# FY1 S2 Junwon Story

# Story

In its founding stage, a new company must define its first business. As the founder of this company, I have the great privilege and the grave duty to lead this process, to propose what we will do such that people can benefit from our existence, such that people will desire us to exist. We can only come into existence if people desire us to exist. 

This season, we tested two interesting ideas, gained important assets, and chose to focus on developing foundation models as the new interface for the internet. 

# Dreams

We dream to make a world where people can dream and make dreams a reality. 

We have an opportunity, for the first time in human history, to give people the power to directly produce the most economically important product of the time. All people can have the power, with Generative AI, to make programs and contents they wish existed for them and for other people. 

This can be the start of a true economic prosperity. And we can be the ones to start it. 

# Advances

## Ideas we tested

This season, we tested two product ideas: 

- **Life Coach** aimed to provide advice on habits, such as exercise plans, based on information that users record on a data notebook. We learned that people are not willing to overcome the friction of manually recording information.
- **Character Messenger** aimed to provide a way for people to chat with characters from history or fiction. We decided that we are not in a strong position to provide a superior product in this market.

## Assets we gained

We anticipated in advance that we may be pivoting away from the product ideas that we tested this season, and we took strategic steps to accumulate assets even as we pivoted. See FY1 S2 Product Plan for a detailed discussion of the assets we gained. In summary, we gained assets in three areas. 

**Design: find it or make it**

Before Generative AI, a user searching for a specific information on an online platform either succeeded or failed. We made a new general interaction flow, that we can also apply to our future product ideas, in which users can find the information if it already exists or make it using Generative AI if it does not yet exist, thereby reaching success every single time. 

**Technology: sims software stack**

We made a software stack that connects users on computers and mobile devices with probabilistically behaving AI sims on servers. 

**Economy: freedom and safety**

Using community policy and energy currency, we set up a way to guide members of our economy to enjoy freedom while respecting the safety of others. 

As we keep testing product ideas, we will keep accumulating assets like these. 

## Gratitudes

We thank you eternally for your trust and support. This season, we add special thanks to 

1. friends who answered our interviews, for your candid feedback,
2. members and moderators of online communities, for your candid feedback, 
3. developers of developer tools and AI models that we consume, for boosting our efficiency. 

# Plans

We plan to keep searching for business opportunities made possible by foundation models. Next target is Web Autopilot. 

We plan to differentiate our business in following ways. 

**Spend high on customer satisfaction**

Unlike existing internet businesses that pay for low unit cost of services using low unit revenue from advertisements business model, we will find ways to deliver high quality services to each customer. 

**Speed up customer success**

We will find ways to help customers reach success faster. 

**Find enduring advantages**

Since Generative AI can provide a magical experience to the first customer on a new online platform, contents inventory is no longer an enduring advantage. We will search for ways to deliver increasing benefits as we scale our operations. 

**Automate science**

Some may find success chasing, or even pioneering, consumer trends. That is not our way. We will instead rely on timeless wisdom and invest in creating automated ways to create and test ideas. We will invest sufficiently in tools for simulations and experiments. 

## Calls

1. Suppliers of developer tools, please keep enhancing software developing experience, 
2. Researchers, please keep expanding AI capability, 
3. Regulators and app stores, please review policy and make necessary changes to enable consumers to benefit from amazing products made possible by Generative AI. 

# Responses

**On our upstream dependency on foundation models**

As we predicted last season, OpenAI started losing monopoly as other developers started delivering comparable foundation models with great task completion skills, big context windows, and support for multiple data modalities. 

We expect we will continue to have reliable access to satisfactory foundation models. 

We have no plans to invest in training proprietary foundation models. 

**On our upstream dependency on server capacity**

We are witnessing a supply shortage of processors that specialize in accelerating tensor computation for AI applications. 

We expect this problem to intensify as businesses begin using AI simulations to discover opportunities and assess options. AI simulations will consume surplus computing power. 

In response, we plan to invest incoming capital into securing permanent server capacity in first-party data centers. 

# Character

**Desire difficulty**

For us, difficulty is safety, because we are safe when others have neither the ability nor the desire to replace us. When the work feels hard, we remember: the steeper is the mountain that we climb, the safer is the fortress that we build on top of it. 

**Set correct expectations**

Customers feel satisfied when our products meet their expectations. This means we can increase customer satisfaction in two ways: make better products and set better expectations. We remember to do the latter too. 

**Follow dream**

For us, market opportunity is merely a constraint. Within the set of options that market opportunity allows, we choose the ones that let us follow our roadmap. 

# Fortune

The search for a business opportunity is difficult, not just because it demands great skills, but also because it demands great character. We are sure to fail many times, and we are sure to feel pain each time. 

But let us remember that the rules of the game are in our favor: 

1. We just need to find one great answer. 
2. There are many great answers. 
3. We can try infinite times. 

So as long as we can endure pain and keep trying, while learning lessons each time we try, we are sure to win one day.