A few weekends ago was one of those rainy weekends I look forward to. A weekend where, honestly, the best thing for you to do is stay indoors. Now don’t get me wrong, I love to touch grass from time-to-time, get out, and adopt my role as weekend entertainment coordinator for my 14-year-old. However, it’s weekends like this that give me the chance to learn something or tap into a pet project with little guilt—a bit of weekend upskilling.
My Weekend Upskilling Challenge
Over the years, particularly during my work with Academic Writers Studio, I made a small list of things that I told myself I’d get to. These items focused on personal development and technical skills. One task involved learning how to build a bot in Slack. Another centered on learning how to deploy code to a cloud server.
The main thing that kept me from ever getting to those tasks was life—a lack of time and shifting priorities.
How AI Changed My Weekend Upskilling Approach
Nevertheless, between my MBA and the advent of AI, I’ve amassed a set of skills and resources. These tools have allowed me to achieve things I couldn’t imagine doing a few years ago. Going back to school really helped me think about things in new ways. It improved my ability to connect the dots.
One of many “connect the dots” moments has been using my fundamental understanding of debugging to leverage AI for coding, even if I have ZERO experience in the programming language. My background in coding—though not necessarily in the latest languages—gives me a crucial advantage during my weekend upskilling projects. I can recognize when AI tools get stuck in repetitive loops.
These loops often happen when the AI keeps suggesting the same failed approach repeatedly. Where others might keep feeding the same prompts back, I spot the pattern immediately. My debugging experience kicks in. I recognize the symptoms: circular logic, repeated error messages, or suggestions that don’t address the root cause.
When this happens, I guide the AI down a completely different path. Instead of pushing harder on the broken approach, I reframe the problem entirely. Sometimes this means breaking the task into smaller components. Other times, I ask the AI to explain its reasoning first, then redirect from there. This approach transforms my weekend upskilling sessions from frustrating dead ends into productive problem-solving opportunities.
What I Built During My Weekend Upskilling Session
With that approach, I was able to finally build that Slack bot. I also deployed code to Google Cloud Run and checked both boxes! Along the way, I exposed myself to Visual Studio Code, Python, Google Cloud Run, Amazon Q (which has a seemingly magical AI-driven auto-suggester for code), and Slack Dev.
The weekend upskilling experience felt like solving a puzzle with AI assistance. Each new error message became a learning opportunity. Amazon Q would suggest solutions before I fully articulated the problem. Google Cloud Run’s deployment process surprised me with its simplicity. By Sunday evening, I had working code running in the cloud.
The final result
My efforts produced an AI Slackbot that, unlike their baked-in version, can be set to more advanced models. Most baked-in AI tools use ‘fast’ (i.e. cheap) models that are good for mundane tasks, synonyms, and simple rewrites, but are pretty dumb beyond that. It also has behaviors that stay out of your way like threading its responses instead of filling your main chat window.
It can also accept images and screenshots! And because it’s built on Google Cloud Run without active instances (at the expense of some initial response speed), it’s a pay-as-you-go system.
What task or to-do have you been putting off that you could weave some of your own upskilling into?
If you’re interested in installing this bot yourself, check out the Github repository I made for it.
