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

Kite is saying farewell and open-sourcing its code

kite.com

Post reply on HN
Print view View original

41–50 of 547 posts

Re: Kite is saying farewell and open-sourcing its code

#41

Post by candiddevmike »

ShamelessC wrote: post #28
jrpt wrote: post #18

"Our diagnosis is that individual developers do not pay for tools. Their manager might, but engineering managers only want to pay for discrete new capabilities, i.e. making their developers 18% faster when writing code did not resonate strongly enough." I never used Kite, but I've tried Github Copilot twice, and found it marginal at best (and distracting at worst - which is why I turned it off both times). If Kite wa…

Your anecdote is trivially rebutted with another. I tried Github Copilot _more than twice_ (gasp), and now pay 10$/month for it. Happily.

How does your workplace/compliance officer feel about you using it?

Re: Kite is saying farewell and open-sourcing its code

#42

Post by deforciant »

jrpt wrote: post #18

"Our diagnosis is that individual developers do not pay for tools. Their manager might, but engineering managers only want to pay for discrete new capabilities, i.e. making their developers 18% faster when writing code did not resonate strongly enough." I never used Kite, but I've tried Github Copilot twice, and found it marginal at best (and distracting at worst - which is why I turned it off both times). If Kite wa…

Paying for copilot :) at least in go it’s great to write tests and sometimes some smaller functions :) totally worth paying for it, even from your own pocket if the company wouldn’t allow expensing it

Re: Kite is saying farewell and open-sourcing its code

#44

Post by cowmix »

ivalm wrote: post #40
jrpt wrote: post #18

"Our diagnosis is that individual developers do not pay for tools. Their manager might, but engineering managers only want to pay for discrete new capabilities, i.e. making their developers 18% faster when writing code did not resonate strongly enough." I never used Kite, but I've tried Github Copilot twice, and found it marginal at best (and distracting at worst - which is why I turned it off both times). If Kite wa…

I pay for github co-pilot. It seems surprisingly bad for typescript and excellent for python.

I tried it. Ironically, it was pretty good for Powershell and so so for Python (in my case at least).

Re: Kite is saying farewell and open-sourcing its code

#45

Post by ShamelessC »

lolinder wrote: post #10

> The largest issue is that state-of-the-art models don’t understand the structure of code, such as non-local context. Depending on how local he's talking, this isn't really true of Copilot. In my experience it will use context all the way up to the top of the file, even in very long files. And at least the Rust version even seems to look at the imports—if you have a use declaration it will actually correctly build a…

I had a similar thought. I think in this instance, they mean something like a labeled/supervised training approach where the model is given not just the tokenized code, but also perhaps is grounded in the possible structure (indentation vs. semicolons, function-scope, etc.).

My understanding is that copilot is largely a self-supervised approach. They feed a massive body of (somewhat noisy) code into the model. The model really does learn a lot of structure on its own and this is a testament to deep learning on noisy datasets.

I'm guessing the "hooks" that they have already from IDE's, language-servers, etc. _are_ quite "structure-aware" - so they want the predicted structure as well as the code, so they can improve the typing experience beyond line-completion.

I think the estimation of 100 million for such a task is maybe too high? I don't know - it feels like you could actually get quite close to such a system by simply using thousands of custom prompt engineering tricks that prepend structure examples to the prompt?

Re: Kite is saying farewell and open-sourcing its code

#49

Post by svnt »

throwthere wrote: post #29

> Our diagnosis is that individual developers do not pay for tools. Throwing salt on the wound here but that’s just false. I mean, there’s copilot and it’s alternative that I can’t think of the name right now. more broadly there’s Jet brains ides, visual studio, Productivity apps, etc. look at product hunt or appsumo or popular show hns. Devs pay for tools, just not Kite. Edit: I should clarify, enough devs pay for t…

You think copilot is self-sustaining/profitable?

Re: Kite is saying farewell and open-sourcing its code

#50

Post by candiddevmike »

arcturus17 wrote: post #25
jrpt wrote: post #18

"Our diagnosis is that individual developers do not pay for tools. Their manager might, but engineering managers only want to pay for discrete new capabilities, i.e. making their developers 18% faster when writing code did not resonate strongly enough." I never used Kite, but I've tried Github Copilot twice, and found it marginal at best (and distracting at worst - which is why I turned it off both times). If Kite wa…

Loads of devs I’ve spoken to, from junior to principal level, absolutely love Github Copilot though. Don’t know who is paying it for them, nor if Kite was significantly worse, but I think that at least Copilot has a brilliant future ahead of it.

Until Microsoft grants Copilot users blanket protection over copyright claims from Copilot generated code, I wouldn't even think of touching it.
Post reply on HN