The “do it first, apologize later” will be the general principle with anything. It’s going to be hard and futile to prove even if they don’t do it through ToS first. Amazon has one of the largest corporate training sets out there:)

and it is not end-to-end encrypted if you don't own the keys, avoid bullshit

Hey Martin, can you please work with Product to significantly clarify what is meant by the following language in the settings? Because right now it's nearly impossible for a layperson (or even an average programmer) to understand what this means:

""" Allow GitHub to use my data for AI model training

Allow GitHub to collect and use my Inputs, Outputs, and associated context to train and improve AI models. Read more in the Privacy Statement. """

If the reality is less scary than how it sounds, then the wording needs to be less scary-sounding. It may be that GitHub isn't training models on private repos, but the language certainly suggests that it is. The feedback we're seeing in this post is proof enough of that.

Finally, I read the Privacy Statement, and it's unclear what the applicable language is. "Inputs," "Outputs," and "Associated Context" are terms of art that have no matching definitions in the Statement. (The terms "Outputs" and "Associated Context" don't even appear in the Statement at all. Not even "train.") As an attorney I find this completely baffling.

If Copilot later adds a feature like "Scan your repo for vulnerabilities using Copilot <opt-out>", then that would both fit your criteria, and the baiting outrage of the original poster, in one swoop! Of course, Microsoft would _never_ do that, right?

> If you don’t use Copilot this will not affect you.

How does this work for a private repository with access granted to additional contributors? Which setting is consulted then?

Back in my day someone would post a HN article to the internal slack in order to sway conversation in their favor. Glad to see its still happening! :D

Sorry doesn't help at all but you can still be useful - can you please tell us how many private repos do "users of Free, Pro and Pro+ Copilot" who have used Copilot in the last 90 days exist in the github database?

Because microsuck is about to violate the law that many times

Right, but it shouldn't be opt-out only to begin with. It's a dishonest pattern that relies on people not noticing. Honest use of data is a "Caesar's wife must be above suspicion" moment for me -- if this is how you're acting when engaging with customers explicitly, I don't trust you to resist the temptation to tap into my data privately. AI companies already have trained their models illegally against the intellectual property of all of humanity with little consent along the way.

Honestly, if you work at GitHub, maybe you should focus on your uptime -- it's awful.

Appreciate the clarification. But, it's still not great.

To the PM behind this - developers are sensitive to this kind of thing. Just make it opt-in instead?

Say someone has a very sensitive secret (say, a Bitcoin private key) in their free private Github repo, and uses Copilot on that repo and touches the secret with it. Would you be willing to assure here that toggling that setting would not affect the likelihood of that secret leaking, and that that likelihood is also unaffected by whether the account is Business or Free?

How do you handle accounts that have copilot managed by an organisation? I've seen several cases where people cannot opt out their account because of the org connection (the option just isn't there in the settings). What happens to their account the moment they leave that org?

I think the problem is more with using PRIVATE repos. My letters are also private and I would be pretty pissed if the mail carrier was reading them. Why does GitHub think it has the right to do this?

Thanks for the clarification. The OP here made me think I missed something in both the blog post about the change and in the available settings.

This is a distinction without a difference, according to the text of that enable/disable dialog,

> Allow GitHub to use my data for AI model training: Allow GitHub to collect and use my Inputs, Outputs, and associated context to train and improve AI models. Read more in the Privacy Statement.

“Associated Context” is the repo. If I use copilot, I’m giving it access to my repo.

I don’t know in all the ways copilot can be triggered, and I’m not certain that I could stop it from being triggered, given Microsoft’s past behaviors in slapping Copilot on everything that exists.

Under GDPR, opt-out is not considered informed consent, and repositories can contain personally identifiable information, which fall under GDPR. Do you think differently, or do you think ignoring the law will be worth it?

Nice try. If you're training on "inputs" to Copilot then you are training on the private repos.

