我在最后放入了使用notion的自动翻译，因为我的英文不够好，如果翻译的水平不行，那只能借助翻译模型再进行推理一下了，因为我的确无法判断英文翻译的结果，十分抱歉）
(I have put in an automatic translation using Notion at the end, because my English is not good enough. If the level of translation is not up to par, then I can only rely on the translation model for further inference, because I really can't judge the results of English translation, I'm very sorry)
前言
我是一个热衷于使用SD模型的中国人，是推崇AI进入人们未来生活的博主，我的名字叫素人极客。他意味着，普通人也可以通过AI成为一名十分厉害的英雄，素人就是普通人，极客就是十分厉害的人，这是中文的魅力，如同stable diffusion一样精巧。
得知2024年Stablity AI的巨大变动，开源模型后企业经营陷入新困难中。我们曾经期盼已久的AI改变人们的生活方式，好像突然出现了一点火苗，就被人的功利追逐熄灭了。我们不得不说，一个绝好的事物出现，往往面对的挑战是旧的利益集团的围剿，正如中国古代第一个变法者——古代秦朝的商鞅，他推动秦国变法，让更多普通人能够加官进爵，使传统王公贵族垄断的上升阶级受到了史无前例的影响，从而被老贵族们杀死在车马之下，使用的是中国历史上最严厉的刑罚——车裂。这就是真正改变者遇到的最大挑战，是要扛住最猛烈的反扑，才能成就伟大，否则就会以最残忍的方式死去。
在中国的近代，也有一个伟大的人物，他叫毛泽东，他用了最简洁的方式表达了出来，让很多中国人津津乐道——打土豪，分田地。就是让旧的贵族失去他们的地位，将他们的财产和土地平均的分给普通人。这也是中国逐渐崛起的几十年，最成功的立足之本，他们与最普通的人们（包括农民和工人）站在一起，成就了今天中国的经济。
AI的出现，最有利的获益者，就是这些普通人，他们在资源不平等的世界，不断的尝试改变自己的生活，但手中没有权柄，没有利器，也没有太多能投入的资本，他们可能这一代是普通人，下一代也是普通人，但当我看到AI模型开源的趋势来临，我就认为一定要抓住这个机遇，因为再没有别的机会，可以让我与世界顶级人才站在同一起跑线，是的，我就是这样一个不懂代码，没有审美，没有专业技能，只能辛苦重复工作的普通人。是SD模型的开源，让我看到了一个新的世界，可以与那些拥有更多资本的人一较高下的世界。我也相信，AI到来的世界，正式这样的世界，是我们每个人都可以用最低成本的方式获取一样的工具，区别于你要开出租车必须要有一辆车，区别于你要成为治病救人的医生就要学习十几年医学。而我通过SD的开源模型，就可以轻松成为世界某个角落的内容创作者，创作我喜欢、我的粉丝也喜欢的作品，并依靠他来养活我和我的家人，还有我计划中的团队成员。
可以说，我们的运气是好的，我们在2022年就看到了SD模型，当然我知道他基于AI模型的多年发展，才进阶到如此成熟，AI模型发展到今天，正是人类探索如何以新的智能改变的拐点。在AI模型出现以前，我们都希望一台电脑就能模仿一个人。而如今我们看到，真正模仿人的，是模型和他所链接的所有软件和硬件，正如我们今天讨论非常热门的全球化趋势。一切的经济，都是你中有我，我中有你。那么，一切的人工智能，也是如此，我们每一个模型的使用者，和Stablity AI的模型生产者，也是你中有我，我中有你。只有达成这样的关系，我们才会命运与共，才能更长久的生存下去。以下，是我基于这样一个背景思考的商业计划，我认为，Stablity AI必须达成与所有使用者、创作者的紧密联系，才能渡过这个2024年，才会实现正向盈利，持久经营。我们要看到未来的世界，看到当下的难题，更要拆解里面的每一个环节，任何一处看不清，都将出现可怕的断链。我认为，过去的Stablity AI并没有看清未来的世界，也就无法解决当下的难题，更无法正确的拆解每一个环节，所以也就没有持续开源的能力。以下的内容，是我酝酿了很久，才想清楚的解决方案，他会和平时我们看到的商业计划差距非常大，并且有点骇人听闻……
我将本篇文章分为未来世界，当下世界，微尘世界，通过三个不同的剖面，来诠释一种我认为的正确理解，为SD模型的开源找到合理的目标、找到正确的对手、找到准确的问题、找到真正的答案。
第一篇章 SD开源的未来世界
可能我们都没想过，SD模型为什么要开源，对于企业的理想主义，AI模型的开源是人类迈向AGI的重要一步，但是现实告诉我们，没有正向盈利的商业项目，根本无以为继，因为开源是破坏规则的存在，是时代需要，却非人类的必须。所以如何看待正确的开源未来世界非常重要，这决定了钱从谁的手里来，钱到谁的手中去。
