"Why X are doing Y" articles like these pretend that the premise of "X are doing Y" is true, conveniently skipping to the "Why" before proving that 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
> 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.
The existence of Telegram doesn't negate the fact that WhatsApp is the world's most popular instant messaging platform, and the others aren't even close.
And Telegram is a lot more developed and has a much larger percentage of the global instant messenger marketshare, compared to Github vs CodeBerg.
Oracle had record revenue in its most recent fiscal year, with record user engagement. So whatever connection you're trying to establish between its fate and that of SO or myspace is off-target, both in terms of popularity and revenue.
Stackoverflow was still arguably the best offering at what it provided, but what it provided became obsolete. The need for a repository service is only growing.
It is not clear to me that Github's service has degraded due to incompetence, it also seems possible that they are just struggling to meet demand as the source code backbone of an internet in a critical moment of evolution. I'm not sure any single provider would fare any better.
If there's a trend to leave a platform it won't start with the most entrenched users (largest repos).
They acknowledge your concern in the article and their analysis does apply to those few who are leaving. But to be fair the title can be interpreted either way and the most reasonable read for anyone is "some of them are leaving". I'd find it clickbaity if they said "why developers are leaving en-masse" and then point out to the regular turnover. There's clearly a trend, what's not clear is if it gains momentum.
That's the point being made. Is there a trend? How do we know?
There's always some repos moving between hosting providers for all kinds of reasons. The burden of proof is on the author here to show there's been an increase and they don't do that.
Which early adopters exactly? The most prominent example they gave is ghostty which has existed for just a couple of years, notwithstanding the fact that the owner published a spiel about how he's personally been using it for longer.
An early adopter is someone who is first in line to try something new, usually because they're willing to build 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.
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
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.
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 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
You could drop the verb clause? This would make the headline accurate while keeping it punchy.
> why Americans switch to soy
And
> why developers switch to codeberg
But the cynic in me thinks that the form of the headline that drastically overstates the the phenomenon in question by implication is something that has been workshopped and is commonly used because it turns something kind of boring into a spectacle.
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”.
> 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...
Agreed, if there was something actually better to switch to then I'd be interested. But seems unlikely for that to happen now - it's easier than ever to build a new github, at least the app itself but i doubt someone's gonna bother with the business effort needed to actually build it up as a reliable, trustworthy option that you know will be around for a while, which is a process that takes years when you know you're going to get disrupted. It would probably have to be open source to get early adopters to use it but somehow be nicer to use than GitHub, and there's basically no money to be made
When cars started replacing horses, they didn’t have more features. They just didn’t get sick or temperamental, and they didn’t shit all over the streets.
Going more than 20 miles an hour and not dying of heat exhaustion (or because your oats are too dry) are pretty big features.
But my point was that it’s a bad analogy. People are opting out of GitHub despite the alternatives having fewer features. You can read that as an ideological choice or as a YAGNI one. If it’s ideological, then there’s no competitive or feature angle at all; people are doing it because it seems right to them.
(To be clear, I have no problem with this. I think GitHub only gets better after public pressure, as we’ve seen with the last N cycles of product atrophy.)
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.
I tried codeberg, used it a year, then early this year in all their wisdom codeberg decided to show adverserial random text instead of my repo, reporteldly to mess up llm training to user agents they weren't sure were human.
Codeberg had one job, serve my repo, it didn't do that, when brought up, I was told it was a feature not a bug, they could maybe whitelist me but that wasn't my problem, it was that random people got totally blocked or from accessing the repo. I moved back to github.
Is your complaint about Anubis? I see the "checking if you are human" kind of text too when going to codeberg, but it is usually only a second or two. It is because I run a bit of obfuscation and resistance to browser finger printing.
But you find seeing the text for a couple of seconds too annoying to use codeberg? Maybe it is more than 2 seconds for you?
Everyone has their views on what is acceptable in the world and what they will put up with. Just, to my point of view, I think codeberg is trying to fight the good fight in keeping llms from crawling their website.
Codeberg has some extra "anti-scraping" measures than just Anubis which, judging from the community issues[0] sometimes ends up with false positives that cause people to get garbage (example issue[1], potential page with the garbage i found from Google[2]).
Interestingly, someone mentioned that you may get the garbage when searching Codeberg using Google's `site:codeberg.org`
If only we had research on the effect of a second or two's effect on user experience.
I wouldn't have a problem with it if it was on expensive endpoints like search or deep history dives, where it matters for server load. But it's every single page, out of some strange sense of righteousness.
No, the issue is when i visited i got a page of senseless drivel (maybe https://lib.rs/crates/iocaine) with the message i was a bot. There was no way around it except "contact support" as the message said the fix, to solve it myself, was to not be a bot.
I accessed through mobile, maybe got a bad ip, i don't don't know. But i never could access my repo, and incidently a day before when i tried to share some markdown file in the repo, they told me it wasn't there (i thought it was a them problem then, as i didn't see it).
So that experience and the way it was handled totally ruined my trust. If randomly people can't even access the repo, and that is "working as intended", what's the use to host it there.
They explicitly explained that the problem was not "they find seeing the text for a couple seconds too annoying to use codeberg". Codeberg offered to remove that for them so they'd never see it again. So idk why you're asking about that. Whatever the problem is, it must be something other than that, since that problem doesn't even exist.
The problem is blocking or speed-bumping users in general, everyone else besides themselves.
They didn't put things on a public hosting site for them to be hidden or obfuscated or even to have a tiny friction inserted between a user and that user discovering their stuff, or following a link to something.
Saying "2 seconds" as though that makes it insignificant is completely missing the point. 0.2 seconds, if it's 0.2 more than some other path is the same as a total block. Having a link to something do anything at all other than instantly provide that thing is outrageous and unacceptable, if you care about the experience you want to present to your users or audience.
(Still, I'd say codeberg was at least among the best options if you want someone else to host it for you. The bad job of providing a bad experience for legit users while trying to block AI scrapers can be true at the same time as no one else is doing any better without some other worse strings attached like github.)
You’re so right. I have a public-facing Forgejo server. Before configuring Anubis, scrapers were sending it about 600K requests per day. Copying and pasting from my blog post about it:
* For every Git commit, fetch the version of every file in the repository at that commit.
* See git blame for every file at every commit.
* Attempt to download the archive of each repo at every commit.
* Run every possible pull request search filter combination.
* Run every possible issue search filter combination.
* Fetch each of those URLs at random from some residential IP in Brazil that had not ever accessed my server before.
Afterward, it dropped to several hundred. Expect anti-attack features to keep getting stranger and more visible as scraper get still more aggressive.
This matches my experience of running a public mediawiki server. The bots (mostly Facebook/Meta) will for every single change fetch every single page again, over and over.
What do you use to monitor this? I don't really keep a close enough eye on my services to know what the traffic is doing (and haven't had any issues) but maybe I should start
I just tailed the web logs for a bit and saw that it was wild. For fun, I fed 10 minutes of logs into an AI and it picked up a lot of signal I didn't catch at first glance, like clients claiming to be MSIE 7 on Android 3 and such. I added some reject rules to the webserver in front of Forgejo but that only made a dent in the traffic, alas.
