Can't live without AI assisted coding anymore 👇🏾 #github #copilot #dotnet #entityframework

Probably no need for AI here: how many times have you created a Posts/Todo project? Clipboard history would be enough 🤭

Hey @davidfowl! Quick question (hopefully): I've seen others use the comment technique... If I start with a comment, Copilot just keeps writing more and more comments 😂 Is there some trick to letting it know my comment is the prompt for things I expect to follow?

If you start writing the class it'll finish it. The hardest thing is to built pauses into your development flow to let it think and then spit out code.

The techie in me likes this, so much happening there. The conscious person in me hates wasting network, CPU, gpu, cooling, water, ... for a simple wizard in the IDE powered by local CPU cycles.

I just finished another project in GPT-4 Advanced Data without writing any code, Eying how I can blend with Copilot soon. Prefer prompting code if -Expansive Solution Space: Ideal for new projects with undefined boundaries or when there is intrinsic value in exploring multiple solutions. Ideal for new projects. -Non-Specialist Scenarios: When the language, framework, or tools are outside your expertise, Prompt Coding can bridge the gap. -Polymathic Problem-Solving: Leverage interdisciplinary skills to tackle problems that are beyond the reach of specialized approaches. -Organic Development: If you favor a fluid, evolving methodology where you iteratively build up high-context states that eventually crystallize into functional code. - Complexity-First Approach: Start with a complex, -multifaceted understanding of the problem, then distill it into simpler, executable components. - Low Testing Overheads: When the cost and complexity of testing are minimal, allowing for rapid iterations and adjustments. When to Opt for AI-Assisted Coding - Prescriptive Solution Space: Best suited for projects where the end goal is clearly defined and established patterns are sufficient. Ideal for existing projects. - Expertise-Driven: When you're operating within your realm of expertise, AI can automate routine tasks, allowing you to focus on unique challenges. - Iterative Methodology: If you're looking to develop your project in increments, each building upon the last in a logical progression. - Simplicity-First Approach: Begin with the most basic, functional iteration of your solution, gradually introducing complexity only as necessary. - High Testing Costs: Effective when the stakes of a mistake are high, requiring code to be as accurate as possible upon initial implementation.

My condolences

Coding without it (even with resharper/jetbrains ide) feels like coding naked, just like coding without intellisense felt before it.

Definitely a time saver. It helps with memory, however, I notice when my internet is out, how lazy it has made me.

I’m looking for what I’ll do in the next couple years. How long until AI replace us?

A while yet, some of us before others....