# FY1 S1 Prominence App Plan # Philosophy For the first time in human history, we have a chance to give expressive power to all people. Today, consumers rely on the expressive power of few big brands. A person can only express one’s own unique experiences and interests using unoriginal options that all other people are also using. When a person finds that no existing products can satisfy what they desire or express who they are, the person feels powerless to be an outsider in a world made by others for others. We want to remove this powerlessness from this world. Junwon Prominence will give people the expressive power of Art Generation AI. People will be able to clearly articulate what they want so that they can easily find it or make it. When a person finds no satisfying products, they will instead feel powerful to have an opportunity to propose a great new idea for the first time in human history. The world will belong to people. # Product Junwon Prominence is a shopping app that always satisfies. If a user fails to discover a satisfying design, then the user can use the power of AI to generate a rendering and propose an idea for the first time in human history. Then, the user can also share it with other users to get help on finding it or making a similar design. Junwon Prominence gives users the power to express dreams and shape the world. ### Communication Find and propose inspiring designs on Junwon Prominence! You too are a {Fashion | Furniture | Car | Architecture | Handbag | Accessory | Shoe} designer. Join millions of us who are using AI to make photorealistic renderings of our ideas! Enter Now. Check out amazing designs made by people just like you on our Spellbind community. Make and share renderings of your own creative designs using our Spellcast generator. Give and get recommendations on how to find or make similar products. ### Composition ![Spells.png](FY1%20S1%20Prominence%20App%20Plan/Spells.png) Spellbind, Spellcast, and Spellbook compose one holistic Spell: - Users can discover interesting arts on Spellbind, then load them to Spellcast or save them to Spellbook. - Users can generate interesting arts on Spellcast, then share them on Spellbind or save them to Spellbook. - Users can collect interesting arts to Spellbook, then share them on Spellbind or load them to Spellcast. ### User Story Alex opens the Junwon app Home Discovery Feed and finds interesting sunglasses, snow globes, and tea cups. Alex thinks Roxana will appreciate a tea cup gift. Alex searches “tea cups”, then adds positive tags and negative tags to guide the exploration. After finding none quite captivating, Alex decides to propose a new design. Using Spellcast, Alex renders a black tea cup with Swarovski crystals to resemble the night sky, remembering that Roxana likes to drink tea in the evenings while stargazing. Alex posts the tea cup on Spellbind, then falls asleep. Alex wakes up to find a pile of notifications. Users love the design. Commenters discuss ideas like connecting the stars to portray constellations. Alex sees that the “bring to reality” button is activated, and taps on it. Alex sees that one commenter shared a link to a site where Alex can buy a pretty similar cup for $25. Another commenter offered to make the exact design for Alex for $50. The last commenter suggested that Alex tries making the cup with a $15 craft kit following a tutorial. Alex contacts the artisan to get the tea cup custom-made. Roxana loves the gift. Roxana has many tea cups, including ones from famous brands that are much more expensive, but the tea cup that Alex made for Roxana is priceless. It has a story. Roxana enjoys tea that night using the gift from Alex, admiring its design, recollecting the night Roxana mentioned to Alex about drinking tea while stargazing. # Defense The Generative AI technology is rapidly growing, and we can only expect we will meet unexpected challenges. We anticipate these topics will be critical to protecting our users amidst such violent changes. ### Focus on products Spellbind will focus exclusively on products. We will direct users interested in other topics, such as renderings of humans, artistic illustrations, politics, and memes, to find or create communities elsewhere. If they need our technology to create and operate a great community, we will provide it. But we will moderate any topics other than products to protect our users, who are coming to Spellbind to find products, from unwanted distractions. We will publish this policy and set up a moderation process to let users report violations. We will take down violations and ban users who repeatedly violate our policy. ### Prohibit misinformation We will prohibit users from distributing misleading imitations of famous brands and famous designers. We will permit otherwise legal and ethical emulations. To enforce our policy, we will research ways to identify AI-made renderings using techniques such as watermarking and AI detection models. ### Monitor dependency risks We will monitor following dependency risks to make sure we can provide stable service to our customers. | Dependency | Risk Assessment | Mitigation Plan | | --- | --- | --- | | Server Capacity | Server capacity is a commodity in a perfect competition market. We do not expect it will be possible for cloud service providers to exclude us from server capacity or to unreasonably hike the price. | None needed. Building a 1P data center is expected to cost over $1B. | | App Stores | Apple and Google have the ability to, but no motivation or justification to, cut us off from app stores. | If cut off from app stores, we will distribute on the web. | | AI Models | Model providers have the ability to stop providing models to us and therefore our users. There is no reason for them to do so since we are opening up our AI Store to all developers. Model providers will be able to engage in business with us and our users on our platform. Even if some model providers choose to exclude us from superior AI models, we expect users will choose a whole catalog of diverse models on our open AI Store over a single AI model. | If we end up finding that it is important to have access to superior AI models, then we will partner with a 2P model provider or train a 1P model. Since there exist many model providers but not yet many discovered profitable applications of AI, we expect we will be in a strong position to form partnerships with model providers. Training a 1P model is expected to cost at most $1M and at most 1 month. Capital and compute capacity are unimportant commodities. Training data is the important specialty. | # **Finance** For Spellbind, expected expense is negligible in comparison to expected income. We expect to generate $5 of revenue per daily active user per year from advertising business model. For Spellcast, we estimate each daily active user will cost us $142.95 per year. This high cost is due to the high price of computing capacity needed for AI model inferencing. We will price Junwon Artist Membership for accessing Spellcast at $30 monthly price to pay for the cost of service. # Plan Junwon Prominence will forever change the way our economy works. Today, there is no effective way for producers to understand desires of consumers. Consumers can only hope that producers will offer what they want, and producers can only hope that consumers will want what they offer. Junwon Prominence will give consumers the power to clearly express their desires. Junwon Prominence will give producers the power to reliably deliver the products that consumers really want. Perhaps manufacturing will become an infrastructure utility. Perhaps I will get to design a jacket in the morning, then pick it up from a local shop in the evening. Perhaps that I chose to wear red and white will express me better than will that I was able to wear an Hermes. As for us, Junwon Prominence will give us the power to give even more power to people: - **AI Platform → App Platform:** To fulfill the long tail desires of users that we cannot fulfill with our 1P models, 1P communities, and 1P apps, we can open up our platform to 3P modelists, 3P moderators, and 3P developers. What users can get done on Junwon app will grow from shopping to any images made using Art Generation AI, then any contents made using Generative AI, then any AI, then any apps. - **Compute Demand → Compute Supply:** Storing and processing data at AI-scale remains costly. This is great for us since we prefer credit to cash. That users trust us to store and process so much data will let us negotiate favorable terms with cloud service providers for the benefit of our users. If vendors cannot meet the demands of our users and developers, we will also be able to build data centers and manufacturing plants to supply server capacity and electronic devices. When we were able to choose to sit idly and do no work, we set out to do hard work to help people in impactful ways. I expect Junwon Prominence will be a hard work that unlocks the opportunity for us to do even more hard work for people. We have a world to build, and it all starts with Junwon Prominence. And while we have a Spell to build, we will not keep our users waiting while we are developing it to the fullest of our vision. We will start delivering our harmony as an arpeggio as soon as our first note is ready. After I finish writing this document, we will flare out to explore the solution space, then carve it down to the maximum possible distillation so that we can meet our users in minimum possible time. Then we will grow from there.