Update on "Co-authored-by: Copilot" in commit messages
github.com2 days ago:
> We did catch it internally in testing [1]
Today:
> bug in the code that was not found in testing.
The bug is not about code behavior, but rather about getting noticed by users :)
"We're sorry we got caught"
I work at MSFT. I can understand the incentives behind this change. Although I am not sure how different GitHub culture is from MSFT.
I am sure they are closely tracking this metric of Copilot authored PRs so that everyone down from Nadella to the dev and PM for this can use it to hype up GH Copilot. It’s also a simple and clean metric that goes well in your Connects (performance discussion), you could say the feature I worked on led to xx million copilot authored PRs and there is now an AI usage mandate and you need to mention how you used AI to do something more efficiently blah blah. It’s good old promotion theatre. I don’t think its unique to MSFT though and is probably common across Big Tech.
> I am sure they are closely tracking this metric of Copilot authored PRs
Yeah that number goes WAY UP when you assume all commits are co-authored by Copilot!
This "bug" is an attempt to muddy the waters of what is and is not copyrightable and/or usable for training data. Full stop.
You may be an honest person, but your employers are not.
There are alternative ways to gather telemetry data about your usage, then literally polluting the commit message / PR description of the author. Why even consider doing that in the first place?
All the people there asking the simple question of why it got changed and getting ignored.
There ought to be decent number of people within Microsoft who have "Copilot usage" as a KPI. I don't think this was gamesmanship on their part (no sarcasm, I truly do not suspect malice), but I'm sure if it could have slipped in without backlash, they would have enjoyed seeing their line go up.
Not suspecting malice from Microslop is an untenable and unsafe position.
Sorry, I don't see another plausible motive than KPIs must go up. The change came from a product manager and was also reviewed and approved by one or two apparently senior developers. They may have tried to slip it in accidentally on purpose.
So freaking weird. Is it normal at MS for Product Managers to push code? wtf
Original PR: https://github.com/microsoft/vscode/pull/310226
Isn't this the promise of AI?
Agreed they could be clearer on this
IMO (and I am biased because I have written about this before in https://news.ycombinator.com/item?id=47164481) but I believe it's to make sure they're legally covering their users, and making sure users of AI tools do at least have some attribution for AI-derived contributions
We all know the answer anyway.
I am using a different approach.
`user.email` is always my email.
`user.name` is either my account name, or model name like `gpt-5.5-high`.
I can easily filter & blame which line was written by me or some specific AI
That doesn’t quite work for cases where you’re either the primary author of a commit (asking the model for some touch ups) or when you heavily edit model output. Easier to just say “this is who’s driving the AI” and keep it to your username.
> doesn’t quite work for cases
In that case, the "Co-authored-by: Copilot" method doesn't quite work either. You have to split the commit somehow.
> this is who’s driving the AI
As indicated by the consistent user.email value.
Hm, why? Just attribute it to yourself and be done with it. Is there a use case for classifying commits as AI assisted? Besides corporate bureaucracy.
Nice! I've been doing similar, but prefer using Co-authored-by, especially as it allows for cases where there may be multiple models in play
I'm not sure that anyone wants the scarlet letter of an AI coauthor on their code just because they used something simple like next edit suggestions or AI autocomplete. It seems like the "all" setting basically only exists for people that haven't figured out how to change it to something else yet.
(Funnily enough, I always commit through the command line in VS code anyways...not sure why. But I guess I would have avoided this annoyance, so that's a plus!)
Yeah. I wasn’t angry about this a couple days ago, but I am now.
So the thing that’s on by default and makes autocomplete worse (plain intelligence never changed my s.x = 0 to s.xVInputRadiusDetectionThreshold = 0 if I happened to take my eyes off the screen for a moment) is now stealing credit for my work?
I’m speechless.
Also glad I use a standalone git client.
Agree. I'm more reliant on having a keyboard than on having copilot make tab suggestions and I wouldn't like my PRs to include a tag: "Keystrokes courtesy of: Keychron K3 Max".
Definitely. Can you imagine the kind of world we'd live in if we had to sign each message with each product we used?
This message brought to you by Xfinity Internet.
i think my GPU should get all the credit "Calculated by AMD 395+ MAX"
Are they apologizing? Was it a bug? Why did they make this decision and what's the end goal? It's so unclear from the message - as evidenced by a lot of the responses.
Seems pretty clear, Claude and Codex were getting a lot of free publicity by instructing their models to do the same and MS wanted similar results. However, a bug caused this to be applied to all commits instead of all Copilot-influenced commits.
They did say it's a bug that they even caught during testing yet somehow let go through. Author of the issue mentioned this on HN https://news.ycombinator.com/item?id=47994193
"AI attribution by default" was an intentional feature, but there were also bugs. This is why the post is so unclear.
Probably because it's a PM who coded that bug, so no contest.
Who vibe coded it because who needs coders nowadays anyway.
I don't get why people are upset here. Vibed code is easy to spot even if you don't credit the LLM in Git.
