• A note before this proposal. I have a learning disability that makes English word recall unusually difficult, although it does not stop me from understanding or reasoning about the ideas. I write in Traditional Chinese first, then use LLMs to discuss translation and English expression. So an AI detector may strongly flag the English text, but the thoughts, experiences, order and decisions are mine.

• So if an AI detector marks this text as 100% AI written, I understand why, but the label is also a little strange in my case. The English has a lot of help from LLMs. The thoughts, experiences, order, examples and decisions are mine.

Project summary

This project comes from trying to put two things together. My exploration with AI, and an attempt to improve the situation I am actually living in.

I have been a science fiction fan since I was a child. It is difficult for me to describe how exciting it feels to be alive at this time. Since I started using LLMs, I have basically been unable to stop.

Part of this also comes from some unusual experiences in my own life. From the time commercial models started becoming widely available until now, when something new seems to appear almost every day, I have been following new information every day and trying all kinds of things I imagine. I keep thinking of projects I want to make, then another one, then another one. Sometimes I do this until I forget eating and sleeping and make my actual life into a complete mess.

Even so, I still cannot stop.

And about LLM, I have kept thinking multimodal AI training. What I imagine is more excessive than combining different kinds of media.

There can be the view from the software.
The view from the hardware.
Even the view from the floor.
There can also be the view of another view.

I want every kind of view to be allowed to become part of the record. I also want the record to have room for error, deviation and things that are simply unknown. I do not think everything has to be cleaned into one correct answer before it becomes useful information.

The imagination is beautiful.
Reality is crueler.

Then eventually I thought of this project.

Maybe the thinking I cannot stop and the real problems I should already be facing do not have to keep fighting each other.

Those problems are already full of possible data.

The real mess. Cleaning. Searching for things. Moving things. Animal care. Drawing. Woodworking. Ceramics. Repair. Making things with my hands. All kinds of different tasks that already require my body to do something in the physical world.

None of them were invented for research. They are things I actually need to do.

While doing them, I can also consciously think about how the event should become data. Where did it start. Where did it end. Is this one action or several. What changed. What information should belong to this event. What sensors would actually be useful. What hardware would be needed. Which part is an observation. Which part is my interpretation. Which part is uncertain or still unknown.

Maybe this could eventually become some kind of common structure where very different people, doing very different things in very different situations, can still produce useful multimodal data.

Then people would not necessarily need to be given endless fixed artificial tasks just because researchers need data. Some of the data could come from things people already need or want to do in their own lives.

If that ever works across many people, it might also produce something much more varied than repeatedly recording the same designed action. Maybe that kind of data could eventually help models learn to deal with more kinds of situations, including states and actions that were not cleanly designed in advance.

Of course, that is still the beautiful version inside my head.

I know the full idea is far beyond what I can prove now. So I only want to cut out the first small piece.

For this first experiment, I will start with myself, my real environment, and things I already need to do. I want to see whether I can make the smallest usable data structure, record some real events with it, and put the result in public where other people can actually look at it.

If that small piece works, then perhaps the next piece becomes worth trying.

And if nothing else, I hope doing it will also help push my own life forward.

What are this project’s goals? How will you achieve them?

Here I do not want to chase the far future of the idea yet.

I want to build the first possibility test. I already have some hardware. I will add a few more sensors and use what I know from multimedia and interaction design to make a first recording setup that does not get in the way of the actual work.

I want several cameras looking from different angles, including my own first person view, while I do the real cleaning and organizing work in my room.

At the same time I will record what equipment I use, what I change, what it costs, what goes in, what comes out, and what information turns out to be useful or useless, also want to record the failures.

And build a public website and research log while doing this. I will publish the setup, the changes, selected data, what I learned, what I think should be tested next, and things another person trying a similar experiment may want to know first.

For this first phase, my standard for success is intentionally very small.

I only need to make this start and actually happen.

