Hacker News Board Hacker News, browsed like a classic message board
Live data from Hacker News

GitHub Copilot Workspace: Technical Preview

github.blog

Post reply on HN
Print view View original

1–10 of 333 posts

Re: GitHub Copilot Workspace: Technical Preview

#2

Post by azhenley »

I just wrote up my experiences of using it for a week: https://austinhenley.com/blog/copilotworkspace.html

TL;DR: Copilot Workspace is a great concept. But the UX and framing is entirely wrong and sets the wrong expectation for users. The current prototype is very slow (5+ minutes for one-line code changes) and the generated code is often buggy or has nothing to do with the specification. It doesn’t help me understand the code. The generated plan is usually good. ChatGPT does a much better job in my head-to-head comparisons (assuming I already know exactly what code is relevant). I'm still optimistic of where it can go from here.

I recommend everyone sign up to try it and give the team feedback.

Re: GitHub Copilot Workspace: Technical Preview

#3

Post by studenthrow3831 »

Student here: I legitimately cannot understand how senior developers can dismiss these LLM tools when they've gone from barely stringing together a TODO app to structuring and executing large-scale changes in entire repositories in 3 years. I'm not a singulatarian, but this seems like a brutal S-curve we're heading into. I also have a hard time believing that there is enough software need to make such an extreme productivity multiplier not be catastrophic to labor demand.

Are there any arguments that could seriously motivate me to continue with this career outside of just blind hope that it will be okay? I'm not a total doomer, currently 'hopium' works and I'm making progress, but I wish my hopes could at least be founded.

Re: GitHub Copilot Workspace: Technical Preview

#4

Post by rgbrenner »

azhenley wrote: post #2

I just wrote up my experiences of using it for a week: https://austinhenley.com/blog/copilotworkspace.html TL;DR: Copilot Workspace is a great concept. But the UX and framing is entirely wrong and sets the wrong expectation for users. The current prototype is very slow (5+ minutes for one-line code changes) and the generated code is often buggy or has nothing to do with the specification. It doesn’t help me understan…

Thanks for writing that up. It's what I suspected from my use of CoPilot.

I love copilot as an autocomplete tool... but it frequently gets things wrong, and using the chat feature to ask it to complete some task usually just generates code that breaks things. So until that improves, Im skeptical a workspace tool would work.

Workspace seems like an awesome idea though.. once the tech is further along.

Re: GitHub Copilot Workspace: Technical Preview

#5

Post by HideousKojima »

studenthrow3831 wrote: post #3

Student here: I legitimately cannot understand how senior developers can dismiss these LLM tools when they've gone from barely stringing together a TODO app to structuring and executing large-scale changes in entire repositories in 3 years. I'm not a singulatarian, but this seems like a brutal S-curve we're heading into. I also have a hard time believing that there is enough software need to make such an extreme prod…

Writing code is one of the less important parts of your job as a developer. Understanding what the client actually wants and turning that into code is where the real difficulty lies for most software development. Also, I doubt we'll be able to overcome the "subtly wrong/insecure" issues with a lot of LLM generated code

Re: GitHub Copilot Workspace: Technical Preview

#7

Post by ThrowawayTestr »

studenthrow3831 wrote: post #3

Student here: I legitimately cannot understand how senior developers can dismiss these LLM tools when they've gone from barely stringing together a TODO app to structuring and executing large-scale changes in entire repositories in 3 years. I'm not a singulatarian, but this seems like a brutal S-curve we're heading into. I also have a hard time believing that there is enough software need to make such an extreme prod…

AI can't prompt itself. The machines will always need human operators.

Re: GitHub Copilot Workspace: Technical Preview

#9

Post by idle_zealot »

studenthrow3831 wrote: post #3

Student here: I legitimately cannot understand how senior developers can dismiss these LLM tools when they've gone from barely stringing together a TODO app to structuring and executing large-scale changes in entire repositories in 3 years. I'm not a singulatarian, but this seems like a brutal S-curve we're heading into. I also have a hard time believing that there is enough software need to make such an extreme prod…

In my experience these tools continue to be really bad at actually translating business needs, that is, specific logic for handling/processing information, specific desired behaviors. They're good at retreiving general patterns and templates, and at transforming existing code in simple ways. My theory is that translating the idea of how some software is supposed to function into working code requires a robust world model and capacity for deep thought and planning. Certainly we will create AI capable of these things eventually, but when that happens "did I make a mistake in studying computer science?" will not be your greatest concern.

Re: GitHub Copilot Workspace: Technical Preview

#10

Post by packetlost »

studenthrow3831 wrote: post #3

Student here: I legitimately cannot understand how senior developers can dismiss these LLM tools when they've gone from barely stringing together a TODO app to structuring and executing large-scale changes in entire repositories in 3 years. I'm not a singulatarian, but this seems like a brutal S-curve we're heading into. I also have a hard time believing that there is enough software need to make such an extreme prod…

The hardest part of software development is not writing code, full stop. It never has been and it never will be. The hard part is designing, understanding, verifying, and repairing complex systems. LLMs do not do this, even a little bit.
Post reply on HN