Our CI for our entire org at https://github.com/lightningdevkit was turned off for 3 weeks because an outside contributor who was wrongfully banned made a PR. After multiple appeals we received no explanation and was told it was a permanent ban until we made a stir on twitter. They sadly are no longer a good place to work.
Do they? Or is it that a new account is opened every second? Because I’ve been seeing so many spammers and scammers that those numbers have to be skewed.
Did we all forget that GitHub’s military-industrial complex owners over at Microsoft made sure to send the “business as usual” signal to the USG when they refused to stop helping ICE violate human rights en masse?
This was during the kidnap-and-rape-kids-in-cages days and before they started a general policy of kidnapping and/or summarily executing law-abiding citizens in the street. There are more reasons now to disassociate with collaborators with the US federal government than ever. I guess I could say I dropped GitHub before it was cool?
Microsoft is a morally bankrupt and despicable organization, just like Meta, Amazon, and modern Google and Apple. Anyone still doing ongoing business with them in 2026 is, imho, a fool.
[Slow clap... building to thunderous applause. Standing ovation.]
Bravo! Well done! Encore! Encore!
Now do a foaming at the mouth diatribe about how unethical libertarian crypto scamming shills like you and Trump are ruining the economy and violently widening the gap between the oligarchy and poverty! Extra points for plugging your latest crypto scam as the solution.
I've ditched Github for all personal stuff. I just keep my repositories offline. I have a reliable backup process so what's the point in pushing it there? I don't give a shit about public profile, stars or any of that gamified crap and I certainly don't trust them.
Lately, 9 out of 10 times i click on a github link to access the repo for $random_project_hosted_on_github I get a page saying that I've been rate limited. Even if it's my first click on a github link for the week.
I guess THAT page has the three nines availability.
Why don't open source alternatives just copy the UI to make it easier to switch? Everyone knows the GitHub UI and it's intuitive. I'm happy to get more privacy and freedom, you don't have to make a worse design just to be different.
Fluxer figured this out and they're the best discord replacement imo.
I think they have the same interface. Pull requests are renamed to merge requests, that's all the difference I see. Wait for github to reshuffle the ui in a redesign churn.
Until you have to work with stale GHAS tool configurations, remember whether a project uses rulesets or branch settings or find that comment you wrote on a PR (and then learn that the new PR "experience" fucking hides them above a certain threshold). Those are just the issues I encounter in a typical week.
Gitea did a lot of work over the last view releases (since 1.21 onwards) and are really GitHub-like nowadays UI-wise. Plus it is no SPA anymore and mostly SSR with Go templates + Htmx, its site performance lets GitHub cry by the wayside.
Best descision ever to leave GitHub and selfhost Gitea with some runners in our own datacenter.
Forgejo is pretty close. Its runners are largely compatible
with GH’s, and its issues, labels, tags, releases, wikis, packages, branch protection, secrets/envs, signing keys, repo permissions, etc. are all largely identical. I maintain a lot of mirrors and while none of my repos are particularly complex, I haven’t ever lost anything in a migration from GitHub to Forgejo.
I'm still paying $4 a month for GitHub since I keep some repos private, which was down from the $8 a month or whatever before Microsoft acquired GitHub, are you saying there's a price hike for AI services or something else?
You do if you want wikis and protected branches, the other features I'm not as invested in, and protected branches is just a common sense nice to have.
AI yearly plans got particulalry hasrsh treatment. per official GitHub Blogs and policies, they penalized specifically yearly plans with 50x ~ 100x multipliers on 2026-06-01.
and yearly plans were given options by GitHub: A) leave and get pro-rata refund; 2) switch to new plan.
so what is the point of yearly plans when GitHub can unilatrerally change conditions of your plan? or force you to leave? what is the point of agreement if I can change it later or say "no agreement".
Not good but that’s unsurprising since Thread’s value proposition is indistinguishable from twitter’s. Mastadon and bluesky seem to have healthy userbases though
Now if only the leader of Github would make Nazi salutes in public, regularly piss his pants due to frequent ketamine abuse, and cancel foreign aid causing 14 million brown children and other undesirable riff-raff to die by 2030, then maybe people would be as compelled to cancel Github as Twitter.
"Why X are doing Y" articles like these pretend that the premise of "X are doing Y" is true, but at the convenience skipping to the "Why" before even proving the premise is even accurate in any meaningful way.
This is why I never buy headlines that start out with "Why".
> developers are ditching
Proceeds to list but a handful of remotely meaningful repos against the hundreds of thousands on there
Rolling your car off a cliff isn't the same thing. This would be more like you want to drive your Tesla into the jungle so you build a road as you go.
