Overwhelming reliance on itemized lists, pseudocode, and filler comments

[beratna-steve]

• Copilot Chat Extension Version:
• VS Code Version: 1.90.1
• OS Version: Windows 11
• Logs:

Steps to Reproduce:

1. Ask Github Copilot Chat to answer a question
2. Don't waste tokens and time begging it not to respond with aligned nonsense

I'm not sure if I'm on the wrong side of some kind of A/B testing campaign or not. I did search for this issue and am trying really hard not to complain. After using Copilot Chat in VS Code for 6 months, successfully on large blocks of code, this past Thursday morning Copilot starting limiting how much code it will even accept (well below any token restriction I've seen, I checked). It now responds, on even small function level requests, with very low quality itemized lists, pseudocode, "basic" implementations, and bold statements on what it assumes. How many logs do you want?

[beratna-steve]

Some feedback here would be great. It's still wasting lots of bandwidth and tokens on itemized lists with the extra "feature" now of making lists after I specifically tell it NOT to use lists. Whatever you guys have done in the past few weeks has essentially ruined the entire experience and made this a useless tool.

[VSCodeTriageBot]

Thanks for creating this issue! We figured it's missing some basic information or in some other way doesn't follow our issue reporting guidelines. Please take the time to review these and update the issue.

Happy Coding!

[beratna-steve]

I'm surprised you didn't respond with an itemized list.

[bhavyaus]

@beratna-steve could you please provide samples of the queries you were using?

[beratna-steve]

Here's one where it just steams ahead with a list.