# increment the current number of users — do it by one
n_users += 1> There was a bug in the code that was not found in testing that attributed non-Copilot code completions to Copilot.
The bug is not about code behavior, but rather about getting noticed by users :)
Inserting authorship claims is incredibly tacky. It’s today’s “Intel Inside” sticker. I don’t want your stickers on the computer I bought.
“Sent from my iPhone” isn’t an authorship claim.
Sent from my iPhone is worse than intel inside or claude in the commits in my opinion.
There is something so gross about injecting an advertising message into every single communication a user has on their device.
I recall there was some understanding that it had a legitimate use as well as the obvious marketing, which was to advise the reader that the message may be unexpectedly concise or contain errors because it was sent from a cell phone, something less common before the iPhone came out. BlackBerry phones did this too for the same reasons.
“Sent from my mobile phone” - no need to inject a product name
You misunderstand the purpose of "Sent from my iPhone" - it was a status symbol, it showed that the sender was part of the superior iPhone owning elite. It was trivial to remove, but most didnt "oh, I am too busy too remove it, I guess I'll just leave it and let everybody know I can afford an iPhone".
You are right, it was advertising, but it advertized the user, not Apple.
That's advertising with extra steps. Apple having created an ingroup and an outgroup is very effective advertising on their side.
I always thought this was an implicit request to forgive obvious typos and autocorrect mistakes. Sent from a mobile device (iPhone, Samsung Galaxy, Blackberry, Windows Phone, etc.) with a tiny keyboard and in a setting in which proofreading may not be as rigorous as normal.
At least “Sent from my iPhone” was a factual claim, unlike this mess.
“Sent from my iPhone” is a default signature, but you can change the message under Settings -> Apps -> Mail -> Signature (at the bottom of the options page)
These are the same thing. Marketing, and the ability to track reach. There's no other reason to do this.
On the flip side there are people who believe that LLM-assisted coding changes require attribution in git history.
As I've written elsewhere in the thread, having worked at a large Enterprise in collaboration with Legal, if there isn't tracking of what AI contributions you have, it's harder to be protected legally by ie Microsoft's indemnity clause if you're sued
It's definitely helpful to know whether a PR was AI-assisted or not and the git attribution line is a simple and effective way of communicating that.
I also recommend specifying model name and version so the maintainer knows upfront the level of slop they are dealing with.
What’s the problem with intel inside? That’s perfectly normal.
What’s the problem with intel inside? That’s perfectly normal.
I don't want my computer to look like it's racing in NASCAR.
If your NASCAR car had a sticker, you’d have ad money from STP/Pennzoil/RainX/whatever.
I would take a sticker for a sponsorship. That could be a good deal. Not for free though!
It's only one sliver of the problem here, but -- do you know how often I update my code editor? Like once every five or ten years, to the version that was released a year or two ago.
I do my own commits by hand so it's moot anyway, but there's a fair bit of "leopards ate my face" going in the GitHub thread.
VSCode updates itself what feels like daily so everyone is on the bleeding edge. There are upsides and downsides to that but it doesn’t feel like a trade-off many have made purposefully.
You can disable auto-updates for VS Code, and you can install older versions of it.
VS Code is updated monthly. More and more they also release a bugfix to the monthly release, a week or two after.
They switched to a weekly release cycle, presumably to compete with the perceived iteration speed of the many VS Code forks.
Default to ON is a complete dik move
It's not even default to ON, it's default to ALL (or at least to a lot), even non co-pilot commits, that's what made people made. If it was at least correct maybe it would have gone unnoticed.
Honestly extremely pathetic by a trillion dollar corporation that has a massive, undemocratic, say in how technology is developed in this country.
Microsoft should be broken up into a dozen different companies and it's quite clear they violated their consent decree from the US DOJ a few decades later, so they should get punished extra hard. Maybe nationalize Excel putting it in the public domain for starters.
Yeah break up all the big companies so Chinese state sponsored behemoths can take over everything. This isn’t the 90s where Americans only competed with other Americans.
China is competing so well because it has a central bureaucracy that issues 5 year plans and issues money to get them done. Do you think America should do that too, or do you think America and China are different countries with different values?
Let’s not forget how much US tech came from government programs like ARPA/DARPA. It isn’t exactly a decentralized bureaucracy. Just hasn’t been hitting as well recently.
Breaking up big tech would make US markets more competitive, not less.
GP didn’t say all the big tech companies. Just Microslop.
So you're saying the market is weaker with more competition?
Nope, just break up the one that has been consistently found to be abusing their market position. Microsoft has been embroiled in this since the 90s.
Honestly not sure I find that prospect worse than the American status quo. At least the Chinese regime is a rational actor.
"America is just as bad as China" is not cute or clever; it's trite and objectively wrong. There really is no intellectually honest argument to the contrary. For starters we don't get arrested for saying "Kent State Massacre" - can't say as much for "Tiananmen Square" in China. No matter how atrocious our government may be at times, it doesn't hold a candle to them.