Let the Code Write Itself
Because You're Busy Enough Already 🤖
So who is this talk for?
-
Burnt-Out Builders
- Side Project Enthusiasts
- Time-Strapped Creators
- Curious Skeptics of AI
- Those Searching for Work-Life Balance
Chat Interfaces
IDE Chat Intefaces
Striking a balance
Model
Context
Prompt
Context
Would I be able to solve this problem with these files and this prompt?
Not enough context, the model may start to fill in the gaps with potentially irrelevant data that it was trained on.
Too much context, the model will need to "guess" which pieces are relevant.
Prompt
Too high-level, focus too much on what to do rather than on how to do it.
Too low-level, has too much detail and doesn't leverage the LLM:
Write a function that sorts a list of products.
Write a Python function that sorts a list of products, where each product is represented as a dictionary with at least the keys 'name' (a string), 'price' (a float), and 'rating' (a float from 1.0 to 5.0). The function should take two parameters: the list of products and a sorting key (either 'price' or 'rating'). The sorting should be in ascending order unless an optional boolean parameter
descendingis set…
Model
Speed vs Quality vs Cost (choose two)
Prompt with a Purpose
-
Every input token contributes to the final result
-
Use concise language to activate model "knowledge"
-
Understand how to leverage the models that you are working with.
-
Take the time to reflect on how to improve the original prompt instead of trying to "convince" the model to "change its mind".
What coding assistants do well
💻 Writing Code
- Scaffolding, boilerplate, repetitive logic
- Generating tests
📝 Documentation
- Docstrings, inline comments, API references
🔍 Debugging & Optimization
- Explaining errors & suggesting fixes
📊 Data Transformation & Scripting
- Parsing and converting file formats
🧠 Learning & Research Assistance
- Summarizing docs,
- Explaining unfamiliar code
🦆 Code Review Assistance
What Developers Should Prioritize
🏗 Architecture Design
- Decision-making, trade-offs, and scaling strategies
🧩 Critical Thinking
- Debugging complex issues that require deep context
✅ High-Level Code Review
- Ensuring correctness, business logic, security risks
🔧 Managing Technical Debt
- Long-term refactoring, planning upgrades & migrations
🎨 Innovative Coding
- Experimenting with new ideas, algorithms, and approaches