And the cruel irony is that these are FOSS Git repos I'm publicly sharing. I'd've been fine with them cloning the repo and analyzing away to their heart's content. That's not the way their scraper's wired, though.
In my experience, yes. At the peak of a scraper flood I was dealing with, I'd say about 90% of the traffic was from a unique IP. I'd never seen anything quite like that before.
Today it appears you have two options
1) Anubis or similar
2) Accept that 90-95% of your traffic is LLM scrapers fetching every file for every commit for every repo etc etc. There's some comments above with specific lists.
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.
Github is an extra layer of backup, among normal backups.
[edit]
Notable reasons:
- Github runners went oftenly out of space & they were slow. With self hosted runners I don't have these issues anymore because I control the hardware.
Previously I was paying Docker Build Cloud/Depot for performance + Github Pro for extra minutes. Now it's zero cost, superb performance and unlimited minutes.
- I have a centralized registry with private packages and images.
- It's secure, I don't worry if I accidentally make a repo public or leak secrets. I control the access to it in network level.
- I own everything, in case something goes nuts (eg lose access to GH) I'm safe.
If you're using github as a backup, then whatever you say you own, they also own. So I still don't understand why you are self-hosting and using github as an extra layer of backup. Seems philosophically inconsistent.
Not GP. Probably less dependencies on github, e.g. actions which sometimes don't work. This way github is a "dumb backup".
I selfhost forgejo (gitea fork) on home sever (nuc), similar setup with tailscale. I was planning to setup git mirror on a remote VM for backup, but since I am the only one using it and have everything on dev laptop and remote backups of nuc server I didn't bother to do that (I know I still should).
NUC 11 i5-1135G7 with 32gb of ram and one nvme. Bought it new years ago.
For forgejo I host only code and packages there, everything is built on dev pc since I am the only one using it. It also host postgresql server, jellyfin (for music only), custom OPDS server (for books only) SMB shared directory I can use from multiple PCs, some test environments for personal projects and possibly more things I forgot.
It's a small pc but it works without problems. I do have to clean it from time to time (with cheap battery-powered air duster) to keep the thermals low.
yeah i have 3 NUCs at home, albeit with less ram and smaller hdd than yours. im already using them for dokploy and other services so trying to selfhost forgeo or gitea might bring it to its max.
Yeah I should've asked why they are self-hosting in the first place. I was wrong in assuming that they wanted to host it themselves (that is, avoiding an external dependency).
The self hosting will still be there and working as expected no matter what GH does (fails... again, DMCAs the repo, bans the account, etc.). Self hosting isn't only about being the only one with the data, it's also for the independence aspect. GH as a backup doesn't hinder the independence. Network effects are strong and make a lot of developers still have a GH presence as a secondary platform.
The evolution is when one can finally fully disconnect from GH, the main self hosted platform will continue to operate as if nothing happened.
A migration can have a period of parallel running.
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.
As a developer who ditched Github and decided to self-host, there is only one reason. It's not technical difficulties, politics, nor AI. It's Microsoft. Like Apple, Facebook etc, I have a deep loathing for Microsoft and I want to remove as much of it from my life as I am able.
I now run Git on a pi using Gitea and Forgejo. I can now upload files of a size unheard of in GitHub, Claude can make a PR by itself that I can diff, edit, then merge, and even with the mighty power of a single pi 3b+, it feels more responsive.
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.
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's a losing game to try to keep any FOSS code completely isolated from large scraping data sets. All you can do is never share it, or share it and eventually it'll get consumed by the machine.
Spend your time making better software you and your friends can use... don't waste the time trying to stop others from making another painfully average codegen machine.
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.
Github for open source was always my favourite part of it.
Github for private repos has long had security issues, every time a serious issue announced it makes me wonder how long it's quietly existed and been exploited, and how many other holes are currently exploited that aren't well known.
Before github, people hosted their own repos all the time. Learning about alternatives, even if they aren't for you in all cases, is still worth it.
Has anyone had any issues with codeberg or other alternatives?
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.
Given more code hosting services, I wonder if we'll also see a corresponding increase in the number of alternative VCS or if git is legitimately very entrenched as a tool. I am just being a bit grouchy but I do wish there was more development of alternative VCSs. pijul at least looks cool even if I don't know if it scales well. Git LFS can be somewhat finicky to work with so maybe we'll see perforce like systems. It's obviously not the most practical thing to have a variety of very different VCS's and definitely a PITA to learn multiple tools but git does seem somewhat suboptimal given the number of anecdotes about people just re-cloning the repo. I was recently trying jj and it seemed to work well (excluding the lack of LFS support) so here's hoping.
I do think git serves a lot of people's needs very well. I see a lot of people being quite happy with direct github alternatives. Self hosting may become more popular too, I hadn't heard too many self hosting stories until recently. If LFS is your pain (I cannot stand to use git lfs) I have to mention oxen, which is open source and self hostable, but also has a hosted option. It handles large files right out of the gate and doing so faster than everyone else is one of its core missions. Still using github for distribution though, lol, see https://github.com/Oxen-AI/Oxen. I am also curious about where jj will go, its philosophy is really interesing, and sounds promising.
Genuinely curious here for someone who has tried self hosting git and has found it a pita to maintain...i want to know what is it that devs are flocking to other platforms and how are we sure that they won't pull all the red card signals that github is said to pull off.
Have been self-hosting GitLab for my org a few years by now, with quite a few users (>800 atm). Updates are automatic via the GitLab Omnibus package repos. Once or twice per year some update requires intervention. Otherwise, nothing bad happens. Very happy so far.
Biggest problem at the moment is that AI scrapers (curse them and their owners, pox be upon their houses!) sometimes bring things to a crawl. But nothing that a few firewall rules and anoubis won't solve.
If you are running an org you should be putting all your private services behind netbird or tailscale at minimum. Zero public infra exposure beyond them.
The service isn't supposed to be private, some repos are intentionally public. And including all external collaborators in some VPN scheme is not possible at that scale.
Forgejo is fairly simple to run – way lighter and simpler than GitLab even. (GitLab is quite okay, too!)
If you want a hosted service, go for Codeberg. It’s run by a German non-profit (so it’ll be hard to bite and switch OpenAI-style). Only free/open source projects are accepted, though.
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.)
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?
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".
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.
I'm trying sourcehut at the moment https://sourcehut.org/ and it seems really good - very simple and fast. And does seem to be free for hosting open source projects.
I really do not like that Drew (the owner and developer) is extremely dogmatic and political, and likes to get involved in what is and is jot allowed to be hosted on “his” forge.
Imo, service providers should be neutral and get involved only as far as required by law.
I would not trust Sourcehut. If Drew decides one day that he does not like you, your politics, or your industry, he will just cut you off. That is no foundation to build on.
I use it as my main forge and could not be happier. It also seems to be the only forge that allows its web frontend to work with JS disabled. (That I am aware of)
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.
I co-maintain a popular Linux launcher, Fuzzel, on Codeberg. We have no shortage of contributors despite not being on Github and the quality of contributors there is fairly high. In the past there were performance problems, but it's worked well enough lately.
My personal repos are on Gitlab. The UI is cluttered, but it works well enough.
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.
I've been on Codeberg for my projects since September but will pull them if this happens. I see no definition of what "mostly using AI" means and refuse to be "punished" for the choice of tooling I use. The tone and content of the thread on Mastodon is absolutely toxic.