This suspect denial is why I will get my clients moved off of github.

Thanks for confirming you train on our data

Why not get user consent first?

Can't you just make it opt-in?

No? Because no one would opt-in, you say?

Wow. It's almost like this is a user-hostile feature that breaks the implicit promise behind a "private" repo.

I think you're well aware that people aren't upset at the distinction between training on Copilot data versus training on private repo data (at rest). People are upset because GH is using an opt-out model. Your response is disingenuous not to address this, and the "hope this helps" comes across as condescending (not sure if that was your intention.)

Defaulting to opt-in is a malicious move, no matter how you present things.

>Hope that helps

Honestly, what the fuck? This changes was already pretty bad but this being the apparent corporate response is insane.

Done with Github and Microsoft after this. Just disgusting how little you care for users, ethics, or morals.

As others have pointed out, this is somewhat dishonest. Which is depressing, if you represent GitHub.

Yes you do? If a user uses any form of copilot in one of his repos except ofc enterprise, says so right in the blog post. These aktshually corporate technicality defense posts aren’t helping, they just end up making you personally look a bit fishy.

What a wildly disingenuous take. Speaking earnestly from one human to another: your behavior and work is shameful, and you should feel embarrassed by your actions, Martin.

You’re laundering the code of users who don’t opt-in through Copilot users who do, to read in as many LoC as possible. It’s clear as day to everyone not morally bankrupt.

I am aware of CUI data hosted on github by corporate entities. You’re saying you’ll essentially violate the entire point of CUI?

That’s fucking terrifying.

"hope that helps"

Why the smug sarcastic attitude? nah, fuck github i'm out.

tl;dr: installed gitlab.

I'm not bidding against you to not train on my data.

“Opt-out” is an egregiously toxic and unethical approach to consent and should be illegal everywhere that it isn’t already.

I didn’t think Github had much of a brand left to damage, but here we are.

I think this kind of nuance is useless or even harmful. That might be how it is now but they'll change it when you're not looking.

You see coders have this reasoning flaw where they go "Oh I've understood the system, now I can work out all the ramifications of my actions", and then they get tricked at every step of the life.

It should take 0 seconds, because I shouldn't have to do it.

That's my bar. My time is my time, and anything that takes time from me better have a damn good excuse. Github is not bringing any good reasons to the table to justify making me take my time to protect privacy I've had by default up to now.

  > takes 30 seconds.

No, it takes an hour of perusing HN every day to stumble upon this. That's 20 hours per month, 240 hours per year, shall I bill it to GitHub or to Microsoft directly?

Corrupting Steinmetz' quip to Ford: it's 30 seconds to flip the switch, 240 hours to know that a switch needs to be flipped.

I will join the club. +1 for ruining M$ AI with my garbage code

And even if you read the banner on the site, the email they sent, and the announcement itself, you would not see instructions that mention the specific thing(s) you must change in order to opt out.

Sure, you can poke around in the settings and find one that you believe opts you out, but in lieu of clear and explicit instructions from GitHub, you'll have no way to find out. Only the possibility of finding out later that you guessed wrong.

I've recently started hosting my own forgejo instance. It works so well! Free tailscale for connectivity. I expose mine over fly.io proxy, also free, but not to be done without caution.

It's a fair question, but if you need private repos, I think you need to start considering a paid option, or self-host.

If it's really important to you that the repo is private, I'd self-host.

GitLab would be a good bet here. We started on their free tier and used that for a couple of years, I was very happy with it. Not sure how the tiers might have evolved since.

And according to their PM and privacy policy, they're not training their models on your code[0].

[0]: https://forum.gitlab.com/t/can-i-opt-out-from-my-code-being-...

It doesn't take much power or time to run your own local git server. My first one which lasted years was parts I mangled together from old computers from garage sales.

There's instructions on running a Git server in the git book: https://git-scm.com/book/en/v2/Git-on-the-Server-The-Protoco...