The key problem is not losing the cars but losing the road builders who are now no longer building roads that lead to you, but rather roads that lead away
An early adopter is someone who is first in line to try something new, usually because they're willing to part of the new thing with their own hands. Just look at what projects are on any of the third-party forges, those are the early adopters of post-github tech.
I would also be one of them, but I'm not actually off Github yet. That's because I haven't quiite finished building the thing I'm going to move to.
I’ve seen titles like “Why top scientists are leaving the United States” where the article itself talked about A SINGLE RESEARCHER relocating to France.
If you're talking about the article featured on HN just day(s) ago, that was about a funding effort to get more researchers to move to France from the US, while they interviewed one specific individual. I think maybe you skipped the contents of that article (as it was in French) and instead just read the HN comments which misunderstood the article :)
Are you perhaps talking about "Top researchers leave USA for the Netherlands" then that was also on the HN frontpage just days ago?
There been so many articles, surveys and papers about how scientists and researchers are leaving the US for the last 1-2 years, that it's hard to keep track. Still, I'm curious to understand what you actually read.
No, I’m talking about an article I saw maybe 10-12 months ago on a major news site, perhaps The Guardian.
Individual reports of scientists leaving are meaningless without overall statistics. If there is a net outflow of scientists from the US, I’m not aware of it, and I certainly haven’t seen any actual statistical evidence for it.
> Proceeds to list but a handful of remotely meaningful repos against the hundreds of thousands on there
The trend is what's interesting here. Github has never been threatened by anyone, because their service was too good to bother for everyone but the most ideologically motivated.
Now their service has become so bad there's a github joke at work every time something is down or slower than it should.
Reputation is a very valuable thing, and Github has destroyed a stellar one in a few month, this is newsworthy.
I would expect that's true in any field that moves as fast as software development.
Developer A makes a move, perceives some benefit, tells developer B, who does the same thing and then tells developer C.
Some of the people consider the move, weigh costs, and make an intelligent decision. Many just think "smart people are doing this, I'll do it too". I really doubt this behavior is unique to developers.
Statistically I would assume that if you select a bunch of people on a certain criterium, you're going to see similarities related to that criterium among that group.
I don’t know if it’s because of the weasel words in the article or the extensive use of passive voice in the article, but this article feels in part like it’s AI generated.
I am exhausted with having to figure out whether someone wrote something or let AI generate it for them.
Reading articles like this has become less pleasurable since the advent of generative AI. There’s no feeling or heart in the article, and it’s one of those cases where I can read the headline, read the article, and wonder why I spent my time reading it.
Thank you! I immediately had the same reaction and was disappointed no one else pointed it out, though I'm guessing it's because most people just read the headline before commenting
> Proceeds to list but a handful of remotely meaningful repos
I'm pretty sure, shortly after the motorized vehicle was made commercialy available, there were only a "handful of remotely meaningful" people and companies who stopped using horses.
Do tell: How many horses are around on todays streets?
What exactly is the car equivalent in this analogy? Codeberg is not a car if you're saying GitHub is the horse. Just doesn't really make sense to compare the two like that...
A better alternative might be along the lines of “why developers ditch GitHub for Codeberg and self-hosting alternatives”. That way it doesn’t commit to a trend or exaggerate the situation in your mind, instead making it clear it’s a report on “those X who do Y do it for these reasons”.
Agree with the headline part, but the second part not so much.
As someone mentioned it's about the trend.
I have heard from people at multiple major open source projects that what is keeping them at Github at this point are free GH Action credits that they get and they couldn't really afford CI/CD if they left. Meaning numbers would be bigger if GH wasn't "paying" projects to stay.
I can't help but think at this point that Ghostty's "departure" from GitHub is unserious. It's been two and a half months now and not even a single peep of discussion about where they may be migrating to or what they may be doing.
Currently I self-host Gitea [0], use its registry for Docker, NPM etc and act runners [1] for github actions alternative, everything secured under tailnet.
I'm extremely satisfied with that setup. It is batteries included & fire and forget.
Now I use Github only as backup by mirroring my self hosted repos.
Similar with forgejo.
I mirrored all gh then flipped the ones I was using the most.
The biggest win was on running apple runners in my mac, so the free gh actions can do other stuff.
Getting better, more reliable and faster CI was such an underappreciated gain (from me at least) when moving to a self-hosted git platform. What used to take ~40 minutes end-to-end (from pushing commit to having release binary ready for three platforms) now takes less than 10 minutes, and seemingly that whole slowdown was causing me more headache that I think I was willing to admit at the time.