I'd rather self-host than deal with the ambiguity and threats. Invariably from people who don't even make anything.
I have plenty of concerns around "AI" and project governance and IP law. My projects are copyleft, I'm in favour of free software, bleah bleah bleah.
But this just looks like pitchforks and mob behaviour.
Same here, i've been on Codeberg since 2022 but if this passes, i'll most likely move my stuff elsewhere (my projects do not even have any AI code, i just dislike the idea of telling people what tools to use to make their own stuff).
I dislike GitHub and i'm using a shared host for my website which wouldn't let me install something like Forgejo (this is by choice as i don't want to bother with sysadmin stuff, my site is almost all static HTML with only a couple PHP scripts for some minor tasks). So i guess i'll be migrating to GitLab, even though i do not really like its UX. Or maybe i'll use some other Forgejo instance like CodeFloe. URLs aside, migrating between Forgejo instances should be easy.
Yep. I have a self-hosted Gerrit instance that I was using for other more private stuff. I could in theory move over to there.
The downside is then having to worry about backups. But also lack of all the other "forge" things like issue tracking and so on. And lack of public exposure.
Frustrating, really. I felt welcome on Codeberg up until now. But that Mastodon thread is... wow.
If I understand it correctly, their target is projects like the recent pgrust (PostgreSQL in rust) which are full conversions of projects that no-one else will ever use. It's just noise.
7. You must not share projects that mostly consist of code written by "generative AI"-tools (including services such as *Claude*, *OpenAI Codex*).
Who defines "mostly"? And what's the definition of "written"? If I'm heavily prompting, curating, reviewing, and editing after the agent has gone through, is that somehow now "allowed?" Who makes the decision and how can I appeal?
Almost everything I write now has some agentic aspect to it, and I'm by far not the only one. That wasn't the case when I started my projects but it is now.
But I'm a software engineer with 25-30 years of experience and a computer nerd since 1979 when they wheeled the Apple II into my Grade 1 classroom. I'll reserve the right to decide what tooling I use to write my copylefted free software.
Given the Mastodon thread linked above I don't trust the people involved to be the judge jury and executioner. Insults flew freely and moral absolutism with it. I'll just pull my project off their infrastructure rather than be subject to their personal whims.
No, being "neutral" is not "taking a side", that's the definition of the word. If you pre-moderate you open a whole can of worms. What will they ban next? This makes me not want to use Codeberg ever because it's not plannable.
Calling all uses of such tooling "AI slop" right away shows your hand as only holding a blunt, crude, instrument.
Many of us are capable of more nuance that that and don't apply the label "slop" to everything where generative AI is involved. The linked Mastodon shows no nuance, and neither does your reply.
Two things can be true at once: GitHub has degraded into a "slop" cesspit and that agentic tooling has legitimate use cases.
Finally people can also legitimately object to a host mandating what tools can be used in authorship even if they don't particularly care for those tools themselves.
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.
Yes, but the thing is just that if people are looking around for new providers it's an opportunity for alternative systems to attract attention and users.
because github is subpar, you will try to convince people to migrate their repositories to $NOT_GIT so they can try a different hosting service ? doesn’t sound like a pragmatic choice to me.
But Github is already fully volatile, capricious, fickle, erratic, unpredictable, variable, inconsistent, changeable, unstable, whimsical, protean, fluid, and a polluting poisonous room temperature liquid heavy metal, so why would you also need mercurial?
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.
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.
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.
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.
I’m not disputing how intuitive the GitHub interface is, but seriously, why is it so hard for technical professionals to set aside 10–20 minutes of their time to learn a new interface? Why has this even become an issue worth discussing?
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.
I have been self hosting my repositories on my own VPS using nothing but Git itself (you don't need to install anything else). Sharing write access is a bit of a pain, but I tend to work on things alone anyway.
I hope some developers will see this message. I know that you work however you wish and you do what you wanna do, but there is a lot of people that will refuse to signup for GitHub (for many reasons), and in consequence, won't contribute to your project. I'm sorry if it sounded a bit off or I offended someone by stating this.
If the goal is lowering GitHub dependence without going full self-host, Buddy has built-in Git hosting alongside its CI/CD, so repos and pipelines live in the same place.
You'd be surprised how easy it is to self-host GitLab with Docker Compose, GitLab has an official "Omnibus" Docker image. No need to handicap yourself with Gitea/Forgejo/whatever, you can just use an industry-standard platform without much effort.
Hardware requirements are nowhere close to high either.
Spinning up a docker image with GitLab is easy indeed, but the installation is huge and you can feel it doing weird stuff (and it keeps doing "stuff" all the time even if nobody does anything - before you even bother to create extra users or repositories or whatever). The whole setup seems to be designed to be opaque.
Forgejo is a single self-contained ~110MB binary (it contains all assets) you can drop anywhere and it'll just work - depending on your needs you most likely wont even need to bother with a DB server as it can use SQLite (my local install is doing that).
Also FWIW personally i find Forgejo's UX much better compared to GitLab.
Aren't those about organizational processes instead of just specific software features? For example, I don't think self-hosting gitlab is enough to claim ISO/IEC 27001, just based on this snippet from wikipedia:
> ISO/IEC 27001 requires that management:
> Systematically examine the organization's information security risks, taking account of the threats, vulnerabilities, and impacts;
> Design and implement a coherent and comprehensive suite of information security controls and/or other forms of risk treatment (such as risk avoidance or risk transfer) to address those risks that are deemed unacceptable; and
> Adopt an overarching management process to ensure that the information security controls continue to meet the organization's information security needs on an ongoing basis.
I don’t know if Gitlab is an industry standard, but I’ve never heard of Forgejo. I worked for a headless CMS company and the only three providers we ever had requests for were GitHub, Bitbucket, and Gitlab. Gitlab is big enough to be generally adopted by governments. I think it’s fair to say it’s at least a lot closer to being an industry standard then Forgejo.
(Aside: I would likely never use Gitlab by choice, and would consider looking into Forgejo)
Probably because those are service providers and Forgejo is software. Same reason people ask for Microsoft account and Google account integration, but nobody ever asks for Linux account integration in your cloud software.
I can barely hold the attention of someone, developers included, when discussing my coding projects. I cannot imagine how difficult it would be if it wasn't even on Github. I'm a small developer, I don't really have the privilege to choose.
I left because Github became too LLM obsessed and I find that to not align with my personal ideals. Its not really deeper than that. The programming community will continue to splinter more into those that care about the craft and those that only care about outcomes.
For those who moved away from GitHub: do you miss the social/discovery side of it?
I mean things like comments on issues/PRs, stars, followers, finding existing repos, seeing which projects are popular, and getting drive-by contributions.
Or does that matter less than I imagine once you self-host and mirror public repos back to GitHub/Codeberg?
I used to love browsing GitHub Explore Trending but in the last two years it is almost exclusively AI apps that are slight improvement on the previous trending page of AI apps.
GitHub explore could be way more interesting with a simple filter algorithm.
I think and hope we see a lot more of this before the adversarial imperative returns from the company side.
People using Claude Fable to just make replacements for disgustingly enshittified software. We desperately need browser extensions to help make websites less scummy across the board as well.