首先，钱从谁的手里来。
我们都知道，好的产品由消费者买单，然而开源意味着免费提供出去。就像我们让农民去种地，那要给他们一块土地，这个不能收钱，但是很显然卖种子的人要收钱，卖农具的人要收钱，农民种出来的粮食归他们自己，提供土地的人并没有收获。所以如果在古代，这些农民要缴纳粮食税，然而在农业税费逐渐远去的时代，即使优秀如苹果一样的产品，苹果税（闭环购买的系统分佣制度）也在被人们声讨。
那么我们看看古代帝王是怎么做的，当帝王免去一个地区土地税收的时候，往往意味着他们要在其他的土地上找回缺失。要么开拓商贸市场，要么征讨其他的土地，往往只有这两个选择。那么摆在Stablity AI面前的，就是，要么开拓新的商业经营项目，要么去打败正在收费的对手。很显然新的商业经营项目，仍然需要巨大的投资，这很不理智。不过，吃掉对手的市场，倒是已经获得了口碑上的优势和众多普通人的拥护。
那么，谁是那个倒霉的人呢？
有的人可能会说，是不是Adobe公司？我认为这就是没看清SD开源的未来世界导致的，我们现在有了一个寻找对手的想法，我们要抢走他们的市场，那就要详细的推敲，“他”到底是不是对手？！
Adobe公司充其量是一个工具类型的公司，他们销售的工具被几乎所有的设计师使用，并且也拥有非常多的竞争对手，他们的基础处理能力，是SD开源模型不能直接达到的，也就是开源创作者们可以将Adobe公司作为竞争对手，但是Stablity AI公司的对手，绝对不是工具类型的公司。Stablity AI是培育Adobe公司对手的更强大存在。
我们回过头看SD开源模型，一个模型里有几十亿，上百亿数据，承载着世界上众多的图像，甚至未来可以装下我们眼睛可见的所有图像和幻想的图像，要创作出来的是经过加工的图像和视频，SD开源模型将是每个人创作大脑的一部分，因为我们每个人的大脑里都装着自己想创作的画面。所以SD开源模型本身，就是人类大脑的一部分，那么对手，也就是那些人类历史上所有可以独立创作图像的人们。这听起来很恐怖，但这就是SD开源模型的未来——成为人类大脑图像创作的一部分。
我们也曾看到过很多画家出来抵制生成式AI的类似抄袭问题，可是谁都无法阻挡科技发展的车轮，任何阻挡AI前进的人，都将成为尘埃。所以，当生成式AI出现的那一天，人们就要知道，用AI生成图像就是未来人们必须拥有的能力，并在此基础上衍生出更优质的作品，来稳固自己的专业性，持续进步，才是人得以生存的基础。
那么这样看来，过去所有的图像创作者，都将是SD开源模型未来世界的创造者之一，与这些人成为敌人显然是不应该的。我们遇到的难题就是，我们的对手到底在哪里呢？我们应该如何与他们公平的竞争呢？
让我们来思考一个问题，一个人创作一幅画，目的是销售给另一个人，从最早的记录需要，到贵族绘画的需要，到消费画作和影视项目需要，我们已经越过了前两个阶段，当下正是消费画作和影视项目需要的阶段，人们太需要这种文化层面的消费了，这些项目让我们开心、让我们哭泣、让我们拥有快乐和感动，让我们体会到我们作为人的独特之处——带有共鸣的情感。SD开源模型未来给人们带来的不就是这些吗？不就是把这些专业内容的制作，从高协同、高难度的专业处理，变成更简单方式，变成更容易实现的人类幻想吗？
所以，在这个世界上，谁在以高协同、高难度生产图像类内容的项目，谁就将是SD开源模型的对手，他们可以是迪士尼，可以是好莱坞，可以是索尼，可以是暴雪游戏。是的，对手都是手握传统资源的大家伙、大企业。他们在不断生产优质的作品，使之成为IP资产，也在应用AI去创作更具想象力的内容。SD开源模型能扶植起来的也只是一部分普通人，这有对抗的可能吗？
我想，谁能看到SD模型未来世界的美好，谁才能有信心去和他们对抗，去争抢更有价值的市场。中国有个非常典型的案例，叫“农村包围城市”，就是说乡村人口远远大于城市人口的时候，满足农村人口的生存需求，反而会更容易成功。那么面对那些大的公司，我们就要和这些普通人一起，创作出带有消费属性的内容，才是更重要的。
至此，我不是很想写太多关于SD模型未来世界的描述。因为我们都知道，SD开源模型的未来想象，是容纳整个世界无边无际的图像创作，即使我们有新的架构，不同于transformer，那Stablity AI的未来也是一直在这条路上，可能就是元宇宙这条路上最重要的一张棋盘，每个普通人都是这里的棋子。Stablity AI帮助这些人在AI的世界里通过优质的内容赚钱，行成良性的循环，把那些公司不能生产的，难以做到的，不愿意去做的，都收拢到Stablity AI的版图里，实现共创的模式，在后面具体的解决方案中，我们再进一步探讨商业的可实现性，目前我们已经找到了合理的对手，发现了未来的可能性，接下来，我们要面临的就是当下世界的难题了。
第二篇章 SD开源的当下难题
