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

Launch HN: CodeComplete (YC W23) – Copilot for Enterprise

news.ycombinator.com

Post reply on HN
Print view View original

21–30 of 69 posts

Re: Launch HN: CodeComplete (YC W23) – Copilot for Enterprise

#22

Post by DeathArrow »

sebzim4500 wrote: post #6
anxtyinmgmt wrote: post #2

So if Microsoft/Github ever offers an on-prem version, what will be the advantage of using your product?

Sounds like they'll finetune the model on your existing codebase. Even if Microsoft/Github offers this functionality, it's the kind of thing where it may be worth paying a premium to have good customer support telling you exactly how to make use of this. It's like how there are plenty of people paying for DB software which is 'worse' than mainstream free ones, but they really like the fact they can call experts who w…

>It's like how there are plenty of people paying for DB software which is 'worse' than mainstream free ones, but they really like the fact they can call experts who will tell them what they are doing wrong with query optimization, etc.

And for us, poor peasants, there's always going to be stackoverflow.

Re: Launch HN: CodeComplete (YC W23) – Copilot for Enterprise

#23

Post by lumax15 »

sqs wrote: post #15

Congrats on the launch. This seems very compelling. For anyone asking "Why can't GitHub just do this?", (1) never underestimate an extremely smart and motivated team, and (2) competition is critical and valuable. Let's celebrate when teams build products to fill big needs, instead of dismissing things because of a potential threat from a big company. If we dismiss new things, then we'll just end up in a world where b…

Couldn't agree more :)

Re: Launch HN: CodeComplete (YC W23) – Copilot for Enterprise

#25

Post by ninjaa »

taptaptoe wrote: post #12
swatcoder wrote: post #8

Earlier quoted context omitted.

It doesn't make sense to treat Copilot as a competitive benchmark where Copilot isn't even an option for policy reasons, which is what this product is targeting. Does it provide net positive support to developers? Is that support worth the licensing and maintenance costs? Those would indeed benefit from a demo, but it's also hard to demo something whose value hinges on fine-tuning. More likely, they'll need to sell l…

Thank you!! You've hit the nail on the head and our thoughts exactly. One anecdote is we offer video demos and one potential customer just said, "I don't need to see it. I know what Copilot does." We do offer pilots for customers after fine-tuning on their codebase!

Lol if Swatcoder is available they could be great at devrel

Re: Launch HN: CodeComplete (YC W23) – Copilot for Enterprise

#27

Post by jacobsenscott »

anxtyinmgmt wrote: post #2

So if Microsoft/Github ever offers an on-prem version, what will be the advantage of using your product?

At that point you would need to compete on service and price. They'll also get the first mover lock in advantage. Once a large company deploys your product to 5000 devs you are pretty much in for life. /me waves to jira.

Re: Launch HN: CodeComplete (YC W23) – Copilot for Enterprise

#28

Post by lumax15 »

DeathArrow wrote: post #18

Have you benchmarked it against Copilot? Training data is only your own data, because that might not achieve a great accuracy? Does it work with any programming language?

We don't have any rigorous benchmarks against Copilot, but we're working on building an evaluation framework to do so. We've played a bunch with traditional academic metrics for codegen (e.g. pass @ k) and found they don't correlate super well with real-world performance. Also, want to mention that we are not competing directly with Copilot. A benchmark against Copilot is useful as we further improve our product, but our main value add here is not that we perform better than Copilot, but rather that we serve a customer segment that can't use Copilot. Would love to hear any thoughts you have.

For training, we start with a capable open-source base model, augment it with a bunch of permissively-licensed repos, and then fine-tune on the customer codebase.

We currently support C/C++, Go, Gosu, Java, Javascript, Python, Ruby, and Typescript, but we're continuously adding new languages.

Post reply on HN