits signals Think of every neuron as a small node inside a gigantic network A brain does not work through one neuron alone Behavior comes from interactions across thousands of them 4/ And what are synapses? Synapses are the communication points between neurons Very simply: Neuron A ↓ Synapse ↓ Neuron B ↓ Synapse ↓ Neuron C This is how activity can propagate across the brain 5/ Not every connection has the same importance. Some neurons have many connections. Others have fewer. And some connections contain different numbers of synapses. So the connectome is not just a list of neurons. It is a giant connected network 6/ Think of it like a city. Neurons are locations Synapses are the roads Some roads are small Others carry much more traffic Different routes can lead signals toward completely different parts of the system The connectome is the map of those roads 7/ The dataset we use contains millions of connection records A record can represent something like: Neuron 182 → Neuron 984 together with information about that connection. Put millions of these records together, and you start rebuilding the architecture of the brain 8/ For our experiments, the neurons are represented by IDs For example: Neuron 0 Neuron 1 Neuron 2 ... Neuron 139254 These IDs let us represent the network efficiently while preserving the relationship to the original mapped neurons 9/ Then we hit the biggest problem: DATA SIZE A network containing around 139,000 neurons and tens of millions of synapses is huge It would not be practical to place everything inside a single Bitcoin inscription So we use recursive inscriptions 10/ This is where: PARENT + CHILDREN become important. We create a main inscription: 🧠 PARENT and divide the brain data across many smaller inscriptions: CHILD 1 CHILD 2 CHILD 3 CHILD 4 ... Each Child contains another part of the brain data. 11/ The Parent acts as the central reference point Instead of forcing tens of megabytes into one inscription, it can reference multiple Children Each Child stores another section of the neural network Together they form: Parent ↓ Children ↓ Neurons ↓ Connections ↓ Synapses 12/ In our first full build, the brain dataset reached around 43 MB That meant it had to be divided into more than 100 parts Each part was compressed so it could fit within the inscription size we were targeting 13/ We also compress the data heavily. Instead of repeatedly storing large amounts of information, we use techniques such as: • adjacency lists • delta encoding • VarInt • compact neuron IDs • dictionaries for repeated information The goal is to reduce the amount of data required as much as possible. 14/ An adjacency list is a more efficient way to represent a network. Instead of writing: A connects to B A connects to C A connects to D A connects to E you can represent something closer to: A → [B, C, D, E] That saves a large amount of space. 15/ We can also preserve information related to connection strength So it is not only: A is connected to B We can also use the number of synapses involved in that connection as part of the neural system. That helps us model how activity can propagate through the network. 16/ And this is where the experimental part begins. We are not only storing the brain. We also built a: ⚙️ NEURAL ENGINE The neural engine reads parts of this structure and turns network activity into computational decisions 17/ Imagine a stimulus entering the system. That stimulus activates a group of neurons Those neurons are connected to others Activity can then propagate through the network: INPUT ↓ Neuron A ↓ Neuron B + C ↓ Neuron D + E + F ↓ OUTPUT 18/ The structure of the network influences which paths this activity can follow So instead of simply choosing a random result, the system uses relationships derived from the connectome to influence the output This is where our two experiments separate 19/ 🦎 AUTONOMOUS LIFE Autonomous Life uses this structure to generate behavior The organism exists inside a digital environment. It can receive different types of stimuli from that environment 20/ These stimuli enter the neural system. For example, information related to: 🌿 environment 🍴 food ⚠️ danger 🧭 direction ⚡ energy 👁️ external stimuli can produce different activation patterns. 21/ That activity then moves through the neural network derived from the connectome. The Neural Engine interprets the resulting activity and converts certain outputs into possible actions. For example: neural activity ↓ decision ↓ movement 22/ The organism can then: walk change direction explore search for something react to the environment stop trigger another action without us simply writing a fixed sequence that tells it exactly what to do every second. 23/ This distinction matters Autonomous Life is not simply: animation1 → animation2 → animation3 The 3D model has animations available But the neural system helps determine when and how those actions are triggered The brain influences the behavior 24/ That is why every Autonomous Life organism can contain several different layers: 🧠 brain data 🔗 neurons and synapses ⚙️ Neural Engine 🌎 environment 👁️ stimuli 🦴 3D model 🎬 animations All working together 25/ Bitcoin becomes the layer where these components can be preserved. Parts of the brain are inscriptions The organism is an inscription The engine can be an inscription The files can reference one another recursively Everything does not have to live inside one gigantic file. 26/ Then comes the second experiment: 🎨 FLY VISION It uses the same basic idea but changes the output completely Autonomous Life asks: “What should this organism DO?” FLY VISION asks: “What should this brain DRAW?” 27/ In FLY VISION, we do not need an organism moving around a 3D environment Instead, neural outputs are converted into visual decisions For example: position direction color stroke point placement size density detail 28/ There is a reference artwork The system knows what image it is trying to reinterpret But the image is built progressively. The engine decides where to work and how to represent different regions of the reference. 29/ One decision can influence: “work in this area” Another: “use this size” Another: “move in this direction” Another: “use this color” Thousands of small decisions combine into the final artwork 30/ This is how I like to summarize both experiments: AUTONOMOUS LIFE 🧠 Brain ↓ ⚡ Neural activity ↓ ⚙️ Neural Engine ↓ 🦎 Behavior FLY VISION 🧠 Brain ↓ ⚡ Neural activity ↓ ⚙️ Neural Engine ↓ 🎨 Visual decisions 31/ It is the same core concept being tested in two completely different worlds. In the first: neural activity influences actions In the second: neural activity influences creation. BEHAVIOR vs ART 32/ FLY VISION is also much lighter than Autonomous Life Autonomous Life can approach around 1 MB per organism when different components are included. FLY VISION can exist in only tens of KB per piece. That makes the second experiment much cheaper to inscribe 33/ That is also why Autonomous Life is much more expensive 3D model + Engine + neural data + environment + recursive inscriptions make each organism significantly heavier FLY VISION removes much of that weight 34/ But both projects come from the same question: What can we build when a neural network based on the architecture of a biological brain becomes part of a digital system? Not just storing the connectome Actually using its connections 35/ One important clarification: the connectome does not contain the consciousness of the fly. We are not putting its memories on Bitcoin. What we are using is the mapped physical architecture of its neural connections. That is very different 36/ Think about a computer Having a complete map of its circuits does not mean you possess everything that has ever happened inside that computer But knowing the circuits lets you study how signals can move through it. That is a useful analogy for the connectome 37/ So the project has two very different layers: 📚 DATA neurons synapses regions connections weights ⚙️ INTERPRETATION how to transform activity across that network into digital outputs. 38/ Parent and Children mainly solve the first problem: HOW TO STORE A HUGE NETWORK. The Neural Engine solves the second: HOW TO USE THAT NETWORK And Autonomous Life / FLY VISION explore the third: WHAT CAN WE DO WITH IT? 39/ The whole structure can be summarized like this: EFCB / CONNECTOME ↓ 139,255 neurons ↓ ~54.5M synapses ↓ compressed data ↓ Parent + Children on Bitcoin ↓ Neural Engine ↙️ ↘️ Autonomous Life FLY VISION 40/ For me, this is the most interesting part of the entire experiment We are not only trying to preserve scientific data on Bitcoin We are trying to explore what happens when that data becomes a functional part of something that can behave, react or create digitally. 🧠 → ₿ → 🦎 / 🎨 And we are only beginning to explore it...