I've been using gitosis to manage private repos for almost 2 decades now. It's extremely easy to host your own repositories.

I just looked up gitosis on github though and it was last updated 12 years ago.... still works for me though.

Overall, hosting your own repos is very easy.

Sourcehut comes to my mind: https://sourcehut.org/

I would've recommended codeberg but codeberg isn't the finest to be recommended for free private repos.

I definitely feel like more can be done within this space and that there is space for more competitors (even forgejo instances for that matter)

BitBucket.org (Atlassian)

I use Fossil for mine. Dead easy to set up, and while the workflow might not be great for public contributions like Github is, that doesn't matter on something where I'm the only user.

Any computer you have ssh access to.

I've seen https://codefloe.com mentioned, can't say I've used it myself yet though.

A cheap mini-pc in your closet.

By migrating to another code forge and paying them so they're sustainable.

Which doesn't answer your question at all, but it is the metric they'll pay attention to. And it is the the thing that actually addresses the underlying problem.

The same reason b/c FreeBSD is great, but eventually it's transitioned to Linux at scale: commodity personnel.

You wouldn't believe the amount of people that would list Github, but not git, as a skill.

Sorry, which ones support 2-GB private repositories and are supported by package managers?

AI is maximizing the move fast and break things approach, including not asking for permission from its userbase.

It's consistent with believing that AI is the future -- if a company doesn't perform really well, it loses that race. And if the userbase they piss off is also the userbase that's skeptical about AI, then they're not pissing off anyone that's relevant to the company winning.

Downside: Pissing off users is gross.

The problem is, GitHub is owned by Microsoft, and Microsoft is desperately trying to shove AI into everything in hopes that it will save them.

Probably did. Now comes the legal ass-covering.

https://github.com/settings/copilot/features, it's near the bottom "Allow GitHub to use my data for AI model training"

I believe it is under:

Settings->Copilot->Features->Privacy=>[ Allow GitHub to use my data for AI model training

Allow GitHub to collect and use my Inputs, Outputs, and associated context to train and improve AI models. Read more in the Privacy Statement. ]

There is this distinct lack of giving a shit about the user that you see coming through in a lot of big tech nowadays.

Take this extremely simple example about antenna pod. I can change the order and what buttons show up in the app nav bar. For example I can remove the "home" button or put other things there instead like playback history.

This is a small minor point of the bigger picture. Yet there is this distinct sense in which when using that app I don't feel like I'm beholden to some chain of management in some company deciding they get to decide what I get to do.

Like its almost unthinkable that the YouTube app let you remove shorts or reorder the navigation bar and decide what you wanted to have there.

Can’t wait for copilot to start saying stuff like

// todo… remove this before it goes to prod lol

Probably extremely ineffective, it's an issue of scale and unless you really automate the terrible code generation and somehow manage to make it distinct enough in style that it isn't easy to detect and eliminate wholesale then you just won't have the volume to significantly impact the result set.

I'm absolutely sure that there are state actors with gigantic budgets that are putting a lot of effort into similar attacks, though.

Not private repos.

Now, anything that gets referenced in a copilot chat is fair game

yeah, how can I view the settings on my own personal account if my employer is managing the copilot settings?

Having hardcoded keys is a security nightmare regardless.

I saw that too, it feels like it's worded to make it sound like it's mandatory for Copilot. Based on their blog post the "feature" is them training on your data.

[dead]

This hints at something, that in my opinion isn't not discussed enough:

Say some personal data leaked into training data, where can I request surgical deletion of that data from the LLM? Not only license washing is done using LLM, but also PII washing and consent ignoring is done using LLMs. How will a service provider make sure to not ever have personal data in the training data set and fix earlier mistakes pertaining to personal data? Are they not obliged to have a way of deleting one's personal data? GDPR or something?

God, there's always that moment when you see the most shit code on earth, just as you're typing "git blame" and you just start chanting "please don't be me".