GriGsi

where ideas take shape

The future is not scary,
it is simply unknown. 😉

I’m a 50-year-old ordinary television professional who keeps hearing about how AI could do terrible things, take away jobs, and who knows what else.

I think this is more likely something unfamiliar — something that simply needs to be understood!

And because I’m curious to find out what AI can actually do, what I can do by working with it, how far we can go together, and how far AI itself can go, I decided to run an experiment.

Here are the results so far!

See what happened → See the other tries →

Human + AI

Collaboration.

HUMAN

  • Experience
  • Ideas
  • Problems
  • Questions
  • Judgment
  • Curiosity

AI

  • Research
  • Coding
  • Prototyping
  • Testing
  • Iteration
  • Ideas

Like anyone else, I have problems.
They’re for solving — not for putting off!

So I take one problem, I think of a solution, and I ask AI whether it can make it real.

It gives me suggestions.
Some I take, some I don’t!
Others we talk through until it’s clear what I want and what it can do. And that’s how we get to some kind of result.

Sometimes it comes easily.
Sometimes it doesn’t.
Sometimes it doesn’t work at all.

But that’s the experiment!

And the results of our collaboration are the tools we made.

Here’s roughly what the process of our collaboration looks like.

Question Experiment Failure Revision Prototype Working tool

Results

I started with problems I actually had!

I wanted to cut the risk of mistakes down to a minimum, and bring it all together into one shared danger alert.

Extensions

REAL PROBLEM When I am tired or distracted, I can send the wrong email, copy the wrong phone number or open a dangerous link.

HUMAN + AI I wanted something to check for me when I am tired, distracted or simply not paying enough attention.

RESULT Extensions checks whether a phone, link or email has been used before, whether others have used it, and whether it is known to be safe or marked as dangerous.

Try it →

Then I wanted to bring task reminders and communication into one place.

SENDi!

Send once. Continue if needed.

REAL PROBLEM Sometimes I just want to send something to someone.
I don’t want that to automatically create another permanent contact.

And sometimes I need to remember something, or someone needs to do something.

HUMAN + AI So I put communication and tasks in the same place.

RESULT SENDi! lets you send a photo, file, link, text or voice message once. If you need to continue, you can.

And “!” turns it into a reminder or a task — for me or for everyone.

Try it →

After much hesitation regarding whether something is true — and to what extent — my instinct led me to the idea of creating a tool that determines the level of trust in a given piece of information or a person.

Trust Check

REAL PROBLEM Anything can be faked now — a message, a link, a website, even a video — and I couldn't always tell what to actually trust.

HUMAN + AI I wanted to paste whatever I was unsure about and get a straight answer, not a technical score I'd have to decode myself.

RESULT Trust Check — one field, one plain-language verdict, with the reasons and what was checked shown right next to it.

Try it →

I wanted to learn the real strength of money, and how it changes over time — for me and for the rest of the world: world → country → region → profession → income → expenses → assets → personal resilience, and back again.

How strong is your money?

REAL PROBLEM Prices kept changing, my income stayed roughly the same, and I couldn't tell anymore whether I was actually falling behind — or just imagining it.

HUMAN + AI I wanted to understand the real cost of money — not a headline number, but what it actually means for me.

RESULT Money and finances — I compare moments, see where the money actually goes, and open Trust for the sources behind it, including official income context.

Try it →

Then came the question: could I build my own tools for my profession, so they'd work easier and faster?

Production

REAL PROBLEM I needed to do something professional — check facts live, get the right picture on cue, keep picture and sound together, turn long recordings into searchable text, run a show without a studio, and let people actually watch it happen — but I had none of the usual equipment or crew.

HUMAN + AI I started with Air Check, then kept adding one piece at a time as I ran into the next real need.

RESULT Six pieces, each solving one real need along the way — from the first fact-check to a live show people can actually watch.

Air Check → Illustrate → AVP → Transcribe → Show Maker → Fresh and Now →

I found myself not knowing how, to whom, or in what format to report something wrong — so I decided: help my future self, and maybe make it easier for someone else in the same spot.

WhoTo

I saw something wrong — who do I tell?

REAL PROBLEM I see something wrong — water theft, a diverted river, illegal construction — and I don't know which authority to report it to, or how to word it so it isn't sent back.

HUMAN + AI I wanted to describe it in my own words or by voice, and have the system shape it into a proper complaint, say whether it's sound, and name the competent authority.

RESULT WhoTo records it (video/voice + GPS + time, with a checksum), drafts the text and the assessment, routes it to the authority — and then you track where it got.

Try it →

One problem leads to a solution, or to the next problem or question, and the first answer or result is rarely the last one.

The important thing is not to give up — and to keep looking!

Programming is not my profession.

But I chose not to be afraid. Instead I started looking into whether AI really is dangerous and whether one day it might replace me. For now, the answer this experiment is leading me to is: hardly… At least for now…

I finance the experiment myself and develop it in my spare time, because the world is changing and I don’t want to fall behind. I want to keep up with the times.

Evidence

These tools are not the experiment.

They are the evidence of it.

These above come naturally from my profession. But I don’t limit myself to that!

There are other horizons that interest me.

Here they are.

Also from this experiment

Check

A place to look at a text — what seems true, what doesn’t, what needs another look.

Open /check/ →

Всичко от GriGsi

The experiment is not finished.

I don’t know what comes next.

Maybe another problem. Maybe another idea.

Maybe something completely different!

If you want to share something with me, or help me out or sponsor, write here. :)

Maybe the next experiment starts with your problem.