Help Me Build an AI That Learns From Reality

My name is Ibrahiem Hamed. I am an independent AI researcher and currently pursuing a master’s degree in Artificial Intelligence at The University of Texas at Austin.

For more than a year, I have been building something that started as an experiment and has become the most ambitious project I have ever worked on.

It is called Alpha Yazan

labs.yazanalpha.com

Alpha Yazan already exists. It is a working AI research system that observes real-world data, makes decisions before outcomes are known, records what it believed, and later compares those decisions with what actually happened.

But the mission is much bigger than making predictions.

Why I am building Alpha Yazan

We are entering a future where AI systems will increasingly make decisions that affect our lives.

They will help people make decisions in medicine, transportation, cybersecurity, finance, engineering, science, and many other fields.

But there is a question I believe we have not answered well enough:

How do we know when an AI system actually deserves our trust?

An AI can sound intelligent.

It can produce an impressive answer.

It can perform well in a demonstration.

It can even be correct many times.

But being impressive is not the same as being reliable.

I want to explore a different approach.

An AI should not earn our trust because it sounds confident. It should earn trust through evidence.

That idea is at the center of Alpha Yazan AI.

An AI that must face its own mistakes

When Alpha Yazan makes a decision, that decision is recorded before the future is known.

Later, reality provides the answer.

The system cannot go back and change what it originally believed.

A correct prediction is recorded.

A wrong prediction is recorded too.

A decision that could not realistically be executed is recorded.

A failure is not erased because it makes the system look bad.

Failure becomes part of the evidence.

That is important to me because I do not want to build an AI that simply appears successful.

I want to investigate whether we can build an AI that can eventually understand the difference between:

what it believed,

what it could actually do,

what really happened,

and what it should learn from the difference.

Why use financial markets?

Right now, financial markets are Alpha Yazan’s first real-world laboratory.

Not because the ultimate mission is simply to build another trading bot.

Markets give this research something extremely valuable: the future arrives continuously, and the AI cannot control the answer.

The system makes a decision.

Then the world moves forward.

Prices change. Liquidity changes. Information arrives. Opportunities disappear. Some predictions are right. Others are wrong.

That makes the market a powerful environment for studying an AI system over a long period of time.

The bigger question is not simply:

“Can AI predict a stock?”

The question is:

“Can an autonomous AI system learn from reality, understand its own limitations, and build a trustworthy record of its decisions over time?”

This is already a real system

I am not asking for help to build an idea that exists only on paper.

Alpha Yazan already has working research infrastructure.

It has processed real market data.

It has made and recorded prospective decisions.

It has gone through historical testing, forward testing, paper-trading experiments, engineering failures, corrections, and redesigns.

Some ideas failed.

Some results have been encouraging.

Many questions remain unanswered.

And that is exactly why I believe the experiment should continue.

I am not claiming that Alpha Yazan has proven a profitable trading strategy.

I am not asking people to fund stock-market speculation.

Donations will not be used as trading capital.

The purpose of this fundraiser is to support the research itself.

What your donation will support

Keeping a real AI research system operating over months and years requires resources.

Your contribution will help support:

real-time and historical research data

cloud computing

secure evidence storage and backups

AI and machine-learning infrastructure

continued prospective experiments

reproducibility and statistical analysis

independent technical or research review

public documentation of results and failures

the infrastructure needed to keep Alpha Yazan running long enough to build meaningful evidence

Until now, I have carried much of this work with my own time and personal resources.

But the longer this experiment runs, the more valuable the evidence becomes.

One week of data can tell us very little.

Months can tell us more.

Years could allow us to study something much more interesting:

How does one persistent AI system change when it repeatedly makes predictions, experiences failure, records what happened, and is forced to confront reality?

What I hope this can become

I do not know what Alpha Yazan will ultimately prove.

And I think that uncertainty is part of what makes this research worth doing.

Maybe some ideas will work.

Maybe many will fail.

Maybe the most valuable result will not be better predictions at all.

It may be discovering better ways to determine when an autonomous AI system should—and should not—be trusted.

I don't want to build an AI that never makes mistakes.

I want to investigate whether we can build an AI that knows when it was wrong, remembers why, and becomes better because of it.

If autonomous AI is going to become part of our future, I believe that ability may matter much more than appearing intelligent in a short demonstration.

Help me keep the experiment alive

I am one person building this system, and I have taken it much further than I originally imagined I could.

Now I need help giving Alpha Yazan something I cannot create alone:

time.

More time observing reality.

More time collecting evidence.

More time learning from mistakes.

More computing power and data to keep the experiment running.

If this idea interests you, I would be grateful for your support.

You do not have to understand financial markets or machine learning to be part of it.

You are supporting a larger question:

What can an AI become when it is forced to learn from the real world—and when evidence, rather than the AI itself, decides whether it is improving?

Every donation helps extend the experiment.

And if you cannot donate, sharing this page with someone interested in AI, science, technology, or independent research can help enormously.

Thank you for giving me the opportunity to keep building.

Ibrahiem Hamed

Independent AI Researcher

Creator of Alpha Yazan AI

The world teaches.

The AI learns.

The evidence decides.