Prompting for Impact: A GitHub Copilot Workshop
This afternoon we hosted Patrick Robinson for a workshop on getting more out of GitHub Copilot. The title “Prompting for Impact” caught our attention, and what followed was an honest, practical discussion about using AI tools effectively in development workflows.
The Real Talk About AI Tools
Patrick kicked things off with refreshing honesty: this wasn’t going to be about “vibe-coding” or blindly trusting AI to do your job. Instead, he focused on how developers can use tools like Copilot to augment their abilities while maintaining ownership of the code they produce. As he put it, “AI won’t take your job, but a developer equipped with AI might.”
The workshop emphasized building context for better results. Patrick’s key principle: “The more complete the context, the more accurate the outcome will be.” This resonated with several attendees who shared their own experiences with AI tools producing inconsistent results when given vague prompts.
Beyond Just Code Generation
What made this workshop valuable was Patrick’s emphasis on using AI for more than just writing code. He demonstrated using Copilot for:
- Learning new frameworks through Q&A interactions
- Brainstorming solutions to problems
- Creating documentation from code discussions
- Even debugging his home plumbing (yes, really - he shared a story about using ChatGPT to diagnose a water leak)
The hands-on portion involved adding a testing framework to a React app. While the live demo had mixed results (as live demos often do), it illustrated important points about iterative prompting and the importance of verifying AI-generated code with actual tests.
Custom Instructions and Chat Modes
Patrick introduced the concept of custom instructions and chat modes - ways to give Copilot persistent context about your project and coding preferences. He showed us “Beast Mode,” a chat mode from the awesome-copilot repository that aggressively tackles large tasks. While entertaining, it served as a good example of why you might want different modes for different situations.
Several attendees were particularly interested in how custom instructions could help maintain consistency across team projects. Brandon mentioned he’s been using similar techniques with Claude Code to create PRD documents and manage tasks systematically.
Community Insights
The best part of the afternoon was the organic discussion that emerged. Attendees shared their own experiences:
- One member described using Copilot to program a robotic arm for computer vision tasks
- Another talked about creating “boards of directors” with AI personas for decision-making
- There was an interesting (if slightly uncomfortable) discussion about the privacy implications of AI’s ability to aggregate publicly available information
The conversation meandered through topics like using AI for project management, the importance of breaking down complex tasks, and whether saying “please” and “thank you” to AI actually matters (jury’s still out on that one).
Looking Forward
As the workshop wrapped up, there was genuine interest in exploring these tools further. A few attendees exchanged contact information to continue the conversation, and there’s already talk of future workshops diving deeper into specific use cases.
Patrick left us with this thought: “How can we use AI to increase our cognitive ability as humans?” It’s not about offloading thinking to machines, but about augmenting our own capabilities. That’s a message that resonates with what we’re trying to build here at ATC - a community that embraces technology while maintaining human creativity and critical thinking.
Join Us
If you’re interested in these kinds of practical, hands-on discussions about development tools and practices, come check out our space. We meet regularly, and there’s always room for more voices in the conversation. Whether you’re AI-curious or AI-skeptical, you’ll find thoughtful discussion and real-world experimentation happening here.
Special thanks to Patrick for sharing his knowledge and to everyone who attended and contributed to the discussion. These community-driven learning experiences are what make our local tech scene special.
The Alamo Tech Collective hackerspace is open to members 24/7. Visit us to learn more about membership and upcoming events.