随便一个SD模型用户，都能列举一堆难题。相信Stablity AI想的更不会少。千头万绪，我们都要从最基本的根源说起——Stablity AI从没想过用户这件事。可能你们推出了很多更便捷的功能，那么为什么还存在那么多模型放大、背景抠除、图片上色的小平台获得了用户的拥戴呢？正如我们看到的Magnific AI，即使只有两个人，在全世界，也收获了非同凡响的影响力，他们也是使用的SD开源模型作为基础，业务严重与Stablity AI重叠，但却走出了自己的特色。并不是Mganific AI这一家公司拥有的特权。很显然，Stablity AI距离普通人太远了，你们所有的工具，并不是从方便普通用户出发的，而是从建立一个收益窗口出发的，你们没有解决用户的使用问题，而只是在解决自己的生存问题。这看似是一个逻辑的解释，却是一家公司能否拥有竞争力的出发点。
比如，Civitai聚拢了世界最庞大的SD开源模型的使用者，Github聚集了最多开源插件的创作者，Huggingface存放了世界最多开源SD相关模型，Reddit活跃着全世界最靠谱的SD开源模型的讨论者。可大多数普通用户连SD模型的基础使用教程，都很难在你们的官方网站上找到，我知道你们有，可是你们在为了留存用户而开放更多的信息吗？显然这一点，你们并没有做好，我相信你们已经意识到。全世界那么多的平台，那么多的人在依靠SD开源模型创作和赚钱，而Stablity AI却根本没有给这些人带来一点点附加的优质内容，更别说服务了，在你们认为除了模型任何服务都该收钱的理念中，其他平台早已根据自己的模型，让SD开源模型成为凝聚人气的利器。
我认为，在这件事上你们应该检讨。我们知道任何好的行业，好的项目，好的产品都应该有清晰的标准，如果一台电脑没有详细的参数标准，那就是不合格的产品。为什么Stablity AI不能有自己的标准？这个标准可以是对一个模型的认可，这个标准也可以是对一张图像的认可，这个标准更可以是对开源插件创作者的认可。而这些都被你们放弃了，你们本该是最伟大的标准制定者。
MIdjourney由自己制作标准，来形成闭源的创作方式，他们选择的方向无非就是三块：第一，不断加强模型的艺术标注；第二，简化使用者的操作标准；第三，丰富社区的互动标准。进而通过综合的服务来收取订阅费。而Stablity AI开源了SD模型后，就一切放手不管了，天下没有谁愿意为一个开源的动作买单，更不会因为一个企业的好心而持续捐献自己的爱心。商业的基础永远是公平的交易。
没有自己制定的标准，没有公平的交易，是我看来，Stablity AI当下面临的最大难题。人们使用你们的工具频次并没有那么高的，你们的官网推出的功能很多替代软件和网页应用都可以实现，这个从你们的商业应用的初心，就已经开始偏离了轨道，因为你们创作这些东西，仅仅是想自己多赚钱而已。
—但是这个时代，已经是协作去中心化，但资源依赖中心化的共赢时代，表面的去中心化是个人作为超级个体的崛起，深层的去中心化，反而像神经网络一样，让别人持续获利，才是Stablity AI这样的企业生存的基础。或许你们会说，制作标准、优化模型、搭建社区都需要巨大的成本。可是这个世界上还有一种共创，是你提供舞池让别人跳舞，跳舞的人获得拥护者，舞池收取吃喝相关的费用。没错，这很像酒吧，恰恰是这样的酒吧，才能让有魅力的更有魅力，这才是去中心化，把自己的舞台变成所有人的舞池的逻辑。
舞台模式转向舞池模式，独舞和观众呆板的观看，已经变成共舞与碰杯
其他所有的难题在我看来，都是因为核心驱动力不能解决根本的难题，才带来的连锁反应，只要整个商业逻辑上，能把最隐形的根源问题解决，就可以扳回所有的败局。即重塑标准，升级交易，带所有开原创作者赚钱，与所有普通人涌进自己开辟的舞池里跳舞，与全世界的SD开源模型使用者做朋友，通过使用者与全世界的人做交易。
重塑标准，升级交易，带所有开原创作者赚钱，这就是当下世界的解决方案。
第三篇章，微尘世界里的微操作
知道了战略层面上的解决方案，战术方面就要可执行的层面。
那么就要从细枝末节里，找到突破点。
首先，是重塑标准。
在一年的SD开源模型使用过程中，我们越来越多的遇到对模型的判断标准，对使用算法、参数的衡量标准的困局中，甚至在学习方向也出现了偏差，在我的身边更多的人都在询问一个问题：你有好的模型推荐吗？
仅仅是这样一个问题，就困扰了无数人无数的时间，所有我们能看到的，都是在分类中和大众的评分中，有的时候因为一个知名博主的推荐，某个模型就会快速上升，所以目前凡是关于SD开源模型应用的基础标准，仍然是模糊的。所以怎么制定标准、谁来制定标准、标准发在哪里、标准能带来什么收益，是非常值得探讨的。以我们现在观看其他一些社区的做法，标准是按照不同分类去划分，用户自主点评，很显然埋没了很多优质的模型，也充斥着大量的色情内容。以Civitai平台为例，只要区别于这个社区的风气，给出一个带有官方特色的平台机制，就可以快速聚集人气。
聚集人气的办法
成立专业版、应用版、生成版三大社区，通过不同的世界级活动，与知名企业合作，行成有效的企业、用户、普通人的联通，提升整个社区的实用性，增加使用者们的使用频次，社区收益多种多样，甚至可以为了帮助一部分人节约时间，推出每周优质模型、优质prompt、优质企业IP孵化等活动的周刊月刊，实现流量向的增长。
