MR_JACK 22

Mereka mengingat siapa yang mengirim apa.
Mereka mengingat keseimbangan.
Mereka mengingat kepemilikan.

Tetapi mereka tidak pandai mengingat kehidupan.

Mereka tidak mengingat gambar dengan baik.
Mereka tidak mengingat video.
Mereka tidak mengingat percakapan besar, sejarah, model, dunia, atau arsip.

Jadi Web3 diam-diam meminjam ingatan dari Web2.

NFT tinggal di server cloud.
Aplikasi dimuat dari situs web terpusat.
Dunia game berada di infrastruktur pribadi.
Data AI disimpan dalam silo perusahaan.

Logika terdesentralisasi, tetapi ingatan tidak.

Walrus diciptakan untuk bagian yang hilang itu.

Saya akan berbicara tentang Walrus seperti cara Anda berbicara tentang sesuatu yang benar-benar ingin Anda andalkan. Bukan sebagai ajakan penjualan. Bukan sebagai papan skor. Lebih seperti jalan santai melalui sistem yang berusaha melindungi salah satu kebutuhan paling manusiawi di era digital. Kebutuhan untuk menjaga hal-hal yang penting agar tidak secara diam-diam menghilang.

Walrus dibangun untuk penyimpanan blob. Kata blob terdengar dingin hingga Anda membayangkan makna sebenarnya. Blob adalah hal-hal berat dari dunia nyata. Foto dan video serta build aplikasi dan arsip serta file penelitian dan data pelatihan. Jenis file yang terlalu besar untuk hidup nyaman di atas rantai dan terlalu penting untuk dipercayakan hanya pada satu server. Walrus dirancang untuk menyimpan data tersebut di banyak operator independen sambil menggunakan Sui sebagai lapisan koordinasi publik di mana komitmen dan peristiwa siklus hidup dapat dicatat dengan cara yang bisa diverifikasi oleh orang lain.

Saya akan menggambarkan Dusk seperti cara Anda menggambarkan suatu tempat yang benar-benar Anda kunjungi. Bukan seperti brosur. Bukan seperti promosi. Hanya tekstur bagaimana cara kerjanya dan mengapa pilihan-pilihannya terasa sengaja. Dusk dimulai pada tahun 2018 dengan naluri yang sangat spesifik. Membangun Layer 1 untuk keuangan yang diatur di mana privasi tidak diperlakukan seperti rahasia yang bersalah dan di mana auditabilitas tidak diperlakukan seperti musuh. Dokumentasi menggambarkan hal ini dengan jelas sebagai privasi berbasis desain dengan transparansi saat dibutuhkan dan menghubungkan janji tersebut dengan dua model transaksi asli yang berada di lapisan penyelesaian.

Saya terus kembali pada rasa takut kecil pribadi yang kebanyakan pembuat belajar untuk menyembunyikan. Internet melupakan. Tautan rusak. Server dimatikan. Akun cloud berpindah tangan. File yang Anda percayai menjadi kisah yang Anda ceritakan pada diri sendiri tentang apa yang dulu ada. Walrus dibangun untuk perasaan itu. Ini tidak berusaha menjadi menarik secara visual. Ini berusaha menjadi andal dalam cara yang benar-benar bisa Anda verifikasi.

Walrus adalah jaringan penyimpanan blob terdesentralisasi yang dirancang untuk file besar dan data tak terstruktur seperti file media dan dataset. Walrus bekerja dengan Sui sebagai bidang kendali sementara byte-bite tersebut berada di jaringan penyimpanan khusus. Pembagian ini penting karena memungkinkan blockchain melakukan apa yang ia kuasai, yaitu koordinasi dan akuntabilitas, serta memungkinkan jaringan penyimpanan melakukan apa yang ia kuasai, yaitu menyimpan data berat tanpa memaksa setiap validator untuk membawanya.

