# Jan as a dreamer mission

This document is for how we tell Jan's new story and how can we add the story on the homepage - jan.ai.

Video version of the ideas: https://share.descript.com/view/DIKZ9elmhp8

First thing first

- I avoided using technical language as our audience is not people who directly know AI, master LLMs or want to learn; it is people who are looking for productivity and want to do it with AI.
- I aimed for a clear and sincere tone similar to that of Discourse.org.
- I focused on explaining precisely what they can gain from Jan and what they should expect from us, avoiding over-commercializing the product.
- I applied Obsidian.md's perspective on data ownership to Jan in terms of AI ownership, conveying how they should position us as the main message. To justify how such a tool could be free, I emphasized its open-source nature.

---

## Context

It's crucial we align on the context first, as without a shared understanding, the rest might not hold as much meaning.

- Jan (Now): Jan is a tool that runs Large Language Models (LLMs) locally on your device, enhancing your efficiency.
- Jan (Near Future): Jan will change how you use your computer by learning from your interaction, getting things done for you and making suggestions.
- Jan (Distant Future): Jan will connect humans and robots as an intelligence behind robots.

Steve Jobs and Steve Wozniak, in their youth, created a blue box that could hack phone systems for free calls. Jobs later reflected that without this experiment, the Macintosh would not have reached its pinnacle.


Similarly, the assistance LLMs provide is evident, **Jan made these technologies accessible to everyday users**.

---

When you want to do something on a computer, you see your mind as a part of that computer.

That means you speak the language of computers. We learn computers language to use the programs. The place we click, the moment they appear on the screen, etc.

**Computers language -> Coding -> Apps + Coding -> Human language**

- Yesterday, we did this with knowledge of coding. Today, we do it knowing the tools.
- Tomorrow, we will do it just by typing/speaking and even thinking. That is, we won't speak their language; they will speak ours. **Jan will help computers to talk with people’s language.**

Our vision is to build this future. People have always needed tools that help them and do things for them, but for the first time they will have a tool that is so powerful and personalized. And her(?) name is Jan.

So, we will have something that gets things done for users - which we'll call a … that learns from you and gets things done for you.

Then Jan will build the future with these features:

- Jan helps you solve everyday problems on your computer.
- Jan accomplishes tasks across applications.
- Jan makes suggestions before you ask for them.

### Example Use Cases

There will definitely be features like question answering, translation, summarizing, commenting, talking to files, but in addition to these:


We will have tens, hundreds of use cases. Using these features, sky is the limit, but we should have content that we need to guide people to.

> It's like having a personal assistant who knows exactly what you need, when you need it.

> Jan’s continuous improvement makes it not just an assistant, but a learning companion that grows with you.

## Then how to explain Jan's story?

Before moving on to the main message, I will explain the areas under the main message in the homepage.

**Get things done for you**
From scheduling meetings to organizing your digital workspace, Jan can accomplish your tasks across applications., executing tasks with precision.

**Learn and remember from your interactions**
With every interaction, Jan becomes more attuned to your preferences and working style. It learns and from the way you use it - also remembers!

**Full ownership - your own AI**
With Jan, you're not just using an AI - you're owning it. It runs on your computer, ensuring all interactions remain private, secure, and completely offline*.
*if you use local models. Jan also works with GPT-4.

**Works like a natural extension**
Jan fits into your natural workflows. Summon it with a hotkey, otherwise it stays in the background - it works like a natural extension of your workflow.

**Voice and type commands**
You can communicate with Jan in any language you prefer, either verbally (including voice) or in writing. It understands, responds quickly, and takes action.

---

Then we need to address any concerns people might have after we tell them what Jan can do for them. That's why I talk about privacy and open-source.

**Privacy-focus**

Jan stores data in your local filesystem. Your data never leaves your computer.

**Open-source**
Jan's code is openly available to everyone, you can edit and tinker.

**Unlimited Access**
You can use Jan with no restrictions on usage.

--

After explaining all the main features, let's move on to the main message. This main message was one of the most challenging topics for me.

Because it had to be a clear statement of both Jan's vision and his recent features. I reread Bill Gates' articles, I went through Sam Altman's interviews, I went all the way back to the machine quotes in Dune.

In fact, Jan's biggest features point to many things.

## Main message

In an interview, Sam Altman describes the future of ChatGPT, saying: "... my super assistant for all of my cognitive work." Sam Altman

> "... my **super assistant** for all of my **cognitive work.**"

Then I dug deeper into the concept of cognitive assistant, and I saw this definition in an article:

I came across many articles while researching the topic. I tried to understand the concept of cognitive assistant. In the simplest definition, they point to something like this:

https://www.sciencedirect.com/science/article/abs/pii/S1071581918300624#:~:text=According%20to%20a%20definition%20by,their%20ordinary%20power%20and%20reach.

This is exactly the kind of artificial intelligence that would complement our human beings. So it's not an enemy of man, but something that empowers him.

Here is Jan's main message and how to introduce Jan in one sentence:


Not quite good, right?

When you say cognitive assistant, people will understand that it is something cognitive. But it's still the biggest feature to pass the personal AI agent in the subtitle.

--

Afterwards, I started to think that 'cognitive AI' might not be easily understood. Not only is 'cognitive' not a word we encounter in everyday life, but it also isn't a 'warm' expression. It seemed too scientific.

* Cognitive assistant
* Personal assistant
* AI assistant
* Personal AI
* Personal agent
* Copilot

I wasn't very keen on the word 'assistant.' Because not only is everyone positioning their tool as an Assistant these days, but the assistant can also remain in minds as a role that does things you don't want to do, organizing rather than helping with complex issues. Jan is going to be much more than that.

So I started thinking about statements that answer yes to the following questions:

1. Is it understood to be a personalized copilot?
1. Is it short enough
1. Is it understood that it can learn from you?
1. If I say this phrase over coffee with my friend, will it be understood?
1. Does it convey doing tasks on your behalf?

I scored all the main messages in my mind according to these questions and tried to find out which one suits us.

I created [such a sheet](https://docs.google.com/spreadsheets/d/1gFv0UiHgA4lmCm_2WoNpcmZ3FMXgym_4Pnovtt6_yDU/edit#gid=0) and filled it out.


I filled out this table myself, explained the context, and also had ChatGPT fill it out, taking its suggestions. You can also go to this table and score from the tabs I created for you: https://docs.google.com/spreadsheets/d/1gFv0UiHgA4lmCm_2WoNpcmZ3FMXgym_4Pnovtt6_yDU/edit#gid=0

Looking at all the criteria, some main messages had already started to stand out.

In addition, I thought we could directly explain the subject in the subtitle.

After these scorings, a few messages began to stand out. Among them were these four:

* Your AI sidekick
* Your AI companion
* Your smart assistant
* Your AI assistant

However, due to being too generic, I eliminated AI assistant and smart assistant.

Only sidekick and companion remained. I saw that 'companion' was used a few times by Rabbit and since such an expression was on its homepage, I gave up on it and Your AI sidekick remained.

Even though this expression is understandable according to my criteria, I thought that a complementary message was needed, and for the sub-message, I added something that has been on my mind for a long time.

We can clarify this with your scores and the tests we will conduct among these messages.


or...


So, in this perspective, what our homepage looks like:


Excalidraw+ import: https://app.excalidraw.com/import?excalidraw=lzN_aTa1A34v,FneyfsivcleD_7PBisxstQ