My Boss Thinks I'm a Data Analyst. I'm Actually Building an AI Empire

You know that moment when life throws something at you, and you're not sure if it's an opportunity or a trap? For me, it landed somewhere in the delicious middle. I recently switched jobs and no, I'm not back as just another data analyst grinding dashboards and working for clients. What hooked me was the chance to actually coming out AI ideas for the company. I have prepared two proposals, ready for the pitch. Just need to form a team to present and power point slides.

Here's the thing: I've been treating AI as a tool for a while. But lately? I'm diving into AI as agents. With an 's' at the end. Like a whole damn company of them. Because if you're not thinking about multi-agent orchestration in 2025, you're still living in the Stone Age of productivity.

Back in my trading days, I had this script for my trading. Every weekend for 6 to 7 weeks wasted. Debugging, tweaking, and staying many late nights. Then I started feeding the problem to AI. Not just ChatGPT asking for recipes, I mean building the script from scratch with AI. Now it just took me 2 to 3 weekends, done!

The math is simple: more free time = more experiments = more stuff that might actually make money. I'm not getting younger, and neither is my patience for manual grind.

Naturally, I went for to this path since I already have my home lab: "Run everything locally. Full control. No censorship. Privacy." Great idea in theory. Then I checked GPU prices. RAM prices. The cost of electricity when you leave a machine running for three days straight.

Yeah, this is not happening on a hobbyist budget. Not unless I'm willing to eat instant noodles for six months. :/
Here's where things get cheeky: Currently Anthropic and OpenAI are basically throwing punches at each other in the ring, and we're the ones catching the confetti.

More tokens. Better models. Cheaper APIs. They're competing so aggressively for market share that they're practically paying us to use their stuff.
I'm about to take one of them for a test drive. Not sure if its permanent. I'll calculate the cost, measure the output, and decide if this worth as a subscription or just buy the hardware.

My plans are to run multiple agents that handle everything I need. Need to complete start a new project? Just spawn up an agent to assist me.

Ambitious? Sure. But I've already cut my scripting time by 60%. Why stop there?