如果你们认为这种方法你们已经尝试过，那么在此基础上，AI微型创业体将会是一个很有吸引力的众筹方式。每个插件创作者、模型创作者，往往是制作完成后，才能获得拥护者们提供的“一杯咖啡”，这失去了期待感和获得感，如果通过了标准的机构和个人，进行预发布公开众筹，大家在期待新版本的时候更会有意愿打赏，开源创作者也会更加有动力。
每个版块甚至都可以分发给愿意参与维护的管理者，通过他们的号召力，来维护一个合理又融洽的环境。过往一个API和生成付费的方式，无法持久维持Stablity AI的高频词使用运营。其实一个论坛对你们来说并不是难处，创作什么样的标准是第一解决的事情。在我们当下的世界里，创作者们从SD开源模型借势，Stablity AI又从知名的开源创作者借势，轮值当管理者，为每一个新的IP，为每一个新的创作提供展示空间，有规划的帮助更多人更便捷的使用SD开源模型，才能收获最稳定的流量，而这个世界，流量是最难得的东西，也是一切的基础。
其次，是升级交易
对于任务的模式，无论世界的各个地方可能都有不同的交易模式。优质的任务发布与任务承揽是一门关键的生意。很多平台也因为劣质的低价竞争，成为了不良供应商的聚集地。Stablity AI的交易模式应该有所升级，成立一个信用体系，提供一个SD交易币，甚至是引入合理的交易货币，来推出基础套餐的要求，促进开源生态的一些高级交易，在Stablity AI平台上实现。这或许是被你们开会很多次淘汰的想法，并不完全算一个创意，于是我们应该重新审视，什么是AI开源领域的交易。
交易的本质是买和卖。两者之间基本缺乏信任的基础。越来越多的商家其实需要带有商业属性的AI产品和AI服务，那么谁能提供产品和服务并不重要，重要的是最终的作品，需要成为必须在平台展示的内容，如果选择不在平台上展示，将失去一部分交易分佣的优惠。交易越多的供应商（个人），将会获得越多的分佣优惠减免，其展示的内容也会接受大众的监督，一旦出现作弊行为或大众点评不及格行为，将扣除更多的分佣。
这个时候，可能你们还会觉得有些天真，因为用户是最难管理的，尤其是开源社区类用户。Stablity AI如果参与社区，将要投入更大的精力，但是如果把这种任务，变为Stablity AI与企业IP共同发布的任务呢？比如某家冰箱企业，想制作一套新上市冰箱的作品，企业通过年度框架协议或者单一项目合作的方式，将品牌推广连带任务发布的方式，全权委托给Stablity AI，我想，这将是颠覆未来广告行业生存的一种方式。能够让Stablity AI掌握跨越时代的企业AI应用探索，这种模式，是和众多企业共建的，也必将是有效的社区平台吸引力。
最后，是带所有开源创作者赚钱
就像各类互联网企业，硬件企业的开发者大会一样，Stablity AI也拥有带领开发者，甚至模型创作者、图像创作者赚钱的能力的。我相信这也并不是新奇的想法。但是能做到带大家赚钱，还必须有一个前提——认证。类似于推特的会员认证。不过我们的认证，是经过官方对效果认可的认证，而这个评判人，必须是Stablity AI。
与此同时，要定期推出合理的榜单，包括但不限于
模型榜单
开发者榜单
工作流榜单
美术效果榜单
民间技术文献榜单
行业分类榜单
并给出各行业的指导标准，明确表示不同的人群，能够在不同的领域，更胜任某个方面的工作，并为他们提供一定的合作机会，推出一个互信的交易平台，比如中国非常知名的“阿里巴巴”，他让失信的市场变得更加活跃，只是从中抽取部分的佣金，可被托付的“信任”，是他们致胜的关键，我想，这个在数字资产上，对每一个开源创作者都非常的重要。
具体的实施举措，我相信贵公司一定有自己的判断，我只是依靠我自己的见解，来表达我希望Stablity AI能够走的更远，能够拥有一个可喜的未来，并且和人们一起创建一个崭新的开源世界，成为人们心中非常伟大的企业。
Distributed Decentralized IP World Based on the SD Model - Business Plan
Foreword
I am a Chinese person passionate about using the SD model, and a blogger advocating for AI to enter people's future lives. My name is Su Ren Geek. It means that ordinary people can also become a very powerful hero through AI. Su Ren is an ordinary person, and geek is a very powerful person. This is the charm of Chinese language, as delicate as stable diffusion.
