Hacker News Reader

Yes.

It feels like how auto complete should actually work, but I haven't seen any exceptional performance from Copilot.

Like with Phind, the longer the code examples, the more errors.

But it feels a more fluid when typing boilerplate heavy code, which is nice.

Pro: writing lots of boilerplate is less painful.

Con: writing lots of boilerplate is less painful.

(So you get more boilerplate code)

Are you using GPT 3.5 or 4? Huge difference in quality with 4.

3.5 — I've found it tends to hallucinate when dealing with APIs as well.

Yeah it's pretty useless. Spend the $20 for a month to try GPT-4, you'll be surprised. It definitely saves me more than $20 of time each month.

Interesting, I've found it's JS support pretty great. Wonder how much the codebase has a factor and how well things are typed/documented.

Thing is, Copilot doesn't seem to take the code of the actual installed packets into account.

The moment you work with something a bit more obscure, it just guesses.

Especially with TypeScript it's kinda funny, since IntelliSense which is right there, in the IDE, tells the truth, but is completely ignored by Copilot.