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
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.
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.
Sentiment for/against GitHub aside...
"Why X are doing Y" articles like these pretend that the premise of "X are doing Y" is true, conveniently skipping to the "Why" before proving that 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
> 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.
Yeah, like how developers were en-masse ditching GitHub to go to GitLab when Microsoft acquired GitHub.
Maybe those people weren't wrong to do that...
I think the parent comment was sarcastic, no?
It's so weird the herd behavior of developers.
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.
Is it really?
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.
Except the article doesn't prove any trend
Here you have (albeit small) proof of some sort of trend: https://trends.google.com/explore?q=codeberg%2Cforgejo%2Cgit...
Still, doesn't come close to popularity of GitHub itself today (https://trends.google.com/explore?q=codeberg%2Cforgejo%2Cgit...), but I think the trend of moving away from GitHub is clear both in data and sentiment, both qualitative and quantitative.
If only the article used any of that. And if it did, I still don't think the headline was warranted.
Also Google search trends are no evidence of adoption or migration. High chance of correlation, sure.
The trend i see in Google is explosion in git(hub) related searches and small blip in non-github alternatives
The existence and growth of the codeberg project does, however.
The existence of Telegram doesn't negate the fact that WhatsApp is the world's most popular instant messaging platform, and the others aren't even close.
And Telegram is a lot more developed and has a much larger percentage of the global instant messenger marketshare, compared to Github vs CodeBerg.
And what level of 'growth' constitutes a trend that warrants "developers are ditching GitHub" without providing any numbers at all?
Stackoverflow usage didn't fall overnight either. But it has gone the way of MySpace and Oracle.
> the way of myspcace and Oracle
Oracle had record revenue in its most recent fiscal year, with record user engagement. So whatever connection you're trying to establish between its fate and that of SO or myspace is off-target, both in terms of popularity and revenue.
Stackoverflow was still arguably the best offering at what it provided, but what it provided became obsolete. The need for a repository service is only growing.
It is not clear to me that Github's service has degraded due to incompetence, it also seems possible that they are just struggling to meet demand as the source code backbone of an internet in a critical moment of evolution. I'm not sure any single provider would fare any better.
[dead]
Wow. GitHub gaining the same reputation as DNS ("it's always DNS"), printers, OSI layer 8, and PEBKAC is actually a bit of an achievement.
Hat tip to Microsoft.
> a handful of remotely meaningful repos
If there's a trend to leave a platform it won't start with the most entrenched users (largest repos).
They acknowledge your concern in the article and their analysis does apply to those few who are leaving. But to be fair the title can be interpreted either way and the most reasonable read for anyone is "some of them are leaving". I'd find it clickbaity if they said "why developers are leaving en-masse" and then point out to the regular turnover. There's clearly a trend, what's not clear is if it gains momentum.
> If there's a trend
That's the point being made. Is there a trend? How do we know?
There's always some repos moving between hosting providers for all kinds of reasons. The burden of proof is on the author here to show there's been an increase and they don't do that.
The early adopters are leaving. These are the people that will blaze a trail that others follow.
Which early adopters exactly? The most prominent example they gave is ghostty which has existed for just a couple of years, notwithstanding the fact that the owner published a spiel about how he's personally been using it for longer.
An early adopter is someone who is first in line to try something new, usually because they're willing to build 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 know what an early adopter is. Should we call this HNsplaining?
I'm saying that Ghostty, which is ~5 years old, cannot be called an early adopter of GitHub, which is 18 years old.
You could say that any time two distant psychos roll their Teslas off a rocky cliff and then exclaim "Why Tesla drivers are dumping their cars".
And not provide any other meaningful data
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
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.
GitHub is still the only option on their download page: https://ghostty.org/download
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.
The trend may be there but the article does nothing to back that up
howtogeek is low-effort content mill, it's just upvoted here because of the headline, there is 0 actual effort in the article
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
You can just insert the word "some" as required.
Agreed, but the headline wouldn't travel nearly as well, if at all.
> Why some Americans are switching to soy
Would be more accurate than
> Why Americans are switching to soy
But wouldn't garner nearly the same amount of clicks.
There is conscious exaggeration in omitting 'some' - a fluff-blog click-farm trope I don't enjoy seeing in the developer space.
You could drop the verb clause? This would make the headline accurate while keeping it punchy.
> why Americans switch to soy
And
> why developers switch to codeberg
But the cynic in me thinks that the form of the headline that drastically overstates the the phenomenon in question by implication is something that has been workshopped and is commonly used because it turns something kind of boring into a spectacle.
Some means at least one.
If I didn't know any better, I would think this is a thinly veiled marketing piece by "Codeberg", a company that I never heard of until this headline.
https://www.paulgraham.com/submarine.html
Codeberg is a non-profit. They only host public FOSS projects. I don't think they're looking for marketing.
It’s just a PR article to promote alternative
Given the context, I had to read this twice to understand that PR is not referring to Pull Request here.
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”.
> 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...
This is not an effective argument, given that these smaller services are uniformly retrograde in functionality.
Agreed, if there was something actually better to switch to then I'd be interested. But seems unlikely for that to happen now - it's easier than ever to build a new github, at least the app itself but i doubt someone's gonna bother with the business effort needed to actually build it up as a reliable, trustworthy option that you know will be around for a while, which is a process that takes years when you know you're going to get disrupted. It would probably have to be open source to get early adopters to use it but somehow be nicer to use than GitHub, and there's basically no money to be made
If only Github was reliable and trustworthy, we wouldn't be having this discussion.
When cars started replacing horses, they didn’t have more features. They just didn’t get sick or temperamental, and they didn’t shit all over the streets.
Going more than 20 miles an hour and not dying of heat exhaustion (or because your oats are too dry) are pretty big features.
But my point was that it’s a bad analogy. People are opting out of GitHub despite the alternatives having fewer features. You can read that as an ideological choice or as a YAGNI one. If it’s ideological, then there’s no competitive or feature angle at all; people are doing it because it seems right to them.
(To be clear, I have no problem with this. I think GitHub only gets better after public pressure, as we’ve seen with the last N cycles of product atrophy.)
Agreed. But if I comment on this, I'm promoting the article. What do I do?
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 :)
> If you're talking about the article featured on HN just day(s) ago
I’m not.
Engaging :)
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.
[dead]
I tried codeberg, used it a year, then early this year in all their wisdom codeberg decided to show adverserial random text instead of my repo, reporteldly to mess up llm training to user agents they weren't sure were human.
Codeberg had one job, serve my repo, it didn't do that, when brought up, I was told it was a feature not a bug, they could maybe whitelist me but that wasn't my problem, it was that random people got totally blocked or from accessing the repo. I moved back to github.
Is your complaint about Anubis? I see the "checking if you are human" kind of text too when going to codeberg, but it is usually only a second or two. It is because I run a bit of obfuscation and resistance to browser finger printing.
But you find seeing the text for a couple of seconds too annoying to use codeberg? Maybe it is more than 2 seconds for you?
Everyone has their views on what is acceptable in the world and what they will put up with. Just, to my point of view, I think codeberg is trying to fight the good fight in keeping llms from crawling their website.