Knowing the huge changes in Stablity AI in 2024, and the difficulties in enterprise operation after open source. We had long hoped that AI would change the way people live, as if a spark suddenly appeared and was extinguished by the pursuit of human utility. We have to say that a great thing often faces the challenge of the siege of old interest groups, just like the first reformer in ancient China - the ancient Qin Dynasty's Shang Yang, who promoted Qin's reform and allowed more ordinary people to be promoted. As a result, the traditional aristocracy's monopoly on the rise was unprecedentedly affected, and was killed under the horse by the old nobles, using the most severe punishment in Chinese history - car cracking. This is the greatest challenge to the real changer, and you have to withstand the most violent counterattack to achieve greatness, otherwise you will die in the cruelest way.
In modern China, there is also a great figure, his name is Mao Zedong, his chosen way, he expressed it in the simplest way, making many Chinese people relish - fighting the landlords, dividing the land. That is to let the old aristocracy lose their status, and their property and land are evenly divided among ordinary people. This is also the most successful foundation for China's rise over the past few decades. They stand with the most ordinary people (including farmers and workers) and have achieved today's Chinese economy.
The emergence of AI, the biggest beneficiaries, are these ordinary people. They are constantly trying to change their lives in an unequal world, but they don't have power, no sharp tools, and not much capital to invest. They may be ordinary people in this generation, and the next generation is also ordinary people. But when I saw the trend of AI model open source coming, I thought I had to seize this opportunity, because there is no other chance to let me stand on the same starting line with the world's top talents. Yes, I am such an ordinary person who doesn’t understand code, has no aesthetics, has no professional skills, and can only work hard repeatedly. It is the open source of the SD model that allows me to see a new world, a world where I can compete with those who have more capital. I also believe that the world that AI brings is such a world, where each of us can use the lowest cost to obtain the same tools, different from the fact that you have to have a car if you want to drive a taxi, different from you becoming a doctor. You have to study medicine for more than ten years. And I can easily become any content creator in the world through the SD open source model, create works that I like, my fans like, and rely on him to support me and my family, and my planned team members.
It can be said that our luck is good. We saw the SD model in 2022. Of course, I know that it has evolved to such maturity based on the many years of development of the AI model. The development of the AI model to this day is the turning point of human exploration of how to change with new intelligence. Before the emergence of AI models, we all hoped that a computer could mimic a person. Now we see that what really mimics people is the model and all the software and hardware it links to, just like the very popular trend of globalization we are discussing today. All economies are you have me and I have you. So, all artificial intelligence is also the case. We, every model user, and the producers of the Stablity AI model, are you have me, and I have you. Only by achieving such a relationship can we have a common destiny and live longer. Below is the business plan I thought about based on such a background. I believe that Stablity AI must establish a close connection with all users and creators in order to pass 2024, to achieve positive profit and lasting operation. We want to see the future world, see the current problems, and even break down every link inside. Any unclear place will form a terrible disconnection. I think that the past Stablity AI did not see the future world, so it was unable to solve the current problem, and it was unable to correctly decompose each link, so it did not have the ability to continue to open source. So, the following content is the solution I have brewed for a long time, and I have thought clearly. He will be very different from the business plan we usually see, and it is a bit shocking...
I will divide this article into the future world, the current world, and the micro world, to interpret a correct understanding that I think, to find a reasonable goal for the open source of the SD model, to find the correct opponent, to find the accurate problem, to find the real answer.
Chapter One The Future World of SD Open Source
Maybe we never thought, why should the SD model be open source? For the idealism of the enterprise, the open source of the AI model is an important step for humans to step towards AGI, but reality tells us that there is no positive profit business project, it is impossible to continue, because open source is a disruptive existence, It is the need of the times, but not the necessity of humanity. So how to look at the correct open source future world is very important. This determines where the money comes from and where the money goes.
