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

Kite Responds to the Minimap and Autocomplete Issues

kite.com

Post reply on HN
Print view View original

41–50 of 75 posts

Re: Kite Responds to the Minimap and Autocomplete Issues

#41

Post by dmix »

ivanbakel wrote: post #27

An apology that starts with 4 paragraphs of self-congratulation is not a real apology at all. Nothing of the article reads as being genuinely regretful of the strategies taken - it's just sorry for being caught. >Unfortunately, while trying to make programmers’ lives a little less complicated, we instead made them more complicated. Barely an admission of guilt, and even that can't get past without patting themselves…

Eh I didn't get the "sorry for being caught" vibe. They admitted they failed to listen to early customer feedback.

But still it sounded like a marketing piece as much as an apology, which is bad because authenticity is everything after a PR blowup like this.

Repeatedly highlighting their open source creds didn't make me want to trust them more, it made me question how they didn't know it was a bad idea in the first place... and if business people were the ones making the calls instead of the OSS developers.

This was a risky thing to do, even despite the luxury of looking back in retrospect, this seems like something you have to do very carefully if you are going to integrate ≈remote cloud services into text editor plugins, let alone commercial backed features.

Clearly there was going to be a wide gap between the users expectations of a static local OSS plugin (as they almost always are) and what they changed it to. It had to be handled very carefully - if at all. They should have made a separate add-on plugin IMO and pushed it via the documentation of other plugins they own...

Re: Kite Responds to the Minimap and Autocomplete Issues

#42

Post by danso »

It is nice to see a more extensive explanation, even as the motive is to save face. But the apology, like the incident itself, feel like symptoms of a flawed strategy.

> How did we get ourselves here? We started Kite with the idea that machine learning could help eliminate the repetitive parts of programming. We spent three years building the initial product - and it works. Our software has really great completions, conveniently sorted by relevance instead of the alphabet, among other features that are proving useful to coders.

> We’re proud of the tools we’ve built - the problem we faced was finding a way to tell potential users about the thing we created. As we considered our options, we had a novel idea: buy an open source plugin, reward the author for their work, and expose new users to Kite.

Many eminently useful plugins and software have been able to endear themselves via word of mouth and user happiness; Homebrew, Bootstrap, and Atom come to mind. And plenty of programmer-optimized software can even charge good money, such as Textmate and PyCharm.

Advertising isn't a bad way to get exposure, but yeah, I do agree that Kite's approach was "novel". It'd be as if the official CDN version of React wrote console messages about how great Instagram's new Snapchat-like features are. Even if Kite's injection of self-promoting code into a popular plugin was harmless, it felt exactly like the kind of shady tactic that people cynically suspect user-data-in-the-cloud companies to partake.

The critique of Kite was not the only Kite-related article to get a huge number of HN upvotes; Kite's initial announcement and a followup about Python features got 1,138 and 553 upvotes [0] respectively. That (plus the VC funding and connections you already had) is enough to get a critical mass of interest. If Kite hasn't gotten the desired userbase a year later, advertising isn't the solution. In each of those HN discussions, as well as on Reddit, the primary concern was the cloud hosting of code. Maybe it is impossible for Kite to be full-featured as a locally-hosted product, but most users seemed unconvinced because they were apparently unable to see the value of Kite over what offline IDEs are able to do.

Kite's response shouldn't have been "the same concerns were raised for tools like Dropbox and Github [which]are now used without hesitation" [1], but to focus on a minimal viable local product that would become popular enough to have the same kind of trust/popularity that Github and Dropbox earned (hard to imagine either being successful if not for their generous free plans). Undertaking a strategy that re-emphasized people's concerns about the cloud and Kite's unclear privacy policies is just not a good look.

[0] https://hn.algolia.com/?query=kite&sort=byPopularity&prefix&...

[1] https://www.reddit.com/r/Python/comments/4erjy4/kite_program...

Re: Kite Responds to the Minimap and Autocomplete Issues

#43

Post by eridius »

bauerd wrote: post #32
dmitrygr wrote: post #26

I interviewed with Kite a number of years back. From that experience, I am entirely sure that they really had no malicious intent with any of this. That being said, I did tell them that IMHO they were severely underestimating the privacy concerns of developers. Looks about right.

They had no malicious intents? Like in they had no intentions of hiring open-source maintainers, taking over their codebases and quietly injecting their product? Point is, if we don't consider this sort of behaviour malicious then we can expect to see this more and more often.

Whether or not we consider this behavior to be malicious has no real bearing on whether they had malicious intent. I'm sure they think that adding Kite support to these plugins is great for everybody. But they have different priorities and opinions. I can easily accept that they had no malicious intent, but simply failed to realize that other people would feel differently about the changes they made. And that's definitely a problem, they should absolutely have realized that they were doing something people would not like. But when you're excited about your own product it can be hard to recognize that other people aren't.

Don't get me wrong, I think they're absolutely in the wrong here. I just also think it's important to understand motivations and reasoning instead of just attributing everything to malice. Otherwise you can't prevent this from happening again.

Re: Kite Responds to the Minimap and Autocomplete Issues

#44

Post by revmoo »

jdiaz5513 wrote: post #20

At companies like my former employer, developers were given the power and responsibility to vet their own tools. Despite somewhat strict compliance needs we were able to allow junior developers to customize their own IDEs and plugins, since personalization is crucial for productivity. One of those compliance needs was, crucially, that _source code in its full form must not leave the developer's machine except en rout…

This is a growing problem, not just with IDEs, but with the millions of LOC of vendor libs that everyone is bandying around with these days.

