Sign In

Long-term competitiveness

Here are answers to frequently asked questions about long-term competitiveness.
Can a GPT Wrapper that uses the GPT API as is survive?
Wouldn't ChatGPT be ruined if it updates the same feature?
Can you call yourself an AI company without your own models?
So-called "AI Wrapper" services are popping up every day. Concerns are also being raised, such as whether AI is a bubble, that companies without their own models will be at risk with every ChatGPT update, and that the high cost of LLM models makes it difficult for them to survive independently.

So, are companies that own their own models or research technologies like VectorDB more secure? In fact, model developers have faced a crisis more quickly over the past few years. Specialized models in specific fields have become obsolete with the emergence of large-scale models, and fine-tuning and VectorDB technologies are losing ground with each major model update.

At the same time, all big tech companies with capital can build massive models. Companies like OpenAI, Anthropic, Meta, and Google are fiercely competing, rapidly driving down model costs. STT models became free overnight, and LLM models became more than tenfold cheaper in just one year.

Even companies like Amazon and Salesforce were often asked, "Aren't they just database wrappers?" in the early days of the web. There was also much speculation about what would happen if browser companies took over everything. But we now know this is false.
While the technology is still in its infancy and attention is focused on large-scale models, just as it was with web and app technologies, AI's biggest opportunities will arise in the "app market." This trend has already begun. Cursor AI became the fastest company in history to reach $100 million in ARR. Many Silicon Valley gurus, including Y Combinator, share this perspective.

Ultimately, what matters is who best solves the customer's problem. AI application app companies like ours may fail, but it's not because they failed to satisfy customers, but because they didn't develop models or because of high costs.
Is there a moat?
What if other startups or large corporations copy it?
Among the stories about 'moats', I most sympathize with Square co-founder Jim McKelvey's book 'Uncopyable'.
A moat isn't a pre-built "defense wall," but rather a natural byproduct of constantly solving problems others deem insurmountable. While outsiders attribute Square's ability to fend off Amazon to its various technologies, these technologies were actually merely byproducts of the problem-solving process.
Ultimately, the only way to build and maintain a moat is to solve problems relentlessly and quickly.
Think of products like Notion, Canva, and Linear.
Some argue that Notion's accumulated user data is a moat, but the same holds true for Evernote. Evernote has shrunk, and Notion has triumphed.
What ultimately determined their success was how quickly and persistently they solved users' problems. Even now, despite their size, they continue to improve their products at an astonishingly rapid pace.
Furthermore, startups have a significant advantage over large corporations in newly emerging market opportunities. Walmart has been outpaced by Amazon, and telecommunications text messaging has been outpaced by KakaoTalk. This is the result of differences in decision-making and execution speed.
The LilysAI team is also faster than anyone else at identifying and solving new problems in the field of "data understanding." Although still in its early stages, we already have a user base more than ten times that of products developed by major companies. While many products try to copy ours, we will continue to widen our lead by offering novel solutions to the "data understanding" problem.
Is Google's NotebookLM a competitor?
How can I win?
It is true that NotebookLM is currently addressing the most similar issues to ours in the market.
So we're always carefully monitoring user feedback.
However, the market is just beginning to open, so it feels like tadpoles are cautiously swimming in a vast ocean, trying to avoid collisions. The field of data comprehension is equally vast.
NotebookLM focuses on "reducing the difficulty of materials," and as a result, its podcasting feature has received rave reviews. However, our product's users are those who, for professional or situational reasons, must handle vast amounts of material. For them, not only "reducing the difficulty" but also "increasing the speed of comprehension" is crucial. NotebookLM falls short in this area.
I believe numerous products will emerge in this space in the future, each with its own potential to succeed. Therefore, we are focusing on solving the problems we seek to solve better than anyone else.
So why can LilysAI survive?
What are your competitive advantages?
1.
We are a team with a crazy obsession with our users.
As AI dramatically accelerates engineering speed, the bottleneck in product development is shifting from coding to listening to and incorporating user feedback.
Therefore, for a team to truly explode in competitiveness, all members must move beyond simply understanding users to instinctively experiencing the user experience. This is precisely why, among AI applications, coding-related product groups are evolving the fastest. Products like Cursor AI, where developers are also users, can significantly reduce the time spent on user research and QA.
The fastest way to get information about a user is to:
1. Intuition through dogfooding
2. Team member feedback
3. Feedback from strangers
4. Sending the prototype to actual users
5. A/B testing
Our team has established a culture of "dog food" as a daily routine, and we rapidly refine our products through rapid internal proof-of-concept (POC) and an active team feedback system, repeatedly experimenting. This is what creates a long-term competitive advantage.
And this philosophy isn't just our argument. Interviews with AI guru Andrew Ng and Cursor Team developers repeatedly confirmed the same perspective. This serves as a powerful validation that our approach is the right one, and it gives us confidence that we must continue on this path unwaveringly.
2.
We are a team with leading perspectives and technologies in the field of AI application apps.
We're a team with experience developing AI applications, confident in the potential of AI technology even before the ChatGPT craze.
So, from the beginning of our business, we have established clear perspectives and made bold choices on various issues that have been controversial.
For example, despite the perception that model costs were too high and business was difficult, we anticipated rapid cost reductions and made early decisions. We also recognized the importance of resolving trust issues early on and incorporated them into our products.
In the AI field, new keywords like "agent" and "RAG" are trending every day. However, we don't simply follow trends; we deeply understand what truly matters technically and reflect this in our products.
While our team doesn't develop models directly, we pride ourselves on being unmatched in technical understanding and execution capabilities related to AI applications.
Made with Slashpage