Codeberg has some extra "anti-scraping" measures than just Anubis which, judging from the community issues[0] sometimes ends up with false positives that cause people to get garbage (example issue[1], potential page with the garbage i found from Google[2]).
Interestingly, someone mentioned that you may get the garbage when searching Codeberg using Google's `site:codeberg.org`
[0] https://codeberg.org/Codeberg/Community/issues
[1] https://codeberg.org/Codeberg/Community/issues/2603
[2] https://codeberg.org/Codeberg-e.V./requests/!~codeberger~!.g...
It’s not Anubis as others have pointed out, they have more going on that blocks you sometimes.
With that said I never used code berg because of Anubis. Something about anime girls popping up turns off any desire to use the product.
I believe they're referencing iocaine[0], which I believe I've seen on codeberg before. I may be wrong though.
[0]: https://iocaine.madhouse-project.org/
[0] Gives me HTTP 418
A rare event you see a wild "HTTP I'm a little Teapot"
Interesting. Well, for information about the project: https://lib.rs/crates/iocaine
The site linked is excluded from the wayback machine.
> only a second or two
If only we had research on the effect of a second or two's effect on user experience.
I wouldn't have a problem with it if it was on expensive endpoints like search or deep history dives, where it matters for server load. But it's every single page, out of some strange sense of righteousness.
ridiculous to call it righteousness when it's allowing the servers to stay functioning
No, the issue is when i visited i got a page of senseless drivel (maybe https://lib.rs/crates/iocaine) with the message i was a bot. There was no way around it except "contact support" as the message said the fix, to solve it myself, was to not be a bot.
https://news.ycombinator.com/item?id=48845668 explains it better than i did.
I accessed through mobile, maybe got a bad ip, i don't don't know. But i never could access my repo, and incidently a day before when i tried to share some markdown file in the repo, they told me it wasn't there (i thought it was a them problem then, as i didn't see it).
So that experience and the way it was handled totally ruined my trust. If randomly people can't even access the repo, and that is "working as intended", what's the use to host it there.
Ah! Thank you for the more in depth explanation. I was not aware that codeberg had such a system. As others pointed out, it is probably iocaine.
Also, Anubis is cute!
They explicitly explained that the problem was not "they find seeing the text for a couple seconds too annoying to use codeberg". Codeberg offered to remove that for them so they'd never see it again. So idk why you're asking about that. Whatever the problem is, it must be something other than that, since that problem doesn't even exist.
The problem is blocking or speed-bumping users in general, everyone else besides themselves.
They didn't put things on a public hosting site for them to be hidden or obfuscated or even to have a tiny friction inserted between a user and that user discovering their stuff, or following a link to something.
Saying "2 seconds" as though that makes it insignificant is completely missing the point. 0.2 seconds, if it's 0.2 more than some other path is the same as a total block. Having a link to something do anything at all other than instantly provide that thing is outrageous and unacceptable, if you care about the experience you want to present to your users or audience.
(Still, I'd say codeberg was at least among the best options if you want someone else to host it for you. The bad job of providing a bad experience for legit users while trying to block AI scrapers can be true at the same time as no one else is doing any better without some other worse strings attached like github.)
Seeing the Anubis weeb interstitial is super annoying.
Agreed. I installed the NoPow add-on and that worked (in Firefox).
Sounds like a feature. I didn't think Codeberg would have the balls to do something like it. I think I'm sold.
codeberg's one job gets interfered with by freeloaders hammering expensive views (like git blame) inconsiderately.
You’re so right. I have a public-facing Forgejo server. Before configuring Anubis, scrapers were sending it about 600K requests per day. Copying and pasting from my blog post about it:
* For every Git commit, fetch the version of every file in the repository at that commit.
* See git blame for every file at every commit.
* Attempt to download the archive of each repo at every commit.
* Run every possible pull request search filter combination.
* Run every possible issue search filter combination.
* Fetch each of those URLs at random from some residential IP in Brazil that had not ever accessed my server before.
Afterward, it dropped to several hundred. Expect anti-attack features to keep getting stranger and more visible as scraper get still more aggressive.
This matches my experience of running a public mediawiki server. The bots (mostly Facebook/Meta) will for every single change fetch every single page again, over and over.
I can’t tell if it’s incompetence or malice.
What do you use to monitor this? I don't really keep a close enough eye on my services to know what the traffic is doing (and haven't had any issues) but maybe I should start
I just tailed the web logs for a bit and saw that it was wild. For fun, I fed 10 minutes of logs into an AI and it picked up a lot of signal I didn't catch at first glance, like clients claiming to be MSIE 7 on Android 3 and such. I added some reject rules to the webserver in front of Forgejo but that only made a dent in the traffic, alas.
That's crazy. So 600,000 someones are absolutely starved for data.
Or one someone with too much money and too little sense misplaced a decimal point in their ScraPy setup?
I think it was the latter.
And the cruel irony is that these are FOSS Git repos I'm publicly sharing. I'd've been fine with them cloning the repo and analyzing away to their heart's content. That's not the way their scraper's wired, though.
Couldn't they just rate limit them? Are they literally using a new IP for every request?
There are indeed scrapers which use tens of thousands of distinct IPs, and so rate limiting them isn't a solution.
We used to call this snow shoeing.
In my experience, yes. At the peak of a scraper flood I was dealing with, I'd say about 90% of the traffic was from a unique IP. I'd never seen anything quite like that before.
That's the reason I left as well. I complained, was told I need to drop the attitude, or leave.
so I left.
must be harder than I think running a src forge
Today it appears you have two options 1) Anubis or similar 2) Accept that 90-95% of your traffic is LLM scrapers fetching every file for every commit for every repo etc etc. There's some comments above with specific lists.
I built bot detection into the web framework I wrote. It's blocking well over >90% of the abuse.
The secret 3rd option is write some code.
It's been 9 months since I ditched Github.
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
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 self host gitea for my personal-personal projects, the ones nobody will ever see.
For the personal-opensource ones, I am on Github because this is where everyone is when I want to share/collaborate etc
actually good approach and the one that could make me switch to selfhosting
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.
You use github as a backup, why bother self-hosting then?
Github is an extra layer of backup, among normal backups.
[edit]
Notable reasons:
- Github runners went oftenly out of space & they were slow. With self hosted runners I don't have these issues anymore because I control the hardware.
Previously I was paying Docker Build Cloud/Depot for performance + Github Pro for extra minutes. Now it's zero cost, superb performance and unlimited minutes.
- I have a centralized registry with private packages and images.
- It's secure, I don't worry if I accidentally make a repo public or leak secrets. I control the access to it in network level.
- I own everything, in case something goes nuts (eg lose access to GH) I'm safe.
If you're using github as a backup, then whatever you say you own, they also own. So I still don't understand why you are self-hosting and using github as an extra layer of backup. Seems philosophically inconsistent.
Nothing philosophical here. Whatever makes you productive.
I own the git, the runners, the registry etc and just use GH for backup. GH cannot interrupt suddenly my business.
[dead]
Not GP. Probably less dependencies on github, e.g. actions which sometimes don't work. This way github is a "dumb backup".
I selfhost forgejo (gitea fork) on home sever (nuc), similar setup with tailscale. I was planning to setup git mirror on a remote VM for backup, but since I am the only one using it and have everything on dev laptop and remote backups of nuc server I didn't bother to do that (I know I still should).
what are your home server specs to self host forgejo?