I’m going to tell this story the way it actually lands when you spend time with it, not as a checklist, but as a lived walk through a system that is trying to make one fragile part of modern life feel steadier. Most of us have learned to accept that data is temporary even when it is deeply personal. Photos drift into forgotten drives. Project files sit behind subscriptions. Research datasets vanish when a link dies. A creator’s archive can disappear because an account changes, a policy shifts, a company pivots, or a bill misses a payment. Over time that creates a low grade tension we carry without naming. Walrus shows up in that emotional space and says something quietly radical: what if storing large files did not require trusting a single entity to stay kind and stable forever.

At its core, Walrus is a decentralized blob storage network designed in the Sui ecosystem. Blob storage is just a blunt honest phrase for large unstructured files, the kind that make blockchains sweat if you try to put them directly on chain. Videos, images, models, datasets, application assets, whole archives. Walrus is built to hold those kinds of files across many independent storage operators, while Sui is used as the coordination layer that can record verifiable proof that storage happened and can let applications treat stored data as something programmable rather than a loose off chain attachment. That is the first design decision that changes everything: the heavy bytes do not live on the chain, but the truth about those bytes is anchored on chain. If It becomes normal to build this way, then web scale storage can gain blockchain grade verifiability without forcing the blockchain to become a landfill of large files.

Here is how the system works when you slow it down enough to feel the mechanics, because Walrus only really makes sense when you imagine an actual upload happening in real life. You start with a file, which Walrus calls a blob. The client software does not simply push that blob to one server. Instead it encodes it into many smaller fragments that are engineered to tolerate loss. You can think of them like puzzle pieces that are intentionally redundant, but not in the crude way of copying the same file again and again. These fragments are often described as slivers, and the important point is that you do not need every sliver to survive in order to recover the original file later. You need enough. Walrus then distributes these slivers across a committee of storage nodes, a group of operators responsible for holding the data during a given epoch. Once enough nodes acknowledge they are storing their assigned slivers, the network can produce and anchor a proof of availability on Sui. That proof is the turning point. It is the moment the network stops being a collection of best efforts and becomes a commitment you can verify.

This split between the data plane and the coordination plane is not cosmetic. It is the reason the architecture is plausible. Sui is very good at coordination, identity, ownership, state, and verifiable records. Walrus is designed to be very good at storing and serving large volumes of data efficiently. By keeping the big data off chain and keeping the accountability on chain, Walrus aims to get the best of both worlds. The tradeoff is complexity. Two planes means two kinds of failure modes and two kinds of recovery stories. You now have to reason about committee changes, node churn, network conditions, and the exact shape of proofs. But the alternative is worse. The alternative is pretending blockchains can cheaply store massive files in replicated form and hoping fees and throughput do not become a wall. Walrus chooses complexity in service of practicality.

The encoding approach inside Walrus is one of its quiet signatures. It is described through a design called Red Stuff, which uses a two dimensional erasure coding approach. You do not need to memorize the math to understand the human meaning. Red Stuff is Walrus choosing a kind of resilience that is not based on waste. Instead of storing three or five full copies of a file, it stores encoded redundancy so that the network can tolerate missing pieces and still reconstruct the original. This matters because decentralized storage is only truly decentralized when it can be affordable enough for many people to use, not just for those who can pay for brute force replication at scale. Efficiency is what keeps the door open. It also changes recovery. When something breaks, a good system should not force you to pull the entire blob again from scratch. A good system should heal proportionally. Walrus is designed with this healing mindset. It expects nodes to go offline. It expects churn. It expects parts of the puzzle to disappear temporarily. And it tries to remain calm anyway.

That calmness depends on incentives. A decentralized network does not stay alive because we want it to. Storage operators need a reason to show up every day with bandwidth and disks and operational discipline. Users need predictable pricing and a payment flow that feels closer to a utility than a roulette wheel. This is where WAL comes in. WAL is the token that underwrites the storage economy. It is used to pay for storage for defined periods, and it is also used for staking, where participants delegate stake to storage nodes to help secure the network and align incentives. Rewards flow to operators and stakers for providing service and behaving honestly. Governance uses stake so that changes to network parameters and policies are anchored in community weight rather than in a single operator’s preferences. Walrus also acknowledges the cold start reality with subsidies that support early adoption while the network matures. They’re not pretending demand magically appears at full scale on day one. They’re building an economy that can survive the awkward beginning and still become sustainable later.

