Blog

"Great companies pursue and, not or." -Jim Collins, ‘Built to Last: Successful Habits of Visionary Companies’ "You can get closer to the truth through thesis and antithesis" -Hegel The co-founders of LilysAI spent over two years talking about "what kind of company do we want to build?" before starting the business. On this page, we share our philosophy of pursuing and, not or. 📦 Product: A company that demonstrates entrepreneurial creativity and originality The reason we started our business was not because we wanted to make a lot of money or build a big company. We started our business with the dream of delivering great value to people through original creations . Since we were young, we were people who wanted to create something unique. CEO John dreamed of becoming a cartoonist and guitarist when he was young, and co-founder Morrie dreamed of becoming a broadcast producer. We, who have come this far chasing the dream of whether there are any creative works that can have a greater impact, are people who feel the greatest joy in creating overwhelmingly cool products that once used, cannot be returned to . We believe that the best products that change lives are born from a subjective perspective . We do not think that the United States is the best and that Silicon Valley is the center of the world. Of course, it is great to benchmark overseas models and grow significantly, but we believe that world-class software with originality can be born in Korea as well . Bong Joon-ho, BTS, and Lilys In fact, we are constantly refining our own framework to increase the chances of people loving new features when we actually release them . We also get input from users every day, watch competitors and market trends, and study new technologies. We need to keep our feet firmly on the ground so that we don't have wild ideas . 3 criteria to check when imagining new features Is it something that corresponds to a fundamental human desire that has not changed for a long time, and is it something that occupies a large area of that desire? Is this something people are actually doing today, and increasingly so? Are there any strong incumbent players whose customers overlap with ours, or can we do better? (Is it technically feasible at this point in time, or is there no product in the market that meets the need?)

It is the era of AI. So-called 'AI Wrapper' services are pouring out every day. At the same time, there are concerns that AI is a bubble, that companies without their own models will be at risk whenever ChatGPT is updated, and that the cost of the LLM model is too expensive to survive on its own. So, are companies that have their own models or research technologies like VectorDB safer? In fact, model developers have been in crisis more quickly in the past few years. Specialized models in certain fields have become useless due to the emergence of large models, and fine-tuning and VectorDB technologies are losing ground with each update of large models. At the same time, all big tech companies with capital can build huge models. OpenAI, Anthropic, Meta, Google, etc. are competing fiercely, and the cost of models is rapidly decreasing. STT models became free overnight, and LLM models became more than 10 times cheaper in a year. Companies like Amazon and Salesforce were often asked in the early days of the web, “Aren’t they just DB wrappers?” There was also a lot of talk about what would happen if browser companies did everything. Products like Notion and Canva didn't grow with any special technology. In the end, what matters is “who solves the customer’s problem best?” AI application app companies like ours may fail, but it will be because they failed to satisfy the customer, not because of their own technology or high costs. We believe that only focusing on solving customer problems and providing innovative product experiences will create a moat. Reference: A business style that focuses on solving customer problems.

There are many predictions that developers will disappear in the AI era. I'd like to make a cautious argument that there will be more developers. In the past few decades, the productivity of developers has increased exponentially. One person can now do the work that ten developers did ten years ago. With the power of open source and the cloud, it has become possible to provide large-scale services with less than five developers. But why is the demand for developers increasing rather than decreasing? The reason is that software is rapidly absorbing other fields. The software market itself is growing faster than the productivity of developers. Now, media, commerce, and finance are all dominated by software companies. What about the AI era? I think this trend will grow even bigger than you can imagine. Psychological counseling, childcare, law, medicine, all areas that were thought to be exclusively human will be absorbed by software. This pace is so fast and so intense that I think it may actually increase the demand for developers. However, development skills may be considered as something like the ability to handle Excel, rather than a specialized skill as it is now. With the power of AI, it may become an area that anyone can learn without much effort or training. So the most important thing is the ability to provide what people want. For that, development skills will become more instrumental. At times like this, I think you should work for a company that can develop application skills and application abilities rather than the peripheral skills of a developer. Lilys AI expects developers to plan, interview, report statistics, and produce results. I think it is the best company to develop the ability to satisfy people by using development as a tool. If you want to develop the ability most suitable for the coming AI era, please check the link below. We are currently hiring!

I would like to share one by one what I learned and felt while creating three AI products: Vrew, Rutton, and LilysAI :) With the incredible advancements in artificial intelligence, there is now an opportunity to ‘automate’ many human intellectual tasks. But ironically, when our team discusses the UX of a product, we often talk about not falling into the “automatic trap.” When you ask AI to do something, it's rare for it to get a perfect score. Most of the time, it ends up getting around a 70. You might think that it would be meaningful if you could just create a draft with 70 points, since it would save a lot of time. However, as you follow the user's workflow, you will often find yourself having to destroy the 70-point result created by AI and create 100 points from scratch to achieve your desired goal. For products like this, users initially react by saying, "Wow, this is amazing," but in reality, they don't reach a usable level, so retention isn't very good. We call this phenomenon the 'automatic trap'. So how should I fill in the remaining 30 points? A lot of things could be solved with the engineering and UX layers. Let me give you an example using our products. LilysAI is a video summary service. There was an issue where the summary was sometimes inaccurate or overly abbreviated, making the reliability questionable. When I used the existing video summary service, I felt inconvenienced because there was no action that the user could take when the reliability was questionable, so I ended up having to watch the video again.