I similarly have been using Gitea for some years. I use it as my main forge and mirror to Github for discoverability and community reports and contributions.
For public projects I have workflows that can publish and push containers to both Gitea and Github.
I like GitHub and I'm not going to ditch anything, but I have to admit it's currently one of the few MS products that still holds up. Curious what it'll look like in a few years. Just in case, I've already reserved my username on Codeberg :)
Good idea to reserve a username early! I will do that too
Anyway, it's not that I don't think GitHub has real issues under Microsoft, but I disagree with the title of the article, as devs are not leaving it in droves, yet.
For private code, it just feels safer to self host that -- ideally behind wireguard for an extra layer of security.
For public code hosting, GitHub have banned too many people/projects for comfort. From security researchers to 18+ game devs, too many have been wrongfully banned.
It reminds me of the time where I deployed Gitea for self-hosting my git projects. In the end, nobody wanted to use it beyond myself. I would love to have a true federation protocol for Git, to decentralize the solution further.
I think there are a few options coming around federation of repos - like extension of forgejo, and then there's a newish player called Tangled i think, etc. I'm no expert, but that is something i would like as well...and i'm pretty sure that we are not alone in that desire.
Quite a few commentors here mentioned using gitea or one of it's forks on a private tailnet. That would mean it isn't publicly available and can't be scraped.
But on GitHub, you agree explicitly that MS can use it. I would assume codeberg has any-scraping measures (not saying they are perfect, but better than knowing Microsoft can train with private repos).
It sucks, but that ship has sailed. I'm not sure how they'll handle the issue of most new repos being AI generated, if they continue using new code for training. If the world accepts LLMs as a valid method for license washing, I don't see how I can fight it.
That is exactly what I would say. He got beat up, it is a fact and happened in the past. Unless you have a timemachine to undo it, nothing can make that undone. Now you can endulge in pain, anger, misery etc. - or you could get up, swallow the pain and look forward.
We've been self-hosting GitLab for about a year now, and I don't remember it ever going down or being unavailable. We self-host almost everything else too (except for online meetings), and it's all been pretty stable as well. Some of the tools we self-host do go down occasionally, but it's usually just a matter of restarting the VM or adding more storage.
Recently on Hacker News: https://news.ycombinator.com/item?id=48833116 (Chatto went open source) so you might be able to self-host your online meetings pretty easily too if you wanted to. (They're not the only option in that space, either, although they are one of the easiest to spin up an instance of that I've seen thus far.)
The appeal of GitHub for me is not only in the git hosting, but also in codespaces. It gives me:
1: An easy way to start a VM
2: A one-click solution to access it via private https access
So for development, I dont need to dabble with spawning my own Hetzner VM or something. And I also do not have to dabble with getting a temporary domain and DNS so I can set up my own letsencrypt certs and point the domain to that VM.
I can just write an index.html, execute "sudo python -m http.server 80", click the link that then opens to something.app.github.dev and test my new web application.
This is why codespaces make starting a new product idea a thing of like 1 minute instead of 1 hour for me.
216 comments
[ 3.6 ms ] story [ 57.1 ms ] threadDo they? Or is it that a new account is opened every second? Because I’ve been seeing so many spammers and scammers that those numbers have to be skewed.
This was during the kidnap-and-rape-kids-in-cages days and before they started a general policy of kidnapping and/or summarily executing law-abiding citizens in the street. There are more reasons now to disassociate with collaborators with the US federal government than ever. I guess I could say I dropped GitHub before it was cool?
https://github.blog/news-insights/company-news/github-and-us...
https://github.com/sneak
Microsoft is a morally bankrupt and despicable organization, just like Meta, Amazon, and modern Google and Apple. Anyone still doing ongoing business with them in 2026 is, imho, a fool.
So that would be almost everyone.
Bravo! Well done! Encore! Encore!
Now do a foaming at the mouth diatribe about how unethical libertarian crypto scamming shills like you and Trump are ruining the economy and violently widening the gap between the oligarchy and poverty! Extra points for plugging your latest crypto scam as the solution.
My email and phone number’s in my profile. Do please get in touch.
I guess THAT page has the three nines availability.
Fluxer figured this out and they're the best discord replacement imo.
https://fluxer.app/
Good luck. The amount of features and screens on GitHub are vast aside from just those code / issues / PRs tabs.