The equipment can be crude. The first data may turn out to have almost no value. The first structure may be wrong.

What I want to show is that there is something real here that can be tested again, improved, or given a larger bet.

How will this funding be used?
The funding would mainly go to a few things.

The first and most important is supporting my basic living costs while I spend real time doing this work.

The second is hardware. I already have some equipment, but I will probably need more cameras, sensors, storage, mounting equipment and other small things once I start finding out what is actually useful.

There may also be software, cloud, hosting, API or compute costs.

Because this first phase is mainly about making the experiment actually start and finding out what the real requirements are, I do not think I can honestly give a very detailed budget before doing it.

At the $5,000 minimum, my plan is to work mostly with the lower end hardware and computers I already have, add only what is necessary, and make the first version happen.

For me, $10,000 means having more room to replace or add hardware that turns out to be useful for the experiment, and possibly equipment suitable for later small scale processing, testing or training experiments.

Who is on your team? What’s your track record on similar projects?
At the moment, the team is just me. If later stages turn out to be worth continuing, I also hope people who are interested may want to join.

I am not from computer science, and I am not someone who has spent years in Silicon Valley accumulating a lot of practical experience. My university major was digital media design.

Along the way, because of some unusual personal problems, which I talk about more on my profile if you are interested, I also do not really have a conventional track record I can point to.

The closest thing I have had to partners are the different models we have now.

In these short two years, with their help, I have tried countless things that before would have stayed only as thoughts. Many of them were things I had already been standing in front of for a long time. I could see the door. I could roughly understand what was on the other side. But because of different barriers in ability, language, memory, execution, or technical access, I could not get myself across it.

I still lack a solid foundation in many technical areas. But through trying, failing, starting over, breaking things, asking again, and trying another way, I have slowly accumulated a strange mix of experience and ideas from many different directions.

The links below are not really a portfolio. They are more like traces of my recent attempts to challenge some of my own problems and make changes in real life.

https://youtube.com/@11-y0w0y?si=6Zx2fK7OvOFE4ujp

https://bsky.app/profile/q0t0p-11-y0w0y.bsky.social

https://www.tiktok.com/@11.y0w0y

https://x.com/1_y0w0y_1?s=11

If there is one thing I can honestly count as my track record, maybe it is this.

Even when I am stopped outside a door, or stuck in front of something I do not know how to get past, I will usually keep trying to find another way, then another way, and keep trying to make it a little better.

What are the most likely causes and outcomes if this project fails?
The most likely reason is probably not that I fail to raise the money.

Whether this fundraiser succeeds or not, this path is already something I thought of today that may help push me to face the problems in front of me.

The more likely reason for failure is fear and anxiety, I think. Real problems keep appearing one after another. Then, together with the problems themselves, all the imagined consequences also start appearing in my head. It becomes very easy for me to stop moving.

Then I start thinking again.

Maybe there is another possibility. Maybe there is another way to change everything. Maybe if I think a little further I can find some different route.

And then the action stops again.

So the most likely failure is that I once again use imagination to escape from something I should actually be doing, and fall back into the same negative cycle I have repeated before.

The difference this time is that I can see this pattern much more clearly.

I have already created this project as one possible way to turn the thing I cannot stop doing into something that forces me to deal with reality.

Even without funding, I think I will still try to build it little by little.

Maybe eventually this stops being mainly about wanting to participate in this enormous AI moment happening around me.

Maybe it becomes a way to use that excitement to make myself deal with the things directly in front of me.

How much money have you raised in the last 12 months, and from where?

For projects like this, none.
The previous thing I actively tried as another way to break my own situation out of its cycle was this fundraiser for the animals and household pressure.

https://whydonate.com/fundraising/23-animals-breathing-room

It has received one $23 donation so far.
That donation was from me, mostly as a slightly self-mocking test to make sure the donation system actually worked.

So in terms of outside funding, I have raised $0.