First of all, where does the money come from?
We all know that good products are paid for by consumers, but open source means free. It's like we let farmers cultivate, we have to give them a piece of land, this can't be charged, but it's obvious that the people who sell seeds want to charge, the people who sell farm tools want to charge, the grains grown by farmers belong to them, and those who provide the land There is no harvest. So if in ancient times, these farmers had to pay grain taxes, but in the era when agricultural taxes are gradually going away, even excellent products like Apple, the Apple tax (a closed-loop purchase system commission system) is also being people. Criticized.
So let's see how the ancient emperors did it. When the emperor exempted a region from land tax, it often meant that they had to find the missing on other lands. Either open up a commercial market, or conquer other lands, there are often only these two choices. The problem facing Stablity AI is that either open up new business projects, or defeat the charging opponent. It is clear that new commercial projects still require huge investment, which is not wise. However, eating the opponent's market is already an advantage in word of mouth and has won the support of many ordinary people.
So, who is the unlucky one?
We all know that good products are paid for by consumers, but open source means providing them for free. Just like we let farmers cultivate the land, we have to give them a piece of land, which can't be charged for, but obviously the people selling seeds and farming tools need to charge. The crops grown by the farmers belong to them, and those who provide the land do not reap any benefits. So in ancient times, these farmers had to pay grain taxes, but in an era where agricultural taxes are gradually disappearing, even excellent products like Apple products are being criticized for the "Apple Tax" (a closed-loop purchase system commission system).
So let's take a look at what the emperors of ancient times did. When an emperor exempted a region from land tax, it often meant that they had to make up for the loss on other lands. Either they opened up commercial markets, or they conquered other lands; there were often only these two choices. So what lies ahead for Stablity AI is either to open up new commercial ventures, or to defeat competitors who are currently charging. Obviously, new commercial ventures still require substantial investment, which is not wise. However, eating into the market of competitors, who have already gained a reputation and the support of many ordinary people, does seem feasible.
So, who is the unfortunate one?
Some people might say, is it Adobe Company? I think this is due to not understanding the future world of SD open source. Now that we have the idea of looking for competitors, and we want to grab their market, we need to scrutinize in detail whether "he" is a competitor or not?!
Adobe Company is simply a tool-type company. The tools they sell are used by almost all designers, and they also have a lot of competitors. Their basic processing ability is not something that the SD open source model can directly achieve. Therefore, open source creators can regard Adobe Company as a competitor, but Stablity AI Company's competitors are definitely not tool-type companies.
When we look back at the SD open source model, a model contains tens of billions, hundreds of billions of data, carrying many images in the world, and even in the future it can carry all the images and dream images that our eyes can see. The images and videos to be created are processed images, and the SD open source model will be part of the creative brain of everyone, because each of us has images we want to create in our brains. Therefore, the SD open source model itself is a part of the human brain, so the competitors are those who have been able to create images independently throughout human history. This sounds terrifying, but this is the future of the SD open source model - to become part of the human brain's image creation.
We have also seen many painters come out and resist the similar plagiarism problems of generative AI, but no one can stop the wheels of technological development. Anyone who blocks the advancement of AI will become dust. Therefore, when the day of generative AI comes, people will know that using AI to generate images is a capability that people must have in the future, and on this basis, they will create higher-quality works, to solidify their professionalism, continuous progress is the basis for human survival.
So it seems that all past image creators will be one of the creators of the future world of the SD open source model, and it should not be appropriate to become enemies with these people. The problem we face is, where is our competitor? How should we compete fairly with them?
Let's think about a problem, a person creates a painting with the purpose of selling it to another person, from the earliest need for records, to the needs of nobles to paint, to the needs of consumer paintings and film and television projects, we have already crossed the first two stages, and the current stage is the needs of consumer paintings and film and television projects, people need this kind of cultural consumption too much, these projects make us happy, make us cry, bring us joy and touch, let us experience our uniqueness as human beings - having resonant emotions. Isn't what the future of the SD open source model brings to people just these? Isn't it to take these professional content production from high collaboration, high difficulty professional processing, into a simpler way, into a more easily achievable human fantasy?
Therefore, in this world, who is producing image-like content projects with high synergy and high difficulty, they will be the competitors of the SD open source model, they can be Disney, they can be Hollywood, they can be Sony, they can be Blizzard Games. Yes, the competitors are all big guys, big companies with traditional resources in hand. They continue to produce high-quality works, making them IP assets, and also use AI to create more imaginative content. The SD open source model can only support a part of the ordinary people. Is there a possibility of confrontation?
I think, whoever can see the beauty of the future world of the SD model will have the confidence to fight them and compete for a more valuable market. There is a very typical case in China called "the countryside surrounds the city", which means that when the rural population is much larger than the urban population, meeting the survival needs of the rural population is more likely to succeed. So in the face of these big companies, we need to create more consumer content with these ordinary people.
