[Snickerdoodle]

I used AI recently to apply for scholarships for school in mass.

Stuff like that or resume building and job applications are always cool resources to use too.

[dreadful.diary]

Really interesting projects you're working on!

I’ve used AI to build several substantial software products, including taking software into production and reaching a substantial ARR before exiting.

Across these projects, I’ve worked on the interfaces people interact with, the systems running behind them, databases, integrations, permissions, automated agentic workflows, testing, deployment, and ongoing maintenance.

AI has been reliable part of some of the process.

I’ve also used AI to investigate my own software. To trace a problem across different parts of a codebase, review architectural decisions, write and run tests, find inconsistencies, and work through security issues. Revisiting things I had already built & changing them when evidence shows they weren’t good enough.

There is a considerable amount of work between a feature appearing to work and actually having enough evidence to trust it. Building products has made me much more aware of that gap.

The honest answer also includes all of the unfinished work. All of the failed projects, the ones I scrapped because the demand wasn't there, the ones I thought were brilliant, but the market didn't.

Building software is its own accomplishment.
But proving that people will pay for it is an even bigger accomplishment.

I think those distinctions matter when we talk about what we’ve accomplished with AI.

My involvement has been very hands-on. I’ve had to define what I wanted, decide how the pieces should fit together, provide context, review the results, question assumptions, and keep working through problems.

AI can produce a convincing answer, while still missing an important constraint. Sometimes the difficult part is recognizing that before it becomes the foundation for everything else.

It has also expanded what I can personally take on. I can work across more than one product, investigate unfamiliar technical areas, and turn an idea into something, I can actually test.

That is probably my most honest answer: I’ve accomplished a great deal with it, and doing the work has made me more demanding about what I’m willing to call finished. The ability to build has expanded enormously. The responsibility for what I build still sits with me.

[Snickerdoodle]

Really interesting projects you're working on!

I’ve used AI to build several substantial software products, including taking software into production and reaching a substantial ARR before exiting.

Across these projects, I’ve worked on the interfaces people interact with, the systems running behind them, databases, integrations, permissions, automated agentic workflows, testing, deployment, and ongoing maintenance.

AI has been reliable part of some of the process.

I’ve also used AI to investigate my own software. To trace a problem across different parts of a codebase, review architectural decisions, write and run tests, find inconsistencies, and work through security issues. Revisiting things I had already built & changing them when evidence shows they weren’t good enough.

There is a considerable amount of work between a feature appearing to work and actually having enough evidence to trust it. Building products has made me much more aware of that gap.

The honest answer also includes all of the unfinished work. All of the failed projects, the ones I scrapped because the demand wasn't there, the ones I thought were brilliant, but the market didn't.

Building software is its own accomplishment.
But proving that people will pay for it is an even bigger accomplishment.

I think those distinctions matter when we talk about what we’ve accomplished with AI.

My involvement has been very hands-on. I’ve had to define what I wanted, decide how the pieces should fit together, provide context, review the results, question assumptions, and keep working through problems.

AI can produce a convincing answer, while still missing an important constraint. Sometimes the difficult part is recognizing that before it becomes the foundation for everything else.

It has also expanded what I can personally take on. I can work across more than one product, investigate unfamiliar technical areas, and turn an idea into something, I can actually test.

That is probably my most honest answer: I’ve accomplished a great deal with it, and doing the work has made me more demanding about what I’m willing to call finished. The ability to build has expanded enormously. The responsibility for what I build still sits with me.

Yeah, I agree with this. AI makes it way easier to build, but actually testing it, proving people want it, and making sure it works properly is the harder part. I’m trying to keep that mindset with what I’m building too.

[Oni]

I was messing around with Codex and made a custom camera mod for Tokyo Xtreme Racer, at the request of @Skullmeat.

Lots of instructions needed, but it has been pretty smooth. I've been very impressed with how it was able to add UI and custom controls for things like camera positioning, blur, and FOV. I'll probably release the mod when I'm done, as it's very polished so far.

[Snickerdoodle]

I was messing around with Codex and made a custom camera mod for Tokyo Xtreme Racer, at the request of @Skullmeat.

Lots of instructions needed, but it has been pretty smooth. I've been very impressed with how it was able to add UI and custom controls for things like camera positioning, blur, and FOV. I'll probably release the mod when I'm done, as it's very polished so far.

Very cool, recently I used ai to build my friend's website nicklynforged.com