AI Rebuilt the Software: What Now for Commercial Programs?

At first I assumed it was a program made just to attract attention and views on YouTube. But it became such a hot topic that I downloaded it and tried it.

It is called SolidCraft.

I was genuinely startled. Really, this level of quality?

What surprised me even more was that it is a one-person project. I heard one person developed it alone. People say updates are slow, but for a solo project that is unavoidable.

Whether you call it development, or recreating existing software with AI, I am not sure. What matters is that the result exists at all, and that it was made purely by using AI.

Honestly, I still find it hard to believe.

I have never used SolidWorks myself. But I had always wanted to study this kind of program, so I looked into various ways to do it.

In the past, you could find a pirated copy if you looked around a little. These days it is not that simple because of server-based authentication and license management. So I assumed this would be out of reach.

Looking for other options, I found that a student version is free for about a year. But I am already in my mid-fifties. Ha. That would not be easy either.

In the end I settled for watching tutorials on YouTube. Oddly enough, I like this kind of professional program. Learning complex features one by one and producing the result I want myself is fun.

Then I forgot about it for a while.

While building the AI hub website, I was looking up various materials and came across this program again on YouTube.

At first, honestly, I thought this.

How good could something like that possibly be?

I downloaded it after reading the description, but I did not expect much. Still, out of curiosity I ran it once. And I am still using it.

Once I actually used it, my thinking changed completely.

Really, this level?

This was not something to dismiss as a few similar features implemented roughly. I have not used SolidWorks myself, so I cannot compare its performance precisely with the original. But the level of completeness I felt while running and using it myself was truly remarkable.

And I thought again.

One person made this?

To be precise, what is even more astonishing is that AI contributed to understanding the structure and recreating the existing software. It seemed to show a possibility on a completely different level from one person developing in the traditional way.

Of course, the fact that AI made it does not mean everything is perfect. There will be problems to solve, and features and stability can be improved. If one person develops and maintains it, update speed will have limits.

But even the current result was enough to shock me.

If it is already this good, how will the software market change in one or two years?

I do not think this is just a story about one free program appearing.

What happens to the existing software market?

The strength of commercial software is not simply that it has many features. The user experience built over a long time, familiar commands, workflows, and professional features all become barriers.

Professional programs like this are especially hard to switch away from once you are used to them.

Even if a better program appears, it is hard to move if you have to abandon the interface and commands you know and learn everything from scratch. People do not change a program they have used for years just because it is slightly better.

But what if a new program offers an interface and workflow similar to the existing one?

Users can reuse much of what they already know. The burden of learning a new program decreases. Similar screens and commands do not mean every feature and compatibility is identical. Still, lowering the entry barrier alone is meaningful.

I think this part matters considerably.

One of the greatest assets of commercial software is that users are already accustomed to it. If AI can analyze the structure and usage of existing software and produce a similar program, the barrier that was hard to cross may become lower.

So what should existing software companies compete on?

Having many features alone may no longer be enough. Stability, precision, compatibility, technical support, ecosystem, and continuous updates will become even more important.

I do not think existing companies will collapse right away. But the development costs and market entry barriers that were taken for granted can clearly change.

It is even more surprising because it is open source

What I pay particular attention to in this program is that it is open source.

For people who use it at work, this advantage can feel even bigger.

When you use commercial software, inconveniences inevitably appear. The feature you want may be missing, the workflow may not fit you, or the program may not support enough to reduce repetitive work.

With commercial software, fixing such inconveniences may mean waiting for the vendor's update or paying extra.

With open source, the story is different.

You can modify the program yourself and add the features you need. You can change the interface to fit your workflow, remove unnecessary features, and automate repetitive work.

Literally, you can build a program for you, by you, of you.

Of course, being open source does not mean you can modify everything freely. You have to consider license terms, code structure, and technical limits. Still, being able to access the source code and improve it yourself is a huge difference.

For professional engineers in particular, that value grows even more.

They know exactly what is inconvenient in their work. If they also have the ability to fix it, the productivity gain can be substantial.

What if an AI agent is attached on top?

This is where I think the real core begins.

This program can be connected to an AI agent, and you can reportedly also direct work with text.

I think this is a stroke of genius.

Professional programs are hard for beginners. There are many menus, complex features, and professional terminology to learn. Simply running the program does not produce the result you want.

But if you can direct work with text, the entry barrier can be lowered. You can automate repetitive tasks, connect multi-step work, and reduce processes that previously required manual operation.

And what about engineers who are already skilled with the program?

For them, an AI agent will not be a mere convenience. It can be like installing a turbo booster.

Experts already know the program's structure, commands, and design principles. They can judge what to build, which conditions must be met, and whether the result is correct.

The same applies when directing work to AI. You must know exactly what you want to get a good result.

I have written about this before, but you can design only as much as you know.

Beginners may not know the professional terms, the order of work, or how to verify the result when trying to direct AI. Explaining to AI may even take longer than doing the work directly. Ha.

Skilled engineers are different. They can describe the desired result concretely, recognize a wrong result, and point out precisely what must be fixed.

Add the advantages of open source, and the possibilities grow further.

You are not limited to using the existing program as it is. You can modify it to fit your workflow, automate repetitive work with an AI agent, and add the features you need yourself.

This is not just about using a program for free.

It is about professionals building their own working environment and raising productivity beyond what existing software offers.

What will it look like in one or two years?

Honestly, this experience made me rethink the pace of AI progress.

In the past, creating a piece of software required considerable time, people, and expertise. Even now, developing and maintaining proper software takes a lot of effort.

But if AI can support the process of analyzing an existing program, understanding its structure, writing code, and implementing features, the story changes.

One person can attempt a project on a scale that was hard to handle in the past. If it is released as open source, others can join and improve it.

Of course, the fact that AI contributed to development does not guarantee the quality of the result. Making software usable in practice and keeping it stable over the long term is another matter.

Still, this experience showed a clear possibility.

Perhaps the era when only large companies can make high-quality software is ending. Small developers and individuals can also use AI to attempt professional programs and offer new alternatives to the existing market.

Existing software companies should not take this change lightly either.

What users want is not necessarily a product from a famous company or an expensive program. It is a program that does the job they want properly.

And what if that program is free, open source, modifiable to fit their workflow, and able to use an AI agent?

I think that possibility is quite large.

Closing

At first it was a program I downloaded without much expectation.

But using it directly was far more impressive than I expected, and above all, the fact that one person used AI to produce this level of result keeps staying in my mind.

Using this program did not just make me think I had found a good free program.

It made me think about how much the way software is made and used could change.

For me, being able to use it for free is good. Ha.

But what are the company that made this and the companies developing existing commercial software thinking?

From a user's perspective it is a welcome change, but from the perspective of the software industry as a whole, it may be a considerable challenge.

And if an AI agent is combined here, and an environment where users can modify and improve the program themselves is in place, the impact will grow even larger.

What kinds of programs will appear one or two years from now?

It is astonishing already, and the future makes me even more curious.

The shock I felt this time is not only about the performance of one program. It is that I directly confirmed the possibility that individuals can use AI to create software at a level that was hard to imagine before.

I want to keep watching how the software market changes.