NUC 11 i5-1135G7 with 32gb of ram and one nvme. Bought it new years ago.
For forgejo I host only code and packages there, everything is built on dev pc since I am the only one using it. It also host postgresql server, jellyfin (for music only), custom OPDS server (for books only) SMB shared directory I can use from multiple PCs, some test environments for personal projects and possibly more things I forgot.
It's a small pc but it works without problems. I do have to clean it from time to time (with cheap battery-powered air duster) to keep the thermals low.
yeah i have 3 NUCs at home, albeit with less ram and smaller hdd than yours. im already using them for dokploy and other services so trying to selfhost forgeo or gitea might bring it to its max.
> This way github is a "dumb backup"
Eh? What GH department do you work in anyways? Training Data Sustainability?
I don't understand your question. If I was working for GH would I be selfhosting forgejo?
edit: If you think I sound like an LLM, I guess that is what happens when most of your interactions in English language are with LLMs.
Given the state of it, you might well ...
I missed the opportunity to add "you are absolutely right".
> If I was working for GH would I be selfhosting forgejo?
Why not. But you are saying github as a dumb backup makes sense when you do self-hosting. I don't think you really believe that.
GH as a backup and a downstream public mirror for a self-hosted gitea or similar is entirely purposful. None of your comments have made any sense.
Yeah I should've asked why they are self-hosting in the first place. I was wrong in assuming that they wanted to host it themselves (that is, avoiding an external dependency).
Because the existence and continued normal operation of the primary is not dependent upon the capricious whims or instability of GitHub.
The self hosting will still be there and working as expected no matter what GH does (fails... again, DMCAs the repo, bans the account, etc.). Self hosting isn't only about being the only one with the data, it's also for the independence aspect. GH as a backup doesn't hinder the independence. Network effects are strong and make a lot of developers still have a GH presence as a secondary platform.
The evolution is when one can finally fully disconnect from GH, the main self hosted platform will continue to operate as if nothing happened.
A migration can have a period of parallel running.
[dead]
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.
As a developer who ditched Github and decided to self-host, there is only one reason. It's not technical difficulties, politics, nor AI. It's Microsoft. Like Apple, Facebook etc, I have a deep loathing for Microsoft and I want to remove as much of it from my life as I am able.
I now run Git on a pi using Gitea and Forgejo. I can now upload files of a size unheard of in GitHub, Claude can make a PR by itself that I can diff, edit, then merge, and even with the mighty power of a single pi 3b+, it feels more responsive.
Microsoft is a mafia that sometimes produces code (poorly).
Microsoft's human rights stain will take decades to wear off (just like Meta, Google, and Amazon).
Eh, the stain will only grow worse over time.
They certainly are not doing anything to start bleaching it right now.
Why Gitea _and_ Forgejo, rather than just one or the other?
Mostly because developers (me included) don't like to be told we are being laid off due to AI that was trained on our free open-source hobby projects.
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.
> It sucks, but that ship has sailed.
Hey, this attitude sucks. Would you tell this to someone who got beat up in an alley?
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.
Genuinely disgusting that your first thought is therapy speak platitudes rather than holding someone accountable.
how switching to an alternative like codeberg would prevent it? ti can be scraped as well
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.
That's where it's all goin'. Private, outta sight. Keep the 'bots out, an' you keep the man out.
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's a losing game to try to keep any FOSS code completely isolated from large scraping data sets. All you can do is never share it, or share it and eventually it'll get consumed by the machine.
Spend your time making better software you and your friends can use... don't waste the time trying to stop others from making another painfully average codegen machine.
Or just stop making FOSS code
Starve the greedy beast
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.
Github for open source was always my favourite part of it.
Github for private repos has long had security issues, every time a serious issue announced it makes me wonder how long it's quietly existed and been exploited, and how many other holes are currently exploited that aren't well known.
Before github, people hosted their own repos all the time. Learning about alternatives, even if they aren't for you in all cases, is still worth it.
Has anyone had any issues with codeberg or other alternatives?
> One new user joins every second
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.
Given more code hosting services, I wonder if we'll also see a corresponding increase in the number of alternative VCS or if git is legitimately very entrenched as a tool. I am just being a bit grouchy but I do wish there was more development of alternative VCSs. pijul at least looks cool even if I don't know if it scales well. Git LFS can be somewhat finicky to work with so maybe we'll see perforce like systems. It's obviously not the most practical thing to have a variety of very different VCS's and definitely a PITA to learn multiple tools but git does seem somewhat suboptimal given the number of anecdotes about people just re-cloning the repo. I was recently trying jj and it seemed to work well (excluding the lack of LFS support) so here's hoping.
I do think git serves a lot of people's needs very well. I see a lot of people being quite happy with direct github alternatives. Self hosting may become more popular too, I hadn't heard too many self hosting stories until recently. If LFS is your pain (I cannot stand to use git lfs) I have to mention oxen, which is open source and self hostable, but also has a hosted option. It handles large files right out of the gate and doing so faster than everyone else is one of its core missions. Still using github for distribution though, lol, see https://github.com/Oxen-AI/Oxen. I am also curious about where jj will go, its philosophy is really interesing, and sounds promising.
Genuinely curious here for someone who has tried self hosting git and has found it a pita to maintain...i want to know what is it that devs are flocking to other platforms and how are we sure that they won't pull all the red card signals that github is said to pull off.
Have been self-hosting GitLab for my org a few years by now, with quite a few users (>800 atm). Updates are automatic via the GitLab Omnibus package repos. Once or twice per year some update requires intervention. Otherwise, nothing bad happens. Very happy so far.
Biggest problem at the moment is that AI scrapers (curse them and their owners, pox be upon their houses!) sometimes bring things to a crawl. But nothing that a few firewall rules and anoubis won't solve.
If you are running an org you should be putting all your private services behind netbird or tailscale at minimum. Zero public infra exposure beyond them.
The service isn't supposed to be private, some repos are intentionally public. And including all external collaborators in some VPN scheme is not possible at that scale.
Sorry, I assumed you were running a typical corp
Forgejo is fairly simple to run – way lighter and simpler than GitLab even. (GitLab is quite okay, too!)
If you want a hosted service, go for Codeberg. It’s run by a German non-profit (so it’ll be hard to bite and switch OpenAI-style). Only free/open source projects are accepted, though.
How is Codeberg connectivity from outside Europe, nowadays?
From my experience, things are looking great from East and Southeast Asia. Why?
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.
Do you serve the instance on the public internet or on a VPN?
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.
VPN, accessible only from inside the corporate network
Just use tailscale
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.)
Do you self host chat?
Yes, Rocket.Chat
The recent abrupt removal of the ability to see who has starred a project isn't a good move. Things like this certainly erode trust.
https://github.com/orgs/community/discussions/201209
+ predatory pricing hikes for AI
+ not honouring yearly commitments plans
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?
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".
In case you didn't know, you don't need to pay for private repos.
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.
I'm trying sourcehut at the moment https://sourcehut.org/ and it seems really good - very simple and fast. And does seem to be free for hosting open source projects.
Anyone else used it and have thoughts on it?
