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

121–130 of 547 posts

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

#121

Post by IshKebab »

avidphantasm wrote: post #77
oxfordmale wrote: post #53

I disagree with their statement that individual developers do not pay for tools. I have paid for tools out of my own pocket on many occasions. However, being able to deliver code 18% faster isn't enough to fork out $9.99 a month. First of all it is relatively expensive. For that amount I can get a personal license for PyCharm. Secondly coding speed never tends to be a bottle neck for delivering a feature or a product…

This metric seems silly on its face. 10 bucks to get 18% more productivity out of a $10k per month developer? If this was indeed the case, everyone who employs software engineers would instantly pay this. Maybe they should have marketed more? Or maybe there are other problems with the technology (e.g., fears over copyright infringement?).

Yeah the problem is the "if" part. It may well be true but productivity is notoriously hard to measure and anyone making any claims about exact productivity increases is clearly pulling a number out of the air. People know this.

We have plenty of techniques that we know improve productivity (e.g. static types) but some people still don't believe it because it's really hard to prove productivity increases.

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

#122

Post by gibsonf1 »

perlgeek wrote: post #59
gibsonf1 wrote: post #47

The key issue is that ml/dl is pure statistics - there is no intelligence or learning or conceptual awareness of space-time, so that technology can never do so many things people try to do with it.

Some ml/dl tools (think gpt3) seem to be able to answer questions that we previously thought you needed gintelligence for. I think the line between "pure statistics" and "intelligence" are much more blurry than they used to be, and might go away entirely.

Actually, I think the industry is finally realizing that there is no intelligence there, especially with gpt3 which can figure out with great precision what statistically comes next, but there is zero understanding of the space-time conceptual meaning for gpt3 in that answer - its not designed to do anything but figure out statistically what is most likely to come next.

Gary Marcus has been doing a good job of exposing this: https://garymarcus.substack.com/

Another key piece of evidence, the failure of all FSD attempts trying to use ml/dl thinking its more than just statistics.

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

#123

Post by epolanski »

candiddevmike wrote: post #50
arcturus17 wrote: post #25

Earlier quoted context omitted.

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.

Same.

Never ever I'm risking breaking copyright, and I also don't like Microsoft not including their own code in the model.

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

#124

Post by imiric »

neilv wrote: post #51

Opensourcing code when you shut down has the nice effect of making it available to the world. It also has the nice effect of keeping the code available to the people most familiar with it, as they move on to other ventures.

Open sourcing when you're shutting down is only marginally better to not releasing anything.

It means that you didn't believe in open source while you were in business, and are only doing so now to score some points with your customers. There's no guarantee that someone will step up and maintain the project for you.

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

#125

Post by lolinder »

adamsmith wrote: post #83
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…

Yes, this was precisely what I was referring to. In small-enough programs (e.g. one file) Copilot has all the context. The other extreme would be something like the Chromium codebase. Because of this, Copilot looks better in quick demos than real-world use. (Though of course it is very impressive and this tech will get there, hopefully very soon!)

But what I'm saying is that it does use imports, at least in Rust. I'm assuming that somehow behind the scenes they're concatenating the contents of the imports into the prompt.

I can imagine this is easier in a language like Rust that has a really strict module system, and to be fair the project that I've been using it on is a side project that isn't over 10,000 lines of code yet. If I were up to 30 imports per file I can imagine concatenating would become much less effective.

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

#126

Post by premun »

jascination wrote: post #74
eloff wrote: post #64

Earlier quoted context omitted.

My experience with copilot has been very different. It easily pays for itself, and getting my employer (seed stage startup) to spring for it for the entire team was an easy sell. Yeah it's pretty dumb most of the time. But I know that, and I don't use code from it without carefully checking it out and modifying it. But it's still a huge help. Just the time saved writing tests alone pays for it. And I've had a few spo…

Out of interest, how are you using it to write tests? Do you just write "make a test for functionX" or something? (Don't have much experience with it)

It is amazing for typing out mock data. Say you're testing parsing of XML - it can easily suggest the the assertions over the data parsed from the XML. Example test that was 95% coming out of Copilot: https://github.com/dotnet/arcade-services/blob/61babf31dc63c...

It also predicts comments and logging messages amazingly well (you type "logger." add 7/10 times get what you want, sometimes even better), incorporating variables from the context around. This speeds up the tedious parts of programming when you are finalizing the code (adding docs + tracing).

Honestly, Copilot saves me so much time every week while turning chores into a really fun time.

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

#127

Post by einpoklum »

dijit wrote: post #109
joshvm wrote: post #99

Earlier quoted context omitted.

"Our diagnosis is that individual developers do not pay for tools." I think this is probably true. If you need a tool for your day job, your company ought to be paying for it. Some companies have slush funds for small purchases like books, but subscription costs for services would normally need to be approved. If you're a solo consultant then perhaps you'd pay for tools that make you more productive. But for personal…

If you'll allow me to go on a tangent here; The sheer volume of subscription services I've signed up for as the CTO for a startup is mind-boggling. $8 here, $19 there, $49 for something important, $99 for something essential. Some tools are easily worth it, especially when you see what is charged for other (less valuable) tools. Gitlab, Confluence, Jira, Asana, 1Password, co-pilot, codepen, sentry, jetbrains, gitlab…

> Some tools are easily worth it, especially when you see what is charged for other (less valuable) tools.

Maybe, but most of the tools you listed are not in that category IMNSHO.

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

#128

Post by hobofan »

perlgeek wrote: post #59
gibsonf1 wrote: post #47

The key issue is that ml/dl is pure statistics - there is no intelligence or learning or conceptual awareness of space-time, so that technology can never do so many things people try to do with it.

Some ml/dl tools (think gpt3) seem to be able to answer questions that we previously thought you needed gintelligence for. I think the line between "pure statistics" and "intelligence" are much more blurry than they used to be, and might go away entirely.

Some of the things GPT3 can achieve are very impressive, but once you work you work with it a little bit, you definitely feel that it just regurgitates the masses of text it has been trained on and tries to piece it together in the most cohesive ways. And for production usage, it (and similar models) have huge problems with halucination where it will confidently spit out "facts" that are just plain wrong.

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

#129

Post by Jenk »

candiddevmike wrote: post #50
arcturus17 wrote: post #25

Earlier quoted context omitted.

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.

We may well find out the answer to that when this lawsuit[0] concludes.

[0]: https://www.theverge.com/2022/11/8/23446821/microsoft-openai...

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

#130

Post by epolanski »

ShamelessC wrote: post #62
jrpt wrote: post #35

Earlier quoted context omitted.

That's why I tried it twice. I've been hearing people say they liked it. But I haven't found it very helpful, and often distracting, so I ended up turning it off. I'll probably try again next year when the models are improved to see if I feel any differently then.

Yeah, I understand. I just see a lot of people on here who seem to be deliberately looking for reasons not to like Copilot. You don't fit that stereotype, of course. So feel free to ignore the following. Developer tools have learning curves. One doesn't simply open vim/emacs for the first time with a full understanding of how to use it (or why it's a good tool to use, even). Historically, we have had _no_ problem wit…

Risking breaking copyright and not supporting a model of code laundering isn't exactly looking for reasons.
Post reply on HN