Re: Kite Responds to the Minimap and Autocomplete Issues

#45

Post by danso »

fooey wrote: post #33
ivanbakel wrote: post #27

An apology that starts with 4 paragraphs of self-congratulation is not a real apology at all. Nothing of the article reads as being genuinely regretful of the strategies taken - it's just sorry for being caught. >Unfortunately, while trying to make programmers’ lives a little less complicated, we instead made them more complicated. Barely an admission of guilt, and even that can't get past without patting themselves…

Buying a popular addon with the express purpose of subverting its original intent for commercial monetization seems like it should be against Atom's TOS This instance is razor thin close to being malware What's to stop the next enterprising company from intentionally uploading all your code to their servers? Modifying your code? Or holding your code hostage? Atom has git support backed in, so how likely is it an addo…

Maybe Github should have an informal policy that punishes bad actors by penalizing how they're listed in Atom's built-in package browser. Or create a convention of documentation that is machine-readable and easy to preview in the editor's package-updater tool. But making blanket rules regarding commercialization would likely have unwanted unintended effects.

FWIW, Atom's community manager wrote a long comment in one of the Kite plugins' Github issues:

https://github.com/atom-minimap/minimap/issues/588#issuecomm...

> As a steward of the Atom editor and the Atom package ecosystem, I'm of course concerned about packages being used for nefarious purposes. But I'm also loathe to impose blanket rules on what is or isn't valid functionality for packages without some very well-thought-out rationalization and justification for those rules. I've been watching this topic hoping that the smart people here would give me some ideas

> [an analogy to the value of advertising in other contexts]

> ...I'm definitely keeping my eye on this kind of thing and if advertisements become a problem, I'm sure that the Atom team will consider specific rules. I just don't think we're there yet.

Re: Kite Responds to the Minimap and Autocomplete Issues

#46

Post by Rantenki »

orclev wrote: post #40
zokier wrote: post #38

Earlier quoted context omitted.

> Or some mention of the upload at all The image they show in the post shows that they were fairly clearly saying "Where enabled, your code is sent to our cloud": https://kite.com/static/media/autocomplete-python-flow.690d3... Sure, it is not exactly hilighted there, but I wouldn't say that they were trying to hide that either.

As a third party to this debacle, all I have to go on is the comments people have made, but a few of them have asserted that that language was added later on and some initial versions of this change either outright lacked that warning, or else used styling tricks to make it difficult to notice. I lack the time and/or motivation to go digging through the commit history for that plugin, but maybe someone more enterpris…

It's been in there for quite some time, and there have been a variety of commits that carried some variation of the same "we analyze your code in the cloud" message:

https://github.com/kiteco/kite-installer/commit/c52c5daa672c...

Taking over the minmap autocompletion projects and adding these features still feels a bit sleazy, but the installer was honest about how it worked.

Re: Kite Responds to the Minimap and Autocomplete Issues

#47

Post by codezero »

fooey wrote: post #33
ivanbakel wrote: post #27

An apology that starts with 4 paragraphs of self-congratulation is not a real apology at all. Nothing of the article reads as being genuinely regretful of the strategies taken - it's just sorry for being caught. >Unfortunately, while trying to make programmers’ lives a little less complicated, we instead made them more complicated. Barely an admission of guilt, and even that can't get past without patting themselves…

Buying a popular addon with the express purpose of subverting its original intent for commercial monetization seems like it should be against Atom's TOS This instance is razor thin close to being malware What's to stop the next enterprising company from intentionally uploading all your code to their servers? Modifying your code? Or holding your code hostage? Atom has git support backed in, so how likely is it an addo…

Picking nits I think you mean adware not malware. I think it's worth making the distinction.

Re: Kite Responds to the Minimap and Autocomplete Issues

#49

Post by rhizome »

holtalanm wrote: post #23
free_everybody wrote: post #19

Earlier quoted context omitted.

If you are going to counter this aggressively, perhaps consider the emotional response you are masking as logical?

"emotional response"? really? I'm not masking anything. If you would like to read between the lines, then by all means, feel free. I'm looking at this whole situation without passing any judgement. Looks to me like we have a classic case of a company completely failing to execute properly on an idea. Did they fail? Yes, spectacularly. Was there malicious intent, or some form of intentional shady practice going on? I…

You are absolutely passing judgement when you equate the blowback against Kite with extrajudicial (and mostly racist) murder. That you doubled down in your flagged comment instead of acknowledging the (being charitable) hyperbole raises questions about your judgement. No offense.

That said, by deed and by words (per their apology) they executed on an "ends justify the means" business model. They wanted more business, they came up with the idea of buying open source projects, took a stab at integrating their functionality with the projects, and people didn't like it.

Don't tell me the people don't have a right to choose how they respond to Kite, business reputation is an actual thing and you can't browbeat people into ignoring their feelings on the issue.

Regardless, whether their strategy was malicious and/or shady is a value judgement best left to the individual, and there appear to be a variety of opinions on the matter. Note that @abe33 does not list Kite as an employer, so it would appear he's hedging his bet on how awesome this idea is.

Re: Kite Responds to the Minimap and Autocomplete Issues

#50

Post by emsy »

codezero wrote: post #47
fooey wrote: post #33

Earlier quoted context omitted.

Buying a popular addon with the express purpose of subverting its original intent for commercial monetization seems like it should be against Atom's TOS This instance is razor thin close to being malware What's to stop the next enterprising company from intentionally uploading all your code to their servers? Modifying your code? Or holding your code hostage? Atom has git support backed in, so how likely is it an addo…

Picking nits I think you mean adware not malware. I think it's worth making the distinction.

Giving ad networks tracking information of your editor usage and possibly access to the code you're working (which they didn't do, but the original statement was 'close to being malware') on is beyond mere adware.
Post reply on HN