If you had asked me back in early 2024 what my AI strategy was, I wouldn't have told you I had one. Like most people, I was just shifting between random browser tabs, experimenting with free tiers, and trying to see what stuck.
Over the last two years, that random curiosity has turned into an intentional, highly strategic multi-tool ecosystem. Instead of a backward-looking retrospective, I am treating this post as a live chronicle of my stack, my training, and how I am actively navigating the cutting edge of AI development—ordered with my latest updates right at the top.
The Evolving Log (Latest First)
📍 2026 (Current Stack) — The Age of Agents & Complete Orchestration
Everything shifts to autonomous development. I am moving away from single-prompt completions to running fully agentic workflows, completely transforming how I handle codebases.
- AIxTech by AI Singapore: Right now, I am enrolled in the flagship AIxTech Programme by AISG and IMDA. This moves me straight into production-grade multi-agent orchestration. Crucially, through the program, I have unlocked access to ChatGPT Pro. This completes my current live loop: when my Copilot agent credits dry up during intense programming loops, I instantly hop into ChatGPT Pro to pick up the slack without losing an ounce of momentum. Open AI Singapore AIxTech
- The Credit Bottleneck & The Claude Pivot: I discover Agents mode inside GitHub—allowing the AI to autonomously navigate entire multi-file codebases and self-correct. Because I use it heavily, I rapidly hit the modern developer wall: running out of agent credits. To keep moving, I subscribe to Claude Pro (USD 25) for a month. This lets me leverage Claude’s advanced reasoning to offload deep codebase architecture mapping while my other limits reset. Open Anthropic Claude
📍 2025 — Upgrading to Premium & Workflow Automation
Ready to scale up my coding velocity and operational efficiency, I move past basic chat interfaces and make my first commercial commitments.
- Hatch AI Bootcamp: I dive into Hatch Singapore's free community literacy program. This training pushes me past simple text prompting to teach me how to automate repetitive reports, build custom workflows, and handle data safety properly. Open Hatch Academy
- GitHub Copilot Pro: I step into premium developer environments, starting with a trial before officially committing to a monthly subscription (USD 10/month) for integrated IDE support. Open GitHub Copilot
📍 2024 — Building the Foundation
Instead of blindly guessing my way through these new tools, I start getting structured. I am using free access paths to pick up frameworks and establish clear workflows.
- The Free Tier Split: I begin strictly compartmentalizing my free tool usage to stay organized. Google Gemini becomes my main playground for personal projects and blogging, while ChatGPT is designated exclusively as my "work buddy" for text data analysis and debugging.
- Google AI Essentials: Taking this structured course gives me my first real look at core prompt engineering and everyday productivity hacks. Open Google AI Essentials
Where the Stack Stands Today
My current daily pipeline is entirely about matching specific model strengths to my available resource caps and credit limits:
| Tool | My Tier | Primary Use Case | Monthly Cost |
|---|---|---|---|
| Google Gemini | Free | Personal projects, daily organization, and blogging. | SGD $0 |
| GitHub Copilot | Pro | Inline IDE support and initiating autonomous agent runs. | SGD $12.94 (estimated) |
| ChatGPT | Pro (via AIxTech) | Core work engineering, data debugging, and secondary agent backup. | SGD $0 (on aixtech credits) |
| Anthropic Claude | Pro (On-Demand) | High-context fallback for deep code reasoning when primary credits are spent. | SGD $0 (Not subscribed) |