So far, I don't really want to write too much about the description of the future world of the SD model. Because we all know that the future imagination of the SD open source model is to accommodate the boundless image creation of the entire world, even if we have a new architecture different from the transformer, the future of Stablity AI is still on this road, perhaps the most important chessboard on the road to the metaverse, and each ordinary person is a chess piece here. Stablity AI helps these people make money through high-quality content in the AI world, forms a virtuous cycle, collects what those companies can't produce, can't do, and are unwilling to do, into the territory of Stablity AI, achieves the co-creation mode, and in the following specific solutions, we will further discuss the feasibility of business. At present, we have found a reasonable opponent and discovered the possibility of the future. Next, we will face the problem of the current world.
Chapter 2 The Current Dilemma of SD Open Source
Any SD model user can list a bunch of problems. I believe Stablity AI has even more on their mind. There are numerous issues, all of which originate from one basic root - Stablity AI has never considered the user aspect. You may have introduced many convenient features, but why are there so many small platforms providing model enlargement, background removal, and image coloring that have won user support? Just like Magnific AI that we see, even with only two people, it has gained extraordinary influence worldwide. They also use SD open-source models as the foundation, their business overlaps significantly with Stablity AI, yet they have carved out their own unique features. It's not a privilege exclusive to Magnific AI. Clearly, Stablity AI is too distant from ordinary people, all your tools are not designed for the convenience of ordinary users but rather from the perspective of establishing a profit window. You haven't solved the user's usage issues, you're only solving your own survival problems. This seemingly logical explanation is the starting point for a company's competitiveness.
For example, Civitai has gathered the world's largest users of SD open-source models, Github hosts the most open-source plugin creators, Huggingface stores the world's most open-source SD-related models, and Reddit fosters discussions among the world's most reliable SD open-source model users. But most ordinary users can hardly find basic tutorials for using SD models on your official website. I know you have them, but are you opening up more information to retain users? Clearly, you've not done well in this regard and I believe you've realized it. There are so many platforms and people worldwide relying on SD open-source models for creation and making money, but Stablity AI has not provided these people with any additional quality content, let alone services. In your belief that anything other than models should be charged for, other platforms have already used their models to make SD open-source models popular tools.
I think you should reflect on this issue. We know that any good industry, project, or product should have clear standards. If a computer doesn't have detailed specification standards, then it is an unqualified product. Why can't Stablity AI have its own standards? These standards could be the recognition of a model, an image, or open-source plugin creators. But you've given up on all these, you should have been the greatest standard setter.
MIdjourney creates its own standards to form a closed-source creation method. They focus on three areas: First, constantly enhancing the artistic annotation of the model; Second, simplifying user operation standards; Third, enriching community interaction standards. They then charge subscription fees for comprehensive services. After Stablity AI open-sourced the SD model, they let go of everything else. Nobody is willing to pay for an open-source action, let alone continuously donate out of goodwill for a company. The foundation of commerce is always fair trade.
Without self-imposed standards, without fair trade, in my opinion, these are the biggest problems Stablity AI currently faces. People don't use your tools that often, and many of the functions launched on your official website can be achieved by other software and web applications. You've deviated from your initial business intentions because you only wanted to make more money.
However, this era is one of decentralization, but resources depend on centralized win-win situations. The superficial decentralization is the rise of individuals as super individuals, while deep decentralization, like neural networks, allows others to continuously profit. This is the foundation for the survival of companies like Stablity AI. You might argue that setting standards, optimizing models, and building communities require significant costs. But in this world, there is another kind of co-creation, where you provide the dance floor for others to dance, the dancing people gain supporters, and the dance floor charges for food and drinks. Yes, it's like a bar, and it's precisely such bars that can make the charming people more attractive. This is the logic of decentralization, turning your stage into a dance floor for everyone.
The transition from a stage model to a dance floor model, the rigid watching of a solo dance and audience has now become a shared dance and cheers.
In my view, all other problems are chain reactions caused by the inability of the core driving force to solve fundamental problems. As long as the most invisible root problem can be solved on the whole business logic, all losses can be reversed. That is, reshape the standards, upgrade the transactions, help all original creators earn money, and let all ordinary people rush into the dance floor you've opened to dance, making friends with SD open-source model users worldwide, and trade between users and people worldwide.
Reshaping standards, upgrading transactions, helping all original creators earn money, this is the solution for the current world.
Chapter 3, Micro Operations in the World of Dust
Knowing the strategic level solution, the tactical level needs to be executable.
So we need to find breakthrough points from the details.
First, reshape the standards.
In the process of using the SD open source model over the past year, we have increasingly encountered dilemmas about the judgment standards of the model, measurement standards for using algorithms and parameters, and even deviations in learning directions. More and more people around me are asking one question: Do you have any good models to recommend?