I really do not like that Drew (the owner and developer) is extremely dogmatic and political, and likes to get involved in what is and is jot allowed to be hosted on “his” forge.
Imo, service providers should be neutral and get involved only as far as required by law.
I would not trust Sourcehut. If Drew decides one day that he does not like you, your politics, or your industry, he will just cut you off. That is no foundation to build on.
I use it as my main forge and could not be happier. It also seems to be the only forge that allows its web frontend to work with JS disabled. (That I am aware of)
I really really love it but I just can’t get myself to embrace the email workflow. Maybe I’ll make a “pull requests for git.sr.ht” app one day!
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.
Anyone has suggestions for hosting open source hobby projects managed with Mercurial.
Loved Bitbucket's Mercurial offering. Looking for a replacement.
Sourcehut has Mercurial support: https://hg.sr.ht/
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
Just in time. I was about to edit my own comment to mention
https://hg.sr.ht/ that I learned of from a comment by frabcus on this post.
Thanks for links.
I co-maintain a popular Linux launcher, Fuzzel, on Codeberg. We have no shortage of contributors despite not being on Github and the quality of contributors there is fairly high. In the past there were performance problems, but it's worked well enough lately.
My personal repos are on Gitlab. The UI is cluttered, but it works well enough.
We ditched GitHub for self-hosted Forgejo and could not be happier. The experience is smoother, faster and distraction-free.
Extreme generalization, most devs aren't ditching GitHub yet.
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.
It's pretty much broken by AI. Not only your private repos are not private, but also the LLM will leak them.
FYI that Codeberg is currently holding a vote to broadly ban projects written mostly using AI, so its not a neutral space for hosting your projects like GitHub: https://codeberg.org/Codeberg/org/pulls/1253 https://tldr.nettime.org/@tante/116880003584050912
I've been on Codeberg for my projects since September but will pull them if this happens. I see no definition of what "mostly using AI" means and refuse to be "punished" for the choice of tooling I use. The tone and content of the thread on Mastodon is absolutely toxic.
I'd rather self-host than deal with the ambiguity and threats. Invariably from people who don't even make anything.
I have plenty of concerns around "AI" and project governance and IP law. My projects are copyleft, I'm in favour of free software, bleah bleah bleah.
But this just looks like pitchforks and mob behaviour.
Same here, i've been on Codeberg since 2022 but if this passes, i'll most likely move my stuff elsewhere (my projects do not even have any AI code, i just dislike the idea of telling people what tools to use to make their own stuff).
I dislike GitHub and i'm using a shared host for my website which wouldn't let me install something like Forgejo (this is by choice as i don't want to bother with sysadmin stuff, my site is almost all static HTML with only a couple PHP scripts for some minor tasks). So i guess i'll be migrating to GitLab, even though i do not really like its UX. Or maybe i'll use some other Forgejo instance like CodeFloe. URLs aside, migrating between Forgejo instances should be easy.
Yep. I have a self-hosted Gerrit instance that I was using for other more private stuff. I could in theory move over to there.
The downside is then having to worry about backups. But also lack of all the other "forge" things like issue tracking and so on. And lack of public exposure.
Frustrating, really. I felt welcome on Codeberg up until now. But that Mastodon thread is... wow.
If I understand it correctly, their target is projects like the recent pgrust (PostgreSQL in rust) which are full conversions of projects that no-one else will ever use. It's just noise.
Proposed language only says:
Who defines "mostly"? And what's the definition of "written"? If I'm heavily prompting, curating, reviewing, and editing after the agent has gone through, is that somehow now "allowed?" Who makes the decision and how can I appeal?Almost everything I write now has some agentic aspect to it, and I'm by far not the only one. That wasn't the case when I started my projects but it is now.
But I'm a software engineer with 25-30 years of experience and a computer nerd since 1979 when they wheeled the Apple II into my Grade 1 classroom. I'll reserve the right to decide what tooling I use to write my copylefted free software.
Given the Mastodon thread linked above I don't trust the people involved to be the judge jury and executioner. Insults flew freely and moral absolutism with it. I'll just pull my project off their infrastructure rather than be subject to their personal whims.
Well, that's a good thing. Being "neutral" _is_ taking a side. It means you're taking an amoral stance even in big questions.
Every organization has a stance. We're just become used to companies that take a stance of "as long as we get paid".
No, being "neutral" is not "taking a side", that's the definition of the word. If you pre-moderate you open a whole can of worms. What will they ban next? This makes me not want to use Codeberg ever because it's not plannable.
Do you feel this way about everything? What about pre-moderating CSAM?
Clearly some things are harmful enough that they should be banned.
If you're going to jump ship because they're banning AI slop, then I think the feature is working as intended.
Calling all uses of such tooling "AI slop" right away shows your hand as only holding a blunt, crude, instrument.
Many of us are capable of more nuance that that and don't apply the label "slop" to everything where generative AI is involved. The linked Mastodon shows no nuance, and neither does your reply.
Two things can be true at once: GitHub has degraded into a "slop" cesspit and that agentic tooling has legitimate use cases.
Finally people can also legitimately object to a host mandating what tools can be used in authorship even if they don't particularly care for those tools themselves.
Why do you think you can judge my code and how I program from a comment?
[dead]
Today GitHub blocked me until I turned Apple private relay off. I wasn’t logged in
I guess three nines availability is important.
Not even hitting 1 nine at the moment - https://mrshu.github.io/github-statuses/
I object!
The dashboard clearly says 89.15% uptime!
Who says nines need to be leading?
With enough precision in the time metric there are infinitely many nines!
Somewhere in 3.14159...% uptime there are as many nines as you like!
... assuming time is infinitely divisible. (Not a physicist, is that open?)
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.
So sad to see that no articles about this even mention Mercurial. This is a golden opportunity for Hg providers to shine.
this not a `git` failure per se...
Yes, but the thing is just that if people are looking around for new providers it's an opportunity for alternative systems to attract attention and users.
I understand what you convey, however, users are tired of the git GUI, not git itself.
because github is subpar, you will try to convince people to migrate their repositories to $NOT_GIT so they can try a different hosting service ? doesn’t sound like a pragmatic choice to me.
I never liked git and held on the Mercurial until a couple years ago. Now, for all I do, jj is a better Mercurial than Mercurial.
I miss Bitbucket's Mercurial offering.
But Github is already fully volatile, capricious, fickle, erratic, unpredictable, variable, inconsistent, changeable, unstable, whimsical, protean, fluid, and a polluting poisonous room temperature liquid heavy metal, so why would you also need mercurial?
:)
Was very happy to find that Hginit.com has been given a new life here
https://hginit.github.io/
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.
https://fluxer.app/
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.
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.
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.
>it's intuitive
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.
Acquiring github users may not be their highest priority.
I’m not disputing how intuitive the GitHub interface is, but seriously, why is it so hard for technical professionals to set aside 10–20 minutes of their time to learn a new interface? Why has this even become an issue worth discussing?
Not to mention GH's pull request UI is just...
dogshit.
One of the worst code review tools I've ever used. If not the worst.
Inst gitea doing this?
> copy the UI
Good luck. The amount of features and screens on GitHub are vast aside from just those code / issues / PRs tabs.
The appeal of GitHub for me is not only in the git hosting, but also in codespaces. It gives me:
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.
