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.
Kite is saying farewell and open-sourcing its code
51–60 of 547 posts
Re: Kite is saying farewell and open-sourcing its code
#52Ten years too early? no. They got wiped out by microsoft, github copilot, and litigation issues around AI provided code.
I think you'll find that trying to make copilot in 2014 was definitely 10 years too early. Hell, even Copilot is a few years to early at the moment.
Re: Kite is saying farewell and open-sourcing its code
#53I 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 on time. I can see why Engineering Managers are not willing to pay for this.
I do wish the Kite team all the best, and I hope they can re-use their skills in products that are commercially viable.
Re: Kite is saying farewell and open-sourcing its code
#54>> It may cost over $100 million to build a production-quality tool capable of synthesizing code reliably, and nobody has tried that quite yet. $100 million is nothing tbh.
$100 million on zero product-market fit is a lot.
Re: Kite is saying farewell and open-sourcing its code
#55I've tried Kite once, and wasn't really impressed. For example, back when I tried it, it wouldn't offer any kind of autocompletion within a string. Even vim's built-in autocomplete tries to complete words for you there, based on other words you've used before.
Kite did sometimes offer some good suggestions in regular code, but it tried really hard to understand your code, and went belly-up when it didn't.
At that time, I tried some other ML-based autocompletion tool which wasn't specific to python, and which usually worked much better, except that it used far too much memory and caused regular crashes.
Maybe they improved kite since I tried it, or maybe "individuals don't pay for dev tools" isn't the whole story. Or maybe both.
Anyway, kudos for both trying and for open-sourcing the code at the end!
Re: Kite is saying farewell and open-sourcing its code
#56One last reminder that they once hijacked several open-source repos to inject advertisements for their service into the codebases. https://news.ycombinator.com/item?id=14836653
Buying something and then changing it is not "hijacking" that thing.
Re: Kite is saying farewell and open-sourcing its code
#57Condolences to the Kite team. But, congratulations, too - you have some of the highest value engineering experience in the world. I'm sure you'll land somewhere great; try and take some time off if you can afford it! Mulling over business models, and noticing the 'devs won't pay' narrative in the blog post, it's interesting to see the existing business models in AI; basically they seem to be: * API-driven cloud calls…
Devs almost always lack any kind of purchase authority. Any tool that appeals to devs needs to appeal to their management more, either showing some kind of cost savings over existing tools, increased dev productivity, or the new fangled "dev experience" where this tool, by shear awesomeness, will let devs put aside the low salary, process hell, and keep them employed.
I understand that's Kite's perspective (and yours -- "purchasing won't pay for this"), but devs are not paid meager salaries in general, and definitely might care about their code quality when it's put out in 'public' whether that be internal repos, or github.
Payscale estimates average engineering salary as having between $3,000 and $7,000 a month more in disposable income over writers -- and I would guess almost every professional writer pays for grammarly.
But, I agree that this is a new concept, and just spitballing -- right now, these sorts of linters and code formatting tools are mostly open source, so it would be some product marketing work to see if the market would actually pay.
Re: Kite is saying farewell and open-sourcing its code
#58"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."
Are there a lot of businesses where individual developer productivity, with a narrow definition of LOC per hour, is the bottleneck?
I've worked for 10 years as a web dev and the bottleneck is very often at the product management level (tickets not ready, goals changing, haven't got the credentials for the 3rd party API yet..) and a minority of the time it's my brain (yes sometimes I need to think before I write code). It's rarely how fast I can write a function. So if you make me 18% faster at something I do 1% of the time... good luck making money out of me
Re: Kite is saying farewell and open-sourcing its code
#59The 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.
Re: Kite is saying farewell and open-sourcing its code
#60"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 copilot. It saves me a modest number of minutes of time per week. That's worth a small fee.
And before someone jumps in: I and my other co-founder who also uses copilot (We are the only two in the company who do, I think, without checking) are the compliance team. We're both very senior and use copilot basically a line or three at a time as a smart autocomplete. It's still worth it.