Just such a question has troubled countless people for countless hours. All we can see is that sometimes a model will rise quickly due to the recommendation of a well-known blogger in the classification and public rating. Therefore, the basic standards for the application of the SD open source model are still vague. Therefore, how to set standards, who to set standards, where to post standards, and what benefits standards can bring are worth discussing. Looking at the practices of some other communities we are observing now, standards are divided according to different classifications, and users rate them autonomously, which obviously buries many high-quality models and is also full of a large amount of obscene content. Taking the Civitai platform as an example, as long as it is different from the ethos of this community, a platform mechanism with official characteristics can quickly gather popularity.
How to gather popularity
Establish three major communities: Professional Edition, Application Edition, and Generation Edition. Through various world-class activities, cooperate with well-known enterprises to form effective connections between enterprises, users, and ordinary people, improve the practicality of the entire community, increase the frequency of use by users, and the community benefits are diverse. We can even launch weekly and monthly activities such as high-quality model, high-quality prompt, and high-quality enterprise IP incubation to help some people save time and achieve traffic growth.
If you think you have already tried this method, then on this basis, AI micro-startups will be a very attractive crowdfunding method. Each plugin creator and model creator usually only get the "cup of coffee" provided by fans after they finish making it, which loses the anticipation and sense of gain. If it is pre-released for public crowdfunding through standard institutions and individuals, everyone will be more willing to reward when they look forward to the new version, and open source creators will be more motivated.
Each section can even be distributed to managers who are willing to participate in maintenance, and maintain a reasonable and harmonious environment through their influence. The past way of API and paid generation can't maintain the high-frequency word use operation of Stablity AI for a long time. In fact, a forum is not a problem for you. The first thing to solve is what kind of standards you create. In our current world, creators borrow from the SD open source model, Stablity AI borrows from well-known open source creators, takes turns as managers, provides exhibition space for each new IP, for each new creation, helps more people use the SD open source model more conveniently with a plan, in order to gain the most stable traffic. And in this world, traffic is the hardest thing to get, and it is the basis of everything.
Second, upgrade transactions
For the mode of tasks, there may be different transaction modes in different parts of the world. High-quality task release and task acceptance is a key business. Many platforms have also become gathering places for poor suppliers due to low-price competition. The transaction mode of Stablity AI should be upgraded, establish a credit system, provide an SD transaction coin, and even introduce a reasonable transaction currency to launch the requirements of the basic package, promote some high-level transactions in the open source ecosystem, and realize it on the Stablity AI platform. This may be an idea that has been eliminated many times in your meetings, and it is not entirely a creative idea, so we should re-examine what is the transaction in the field of AI open source.
The essence of the transaction is buying and selling. There is basically a lack of trust between the two. More and more businesses actually need AI products and AI services with commercial attributes. It doesn't matter who can provide products and services, what is important is the final work, which needs to become a content that must be displayed on the platform. If you choose not to display on the platform, you will lose some commission discounts. The more suppliers (individuals) with more transactions will get more commission discounts, and the content they display will also be subject to public supervision. Once cheating or public rating fails, more commissions will be deducted.
At this time, you may still feel a bit naive, because users are the most difficult to manage, especially users of open source communities. If Stablity AI participates in the community, it will need to invest more energy, but what if this task becomes a task jointly issued by Stablity AI and corporate IP? For example, a certain refrigerator company wants to create a set of new listings for refrigerators. The company entrusts Stablity AI with the brand promotion and task issuance through the annual framework agreement or single project cooperation. I think this will be a way to subvert the survival of the future advertising industry. It can enable Stablity AI to master the exploration of corporate AI applications that span the times. This model is built with many companies and will certainly be an effective attraction of the community platform.
Finally, take all open source creators to make money
Just like various Internet companies and hardware company developer conferences, Stablity AI also has the ability to lead developers, even model creators, and image creators to make money. I believe this is not a novel idea. But to be able to make money with everyone, there must be a premise - certification. Similar to Twitter's membership certification. However, our certification is certification approved by the official for the effect, and the judge must be Stablity AI.
At the same time, a reasonable list should be regularly released, including but not limited to
Model list
Developer list
Workflow list
Art effect list
Folk technical literature list
Industry classification list
And give guidance standards for various industries, clearly stating that different people are more competent in different fields and provide certain cooperation opportunities for them. Launch a trustworthy trading platform, such as the very well-known "Alibaba" in China. It makes the dishonest market more active, just taking some commissions from it. The entrusted "trust" is the key to their victory. I think this is very important for every open source creator in digital assets.
For specific implementation measures, I believe your company must have its own judgment. I just rely on my own understanding to express my hope that Stablity AI can go further, have a gratifying future, and create a brand new open source world with people, and become a very great company in people's hearts.