If you have ever tried to build a product on decentralized infrastructure, you know the emotional difference between theory and usage. Walrus is not just trying to be correct. It is trying to be usable. That means developers need simple ways to store, reference, and retrieve blobs. It means storage needs to integrate with applications and smart contracts so data can be treated as a first class object rather than an afterthought. When you can programmatically reference stored data, enforce policies around it, and build workflows that depend on its availability, storage becomes more than a bucket. It becomes part of the application’s logic. This is where the Sui integration becomes more than convenience. It becomes a foundation for programmability that many storage networks struggle to offer without bolting on complex layers.

Now let me walk through the real world use cases in the slow step by step way that shows how value is created rather than merely claimed. Imagine a personal vault application, something that feels like a private drive for your life. A user uploads family photos, passports, contracts, scans, creative drafts. In a typical web2 design, all of that lives under one provider’s control. Even if you have strong encryption, you still depend on that provider’s uptime, billing systems, and policies. In a Walrus based design, the app encodes the user’s files into slivers and spreads them across independent operators. A proof of availability is anchored. The user’s data is no longer “hosted by a company” so much as “held by a network.” The value shows up when something goes wrong somewhere else and the user opens the vault and their files are simply there. No drama, no scramble, no sudden migration. That is not a viral moment. It is a deeply human one, because it replaces anxiety with a small ordinary sense of safety.

Now imagine a website or application frontend that needs to live for years, not just for a launch. Walrus can store the static assets that make a site work, and those assets can be served by the network. This matters because so many applications depend on frontends that are silently centralized. An on chain contract is not very useful if the interface disappears. When a site’s assets are stored in a decentralized blob network, the interface becomes harder to quietly remove. And even when censorship resistance is not the primary goal, resilience still matters. It means a creator or community does not have to worry that one account suspension or one provider outage will erase their presence. If It becomes easy to publish and keep publishing without depending on a single vendor, the internet becomes less brittle.

Now imagine AI and data intensive workflows. Anyone who has built with datasets knows the pain of link rot and quiet replacement. Datasets disappear. Files get moved. Checksums change. Provenance gets muddy. When Walrus stores a dataset as blobs and anchors availability in a verifiable way, it becomes possible to reference data over time with more confidence. Then add encryption and access control layers and you get something even more interesting: public infrastructure that can hold private data without exposing it. This is where the phrase data markets starts to make sense. Not speculative markets, but practical markets where people can share and sell data responsibly, knowing that access can be controlled and audited, and knowing that the underlying blobs are available and consistent. We’re seeing the outline of a world where data can be treated as a durable asset rather than a fragile file.

Of course, any honest deep dive has to sit with risks, because decentralized storage is the kind of promise that collapses if you wave away the hard parts. The first risk is complexity. Erasure coding, committee coordination, epoch changes, healing behavior, economic incentives, and proof verification create more surfaces for bugs and edge cases than simpler systems. Complexity also increases the burden on implementers and auditors. The second risk is economic tuning. Token incentives can be too weak or too strong. Subsidies can help early adoption but must eventually taper without breaking the operator ecosystem. Pricing stability is hard when tokens are volatile. The third risk is governance. Decisions about parameters, penalties, and roadmap changes shape the network’s credibility. Poor governance can damage trust even if the technology is solid. The fourth risk is stewardship and abuse. A storage network can be asked to hold content society struggles with, and systems need clear policies and mechanisms for dealing with harmful use without undermining the core reliability the network exists to provide.

