I let GitHub Copilot be the Pilot for my C# code
Video Statistics and Information
Channel: Nick Chapsas
Views: 213,107
Rating: undefined out of 5
Keywords: Elfocrash, elfo, coding, .netcore, dot net, core, C#, how to code, tutorial, development, software engineering, microsoft, microsoft mvp, .net core, nick chapsas, chapsas, clean code, dotnet, github copilot, github co-pilot, copilot, github c#, github, github copilot C#, github copilot .net
Id: XceIKXeIrD0
Channel Id: undefined
Length: 13min 59sec (839 seconds)
Published: Mon Nov 01 2021
Please note that this website is currently a work in progress! Lots of interesting data and statistics to come.
> inb4 inevitable "programmer jobs are at risk"
This alarm has been sounded anytime a tool that made programming easier came along for the last 20 years. The opposite has always been true. Github copilot will not make you understand the underlying concepts of the code that it auto completes, and therefore isn't much better than copying and pasting off stack overflow in terms of writing quality software.
I've seen co workers struggle to write code, completely ignoring perfectly valid Github copilot suggestions simply because they did not understand the language or the feature they were working on properly.
Github copilot only makes your more efficient at what you know, it does not solve problems or code solutions for you.
EDIT: This is not to say that programmer jobs aren't at risk from AI - they are, just not right now and probably not before AI can automate away almost every other job.
If anyone out there wants to try a similar tool to this but for shorter code snippets try tabnine. It's already free and publically available, been using it for a month now and it's made me significantly faster at my job.
I was using this earlier today while messing around on some LeetCode problems.
If you let it, it basically gives you the answers outright.
I mostly used it for filling out the annoying little helper functions that we've all had to write out a million times. Write the name, maybe a simple comment, and BAM an answer pops out that is 90% of the time correct.
Benefits: Ease of entry and baseline quality of code will likely go up
Detriments: Fewer creative solutions, less job security and more risk of outsourcing, and any oversights or security flaws replicated across a wide variety of software
Programmers are both smart enough and dumb enough to replace ourselves with AI.
My real comment is: this looks pretty cool.
My snarky comment would be: "what are you doing, Dave?"
Honestly what he's showing here isn't that impressive because all of this stuff has been done to death. There's a plethora of example project and tutorials that show exactly what you need to do implement these examples. Essentially what Copilot is doing here is guessing which code to copy paste, to be fair the guesses are good though.
The CRUD service is a poor example because in many cases you could just switch it out with a GraphQL service and then as a back-end developer you get to focus on the important stuff. And also CRUD services usually aren't all that complex.
What would be interesting to know is how Copilot performance when you're implementing complex business logic in a service.
Itβs a tool and almost any tool needs a human element.
this is nothing more than a glorified code snippet engine