Until you have to work with stale GHAS tool configurations, remember whether a project uses rulesets or branch settings or find that comment you wrote on a PR (and then learn that the new PR "experience" fucking hides them above a certain threshold). Those are just the issues I encounter in a typical week.
+ not honouring yearly commitments plans
and yearly plans were given options by GitHub: A) leave and get pro-rata refund; 2) switch to new plan.
so what is the point of yearly plans when GitHub can unilatrerally change conditions of your plan? or force you to leave? what is the point of agreement if I can change it later or say "no agreement".
"Why X are doing Y" articles like these pretend that the premise of "X are doing Y" is true, but at the convenience skipping to the "Why" before even proving the premise is even accurate in any meaningful way.
This is why I never buy headlines that start out with "Why".
> developers are ditching
Proceeds to list but a handful of remotely meaningful repos against the hundreds of thousands on there
If there's a trend to leave a platform it won't start with the most entrenched users (largest repos).
And not provide any other meaningful data
The key problem is not losing the cars but losing the road builders who are now no longer building roads that lead to you, but rather roads that lead away
I would also be one of them, but I'm not actually off Github yet. That's because I haven't quiite finished building the thing I'm going to move to.
I’m not.
Are you perhaps talking about "Top researchers leave USA for the Netherlands" then that was also on the HN frontpage just days ago?
There been so many articles, surveys and papers about how scientists and researchers are leaving the US for the last 1-2 years, that it's hard to keep track. Still, I'm curious to understand what you actually read.
Individual reports of scientists leaving are meaningless without overall statistics. If there is a net outflow of scientists from the US, I’m not aware of it, and I certainly haven’t seen any actual statistical evidence for it.
The trend is what's interesting here. Github has never been threatened by anyone, because their service was too good to bother for everyone but the most ideologically motivated.
Now their service has become so bad there's a github joke at work every time something is down or slower than it should.
Reputation is a very valuable thing, and Github has destroyed a stellar one in a few month, this is newsworthy.
Developer A makes a move, perceives some benefit, tells developer B, who does the same thing and then tells developer C.
Some of the people consider the move, weigh costs, and make an intelligent decision. Many just think "smart people are doing this, I'll do it too". I really doubt this behavior is unique to developers.
Statistically I would assume that if you select a bunch of people on a certain criterium, you're going to see similarities related to that criterium among that group.
Hat tip to Microsoft.
I am exhausted with having to figure out whether someone wrote something or let AI generate it for them.
Reading articles like this has become less pleasurable since the advent of generative AI. There’s no feeling or heart in the article, and it’s one of those cases where I can read the headline, read the article, and wonder why I spent my time reading it.
I'm pretty sure, shortly after the motorized vehicle was made commercialy available, there were only a "handful of remotely meaningful" people and companies who stopped using horses.
Do tell: How many horses are around on todays streets?
As someone mentioned it's about the trend.
I have heard from people at multiple major open source projects that what is keeping them at Github at this point are free GH Action credits that they get and they couldn't really afford CI/CD if they left. Meaning numbers would be bigger if GH wasn't "paying" projects to stay.
https://www.paulgraham.com/submarine.html
Currently I self-host Gitea [0], use its registry for Docker, NPM etc and act runners [1] for github actions alternative, everything secured under tailnet.
I'm extremely satisfied with that setup. It is batteries included & fire and forget.
Now I use Github only as backup by mirroring my self hosted repos.
[0] https://gitea.com
[1] https://docs.gitea.com/usage/actions/act-runner
For the personal-opensource ones, I am on Github because this is where everyone is when I want to share/collaborate etc
For public projects I have workflows that can publish and push containers to both Gitea and Github.
Anyway, it's not that I don't think GitHub has real issues under Microsoft, but I disagree with the title of the article, as devs are not leaving it in droves, yet.
For public code hosting, GitHub have banned too many people/projects for comfort. From security researchers to 18+ game devs, too many have been wrongfully banned.
Loved Bitbucket's Mercurial offering. Looking for a replacement.
Heptapod is a GitLab fork that adds Mercurial support: https://heptapod.net/, free for open source projects: https://foss.heptapod.net/heptapod/foss.heptapod.net
Starve the greedy beast
Hey, this attitude sucks. Would you tell this to someone who got beat up in an alley?
I've been thinking about self-hosting myself, but for my purposes (open source), I'm worries about scrapers and other sources of DOS-like traffic.
I can just write an index.html, execute "sudo python -m http.server 80", click the link that then opens to something.app.github.dev and test my new web application.
This is why codespaces make starting a new product idea a thing of like 1 minute instead of 1 hour for me.