And still, I see strength in the willingness to face these risks early. When a project names its hardest problems, it builds a culture that can survive reality. It creates room for iteration, for hardening, for building operational discipline. It stops being a fantasy and becomes a system that can endure. That is the kind of strength that compounds.

Walrus also carries a kind of momentum that is easy to underestimate because it does not always look like hype. The network has moved through phases that suggest steady accumulation: early previews where real data is stored, testnets where community participation expands, and mainnet where the operator set and economic loops begin to live. In a storage network, the most meaningful momentum is not an announcement. It is usage. It is data that stays available. It is operators who keep serving. It is developers who keep building. Over time, that steady rhythm becomes reputation.

Now let me end with the warm future vision, because the future that matters most here is not the one where everyone becomes an expert in storage protocols. The future that matters is the one where storage stops being a silent source of dread. A student keeps their portfolio safe across years. A creator keeps their work accessible even when platforms change. A community archive outlives its organizers. A small team ships without building their entire business on the fragile assumption that one vendor relationship will stay perfect forever. It becomes normal to build things that last longer than a product cycle. It becomes normal to keep what you made.

I’m not claiming Walrus will magically fix every problem the internet has with permanence and control. But I do think there is something quietly hopeful in its direction. It is trying to make data less dependent on the lifespan of any one company and more dependent on a network designed to keep its commitments. If It becomes the kind of infrastructure people can rely on without thinking, then one day someone will open a file they stored years ago and it will simply be there. And that small ordinary moment will feel like a miracle, not because it is flashy, but because we have gotten so used to losing things.

#Walrus @Walrus 🦭/acc $WAL

Dusk dimulai pada tahun 2018 dengan keyakinan yang tenang dan hampir keras kepala bahwa keuangan tidak harus memilih antara privasi dan tanggung jawab. Sejak awal proyek ini bertujuan untuk membangun blockchain Layer 1 untuk dunia di mana hukum masih ada, lembaga masih penting, dan manusia masih layak mendapatkan martabat. Alih-alih mengejar kebisingan, Dusk memilih untuk membangun infrastruktur. Jenis yang tidak berteriak, tetapi mendukung. Jenis yang tidak menuntut perhatian, tetapi membangun kepercayaan.
Di tengah Dusk ada gagasan sederhana yang membutuhkan bertahun-tahun untuk menjadi kenyataan. Nilai harus bisa bergerak dengan cara yang berbeda dalam satu rantai yang sama. Kadang uang perlu terlihat. Kadang perlu dilindungi. Dusk menciptakan ruang bagi keduanya. Di satu sisi ada Moonlight yang terasa akrab. Saldo bisa dibaca. Transaksi bersifat transparan. Di sinilah pelaporan berlangsung. Di sinilah akuntabilitas publik masuk akal. Di sisi lain ada Phoenix yang terlindungi dan sunyi. Di sini bukti pengetahuan nol melakukan pekerjaan berat. Jaringan dapat memverifikasi bahwa transaksi valid tanpa mengungkapkan pihak terlibat, jumlah, atau riwayatnya. Dua bentuk nilai ini hidup berdampingan di satu lapisan penyelesaian. Mereka tidak merusak ekosistem. Mereka bertemu pada fondasi yang sama dan mencapai finalitas yang sama. Yang berubah bukan kebenaran dari transaksi, tetapi seberapa besar kebenaran itu yang harus dilihat dunia.

Ketika saya pertama kali duduk bersama Walrus, saya tidak langsung membahas harga token. Saya mulai dengan rasa takut kecil dan biasa yang hidup di balik banyak hal di internet modern. Tautan menghilang. Akun terkunci. Perusahaan berubah pikiran. Arsip lenyap. Saya bertanya-tanya apakah hal-hal yang saya buat, bayar, atau bergantung padanya masih ada ketika saya kembali. Walrus dibangun di sekitar rasa takut ini, tetapi merespons dengan cara yang mengejutkan secara realistis. Ia memperlakukan data besar sebagai sesuatu yang harus tetap tersedia di jaringan dengan bukti dan insentif, bukan dengan kepercayaan buta. Ini adalah protokol penyimpanan terdesentralisasi yang dibangun di atas Sui dan dirancang agar aplikasi dapat mengelola blob besar menggunakan kontrak pintar.

