~/ai-team-event/2026 $
how far simple prompts
can take you
Well Ronfim
Lead Software Engineer
Panda Team (Store Team)
Ex Otter (Bookshelf Team)
Lead Software Engineer
Panda Team (Store Team)
Ex Otter (Bookshelf Team)
vitalsource · ai team event
~/intro/audience.md
02
$who is this for
Common worries you may have heard — or thought yourself.
- I don't know which problems to tackle with AI
- I don't know how to incorporate it into my workflow
- I don't know what to prompt
- How do I trust what AI generated?
~/journey/01_the_decision.md
03
$the decision
use AI for everything — even where it seems unlikely to help.
goal: find the limits. mine and AI's.
~/journey/02_myrkup.md
04
$myrkup, my laboratory
idea|shape|pricing|MCPs|code|deploy → every wall climbed with the same tool
https://myrkup.com
~/journey/03_into_work.md
05
$then I dragged it into work
- bug investigation
- learning unfamiliar codebases
- writing tickets
- PR reviews
- writing code I could have written myself
a deliberate choice: get exposed, get frustrated, find the edges.
▶ live demo
simple prompts
~/wrap/lessons.md
07
$what I learned
01.
inertia is the blocker — not capability.
any prompt beats waiting for the perfect one.
02.
AI's real value is thinking with you, not for you.
leave the conversation in a better place than you started.
03.
the bottleneck moves: from generating to validating.
cheap to make. expensive to trust.
04.
more iterative than you may think.
start simple, refine as you go.
~/wrap/what_worked.md
08
$what worked for me
- start with a brain-dump prompt — even a bad one
- the more complex and unique the task, the more detailed your prompt should be
- break big work into chunks you'd actually tackle yourself
- validate by asking, challenging, regenerating in different forms
- find the limit. get frustrated. pivot.