Im just glad the wider world has finally snapped out of their GitHub mono culture trance.
I have been self hosting my repositories on my own VPS using nothing but Git itself (you don't need to install anything else). Sharing write access is a bit of a pain, but I tend to work on things alone anyway.
Wouldn't it be nice to have a repo which AI couldn't access?
Github getting bloated!? after MS?
I hope some developers will see this message. I know that you work however you wish and you do what you wanna do, but there is a lot of people that will refuse to signup for GitHub (for many reasons), and in consequence, won't contribute to your project. I'm sorry if it sounded a bit off or I offended someone by stating this.
Anyone using or been trying out Radicle recently?
https://radicle.dev
EDIT: https://news.ycombinator.com/item?id=48147603
If the goal is lowering GitHub dependence without going full self-host, Buddy has built-in Git hosting alongside its CI/CD, so repos and pipelines live in the same place.
Is there a bsky or mastodon extension for serving git repos and commits? Would be interesting to see this.
Doesn't tangled do that (at least for the ATproto)? See https://tangled.org/
You'd be surprised how easy it is to self-host GitLab with Docker Compose, GitLab has an official "Omnibus" Docker image. No need to handicap yourself with Gitea/Forgejo/whatever, you can just use an industry-standard platform without much effort.
Hardware requirements are nowhere close to high either.
Spinning up a docker image with GitLab is easy indeed, but the installation is huge and you can feel it doing weird stuff (and it keeps doing "stuff" all the time even if nobody does anything - before you even bother to create extra users or repositories or whatever). The whole setup seems to be designed to be opaque.
Forgejo is a single self-contained ~110MB binary (it contains all assets) you can drop anywhere and it'll just work - depending on your needs you most likely wont even need to bother with a DB server as it can use SQLite (my local install is doing that).
Also FWIW personally i find Forgejo's UX much better compared to GitLab.
What makes GitLab "industry-standard" and Forgejo not?
one reason (among others) — compliance.
a corporate/b2b saas environment without stuff like this is often a non starter.
- SOC2
- ISO/IEC 27001:2022
- ISO/IEC 27017:2015
- ISO/IEC 27018:2019
- VPAT 508
https://about.gitlab.com/security/
no mention of these on the forejo site, so i can’t put “our internal software is all SOC2/ISO NUMBER compliant” as a bullet point on a slide deck.
it is theatre. but it’s industry theatre.
Aren't those about organizational processes instead of just specific software features? For example, I don't think self-hosting gitlab is enough to claim ISO/IEC 27001, just based on this snippet from wikipedia:
> ISO/IEC 27001 requires that management:
> Systematically examine the organization's information security risks, taking account of the threats, vulnerabilities, and impacts;
> Design and implement a coherent and comprehensive suite of information security controls and/or other forms of risk treatment (such as risk avoidance or risk transfer) to address those risks that are deemed unacceptable; and
> Adopt an overarching management process to ensure that the information security controls continue to meet the organization's information security needs on an ongoing basis.
> I don't think self-hosting gitlab is enough to claim ISO/IEC 27001
probably not. but it does mean you can focus on the rest of your stack rather than having to go through every single process from the ground up.
I don’t know if Gitlab is an industry standard, but I’ve never heard of Forgejo. I worked for a headless CMS company and the only three providers we ever had requests for were GitHub, Bitbucket, and Gitlab. Gitlab is big enough to be generally adopted by governments. I think it’s fair to say it’s at least a lot closer to being an industry standard then Forgejo.
(Aside: I would likely never use Gitlab by choice, and would consider looking into Forgejo)
Probably because those are service providers and Forgejo is software. Same reason people ask for Microsoft account and Google account integration, but nobody ever asks for Linux account integration in your cloud software.
Forgejo is mention in the article, it powers Codeberg. I agree it doesn't have high adoption, but the news is that it is growing.
Popularity.
It's also easy without Docker. apt install gitlab-ee
I can barely hold the attention of someone, developers included, when discussing my coding projects. I cannot imagine how difficult it would be if it wasn't even on Github. I'm a small developer, I don't really have the privilege to choose.
I left because Github became too LLM obsessed and I find that to not align with my personal ideals. Its not really deeper than that. The programming community will continue to splinter more into those that care about the craft and those that only care about outcomes.
No mention yet of Radicle? I'm curious if anyone's tried it for a real-world project.
it's definitely a cool esoteric thing :) I bet some other features would make it gain traction.
I see a lot of self hosting based a whatever solution but I don’t see any nice GitHub proxies to avoid rate limits while just simple reusing actions.
I use zot within my ci for containers and an apt proxy but I’m missing a solution for github api calls for my ci runners
self-hosted gitea/forgejo is still better
For those who moved away from GitHub: do you miss the social/discovery side of it?
I mean things like comments on issues/PRs, stars, followers, finding existing repos, seeing which projects are popular, and getting drive-by contributions.
Or does that matter less than I imagine once you self-host and mirror public repos back to GitHub/Codeberg?
I used to love browsing GitHub Explore Trending but in the last two years it is almost exclusively AI apps that are slight improvement on the previous trending page of AI apps.
GitHub explore could be way more interesting with a simple filter algorithm.
Switching from one centralized git repository to another that also injects political messaging into URLs without consents.
What do you mean by that? Are you referring to Codeberg? Out the frying pan straight to the fire...
I think and hope we see a lot more of this before the adversarial imperative returns from the company side.
People using Claude Fable to just make replacements for disgustingly enshittified software. We desperately need browser extensions to help make websites less scummy across the board as well.
Anyone tried tangle as a replacement? Verdict?
No affiliation, but http://code.storage gets my vote.
People are going to copy GitHub the way people copied Facebook… how is "Threads" doing again?
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
[dead]
I'm hosting my own Forgejo instance and it's great. Coolify as well :) It's fun!
Yeah over on Mastodon, there's a few really nasty threads about Codeberg requesting voting for approved folks to ban ALL "AI".
Of course, it aint going well.
https://tldr.nettime.org/@tante/116880003584050912
What do you mean, it ain’t going well? It looks like the measure is going to pass, that’s the ideal outcome.
Agree with this
“Verifying you are human”
I’m gone.
Any site that does that. I’m gone. I’m not waiting around for your bullshit.
It’s so scammy and dumb looking anyway.
Cloudflare ruined half the web.
Codeberg might be cool? I’ll never know because they use Cloudflare.
Cloudflare makes me leave websites every day and never return to them.
Dumbest technology ever implemented. It’s joke anyway - DDoS is easily handled far more elegantly.
Codeberg doesn't use Cloudflare.
Fake news
[flagged]
[flagged]
[flagged]
[flagged]
[dead]
[flagged]
[flagged]
[dead]
[dead]
[dead]
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.
[dead]
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?
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.
Can’t go a day without propaganda on HN.
Everything is propaganda, including your comment and this one.
[flagged]
> Anyone still doing ongoing business with them in 2026 is, imho, a fool.
So that would be almost everyone.
I mean, I poisoned and then deleted my GH account the day Microsoft acquisition was confirmed, because it was obvious where it's heading.
[flagged]
Don, you’ve got me wrong. I’d love to have a civil discussion with you to clear up your misconceptions.
My email and phone number’s in my profile. Do please get in touch.