In a world of infinite possibility, only creativity will shine.

Let me tell you about my first $100,000 in AI. The story starts back in 2022, when I began digging into what these models could do. I could see enough early signals to believe AI was going to become a huge part of our future, and I wanted to understand it. I didn’t have much of a roadmap. I spent a lot of time figuring things out myself, regularly putting in 12 or more hours a day learning, experimenting, and building. I got familiar with different models, tools, and ways of putting them together. I built side projects, helped people with things they were trying to create, and kept exploring whatever new possibilities opened up. It took roughly two to two and a half years before that started turning into meaningful income. I had other payments coming in that allowed me to keep going, and I lived well below my means. There was definitely a ramen diet involved. I made sacrifices because I enjoyed the work and believed the skills I was developing would eventually have a place. As more people started entering the space, those skills began opening doors. People wanted to understand what AI could do for them. They needed help building things, choosing tools, and figuring out why the systems they were putting together weren’t working. By then, I’d spent thousands of hours getting my hands dirty with those same questions. One of my early roles started with moderating a Discord community. As people got to know what I could do, I became more involved in helping them understand and use AI. When that same company rebranded, they brought me on as a full-time educator. I stayed with the community, but the experience I’d built opened the door to a bigger role. There were development projects, teaching opportunities, consulting, and other paid positions along the way. Each one helped me understand where I fit and what people actually needed from me. These days, people usually come to me when they have a problem they don’t know how to solve with AI. Maybe their agents aren’t behaving. Maybe they know what they want a workflow to accomplish, but don’t know how to design it or make the pieces work together. That’s where I’ve found my role as an AI solutions architect. The field has changed enormously since I started. The simple chatbot experiences that first caught my attention have grown into systems that use tools, coordinate agents, and have conversations like the phone call I used to put this post together. Learning alongside that change has been a huge part of the journey. Each new tool and project gave me more experience to draw on when someone came to me with a problem. Now, here’s the part I need to be honest about: I haven’t actually reached that $100,000 milestone yet. Between the larger roles and smaller projects along the way, I’m getting really close. I’m still on the journey toward my first $100,000 in total AI-related earnings, and I can see the path much more clearly now. I have my own agency, building AI solutions, websites, software, and workflows for clients. I’ve developed a network over the years, and I’ve found work that makes use of the skills I spent so long building. When I look at what I’ve earned so far, I see years of learning, experimenting, building relationships, and gradually finding my place in this industry. There are opportunities in AI that take time to turn into income. Mine took years of putting in the hours, figuring things out, and finding where I could be useful. I’m sharing this while I’m still building, because that part of the story deserves to be told too.

Live with @mikeygnft watching Dev Day @OpenAI xitter.cf/i/broadcasts/1wxWjlOYy…

Finally got the green light to talk about my time testing GPT-5.6 Sol. Been using it on a couple real projects for a little while and it’s a noticeable step from 5.5 in a few places that actually matter day-to-day. The biggest thing for me: it just doesn’t quit. Messy requirements, changing scope, incomplete context. It keeps iterating and course-correcting instead of hitting a wall or hedging. That persistence alone makes it feel different. It’s also way better at understanding sub-agents and handing off work without losing the thread. Feels more collaborative than Fable (which sometimes just disappears into the task and comes back with something that needs heavy editing). Sol feels like it works with you more.On the code side it caught a few edge cases that Fable had glossed over, which was genuinely useful. Terra variant is fast and capable enough that I can see myself routing a lot of the grunt work there. Excited about the cost/speed combo. Downsides so far: still leans pretty technical even when I ask it not to, and yeah the token usage is real. The absolute smartest model I’ve touched, The diligence and reliability feel like a real workflow upgrade.

And a lot more coming. 🫡

Updates from the GlyphGarden: We're on Day 38 right now. The earliest glyphs just started branching for the first time. There are still 327 that haven't been born yet. Free to mint. One new glyph every day, all year long. glyphgardens.art/