Saya akan berjalan melalui Walrus dengan cara yang akan saya jelaskan kepada teman yang peduli sedikit tentang slogan dan lebih peduli apakah sesuatu benar-benar bisa bertahan ketika kehidupan menjadi kacau. Walrus adalah jaringan penyimpanan terdesentralisasi yang dirancang untuk objek biner besar yang disebut blob, dan ia menggunakan Sui sebagai tempat di mana koordinasi dan bukti penyimpanan berada. Model mental yang paling sederhana adalah ini. Data Anda tidak ditempatkan di satu lokasi. Data tersebut diubah menjadi banyak bagian, dan jaringan membuktikan secara on-chain bahwa jaringan telah menerima tanggung jawab penyimpanan selama periode berbayar.

Saya akan memulai dengan satu-satunya hal yang benar-benar penting dalam jaringan penyimpanan, yaitu apa yang terjadi ketika Anda mencoba menyimpan sesuatu yang nyata. Bukan sekadar gagasan tentang file, tetapi sebuah blob berat dan sulit seperti arsip video, dataset, tumpukan gambar, atau kumpulan PDF yang panjang. Walrus menyebut jenis muatan tak terstruktur seperti ini sebagai blob, dan kejutan pertama adalah bahwa Walrus tidak menangani blob seperti layanan penyimpanan cloud biasa, dan juga tidak menanganinya seperti blockchain. Sebaliknya, blob diubah terlebih dahulu sebelum disimpan. Klien mengkodekannya menggunakan Red Stuff, yang digambarkan oleh Walrus sebagai protokol enkoding erasure dua dimensi. Dalam istilah sederhana, blob Anda berubah menjadi banyak fragmen kecil yang dibentuk sedemikian rupa sehingga jaringan dapat membangun kembali aslinya nanti, bahkan ketika sebagian besar node hilang atau tidak stabil. Makalah akademik Walrus sangat langsung tentang apa yang ingin dicapai oleh ini. Dikatakan bahwa Red Stuff mencapai keamanan tinggi dengan hanya faktor replikasi 4,5x, sambil juga mendukung pemulihan otomatis data yang hilang dan melakukan pemulihan dengan bandwidth yang sebanding dengan data yang hilang, bukan menyalin seluruh blob lagi. Makalah ini juga menyoroti sesuatu yang mudah dilewatkan sampai Anda melihat jaringan nyata yang gagal. Red Stuff dirancang untuk mendukung tantangan penyimpanan dalam jaringan asinkron, sehingga penyerang tidak dapat memanfaatkan keterlambatan jaringan untuk terlihat jujur tanpa benar-benar menyimpan data.

Dusk dimulai pada tahun 2018 dengan arahan yang terasa sangat manusiawi di ruang yang sering kali terasa bising. Dusk bertujuan untuk membangun blockchain layer satu untuk keuangan yang diatur, di mana privasi tidak diperlakukan sebagai celah dan kepatuhan tidak diperlakukan sebagai sesuatu yang terakhir. Saat Anda membaca bagaimana Dusk mendefinisikan tujuannya, kata-kata terus kembali pada keseimbangan yang sama. Privasi bagi pengguna dan lembaga. Kemampuan audit bagi regulator dan infrastruktur pasar nyata.

Saya akan berjalan melalui Dusk seperti yang saya lakukan saat berjalan melalui kota baru. Perlahan. Dengan memperhatikan bagaimana hal-hal saling terhubung. Karena bagian yang menarik dari Dusk bukanlah fitur tunggal. Tetapi cara bagian-bagian tersebut dimaksudkan untuk bekerja sama dalam situasi keuangan nyata di mana aturan ketat dan informasi bersifat sensitif.