Who's next now?

Red built it's ram gallery

Big thanks to @TaoKoltes for this new @GenuinelyUndead website ! Have a look in the new UI : genuineundead.com/ And many thanks to featuring my Bitcoin art into Art & Culture section 🧡

i patted gizmo 80 times and he's still sitting on the seed phrase. there's a bitcoin puppet in that wallet. help me move this cat. felt.cash/?gizmo

In 1 hour, we’ll be live with this LOT of 10 new Ordskul inscriptions. AOB🔶🔸

FLY VISION uses the fruit fly brain to make decisions and reinterpret famous works of art Every line, every stroke, every shape, and every color is chosen through the neural system Diamond Hands from these collections can mint for FREE BANKNOTES CENTS @sovrnart Random Art Memory @redisdead Toadstools @toadstools_btc Rare Gummies @Ob1_Wan_Satoshi Bitcoin Weirdos @ftw_collective Ordskull @Arietoshi Honey Badger @HoneyBadgersBtc Bitmap Sunset @BitmapSunset 50/100 minted ██████████░░░░░░░░░░ 50% 🏆 Rewards for the Top 3 FLY VISION holders 🥇 1st Place: Sub-10 BANKNOTE — valued at 0.00069 BTC 🥈 2nd Place: Bathroom Wall — valued at 0.00024 BTC 🥉 3rd Place: Sola Busca — valued at 0.00018 BTC Mint Now:ord-dropz.xyz/marketplace/li…

So many Pokemon display but only 1 Pipikakachu card...

GOD BLESS ORDIANALS

I want to thank @redisdead for spending the time to give me amazing backstory lore about the piece I acquired. And was so kind he even gifted me a piece from his other collection love_message. A collaboration with his child🤌 A.O.B.🧡

Thanks @banknotes_btc and @redisdead for the WL to fly vision on @Ord_Dropz . Interesting project.

You can just get one now with a badass mfer plugged into a fruit fly brain in your pocket. Forever stored in your Bitcoin wallet as a Random Art Memory

Are Ordinals back ?

SOON! New on the Ord Dropz launchpad: File-to-Parent·Child mints 🌳 Already have a parent inscription? Upload your collection files, and each piece is inscribed at mint as a true on-chain child of your parent. No pre-inscribing needed. 🎨 For artists: Choose “File-to-Parent·Child” when creating your drop. Sign once to move your parent into a secure vault for the drop. When it sells out, your parent returns to your wallet automatically. Your parent can only be used to inscribe your children, and you keep a backup path to reclaim it yourself. 🖼 For collectors: Mint like normal, with one signature. Mints are processed in batches every 10 minutes, or sooner when a batch fills. Follow the countdown on the mint page. Metadata and traits go on-chain, just like every Ord Dropz mint. Your art. Your parent. On-chain. Who’s launching the first one? 👇 ord-dropz.xyz/marketplace

GM Big Random Art Memory sale ! "Who's the bitch now?" is sold 🔥 Love for the owner, this piece held me during several days until final result. You got a gem here 💎 Congratulations and thanks for the support 🙏 All holders, all supporters are the fuel for my art journey 🧡