ZH version is available. Content is displayed in original English for accuracy.
What is a bit unique is: 1) we built it in Rust and 2) the website is a little odd. Its design is inspired by CLIs (e.g., fzf, broot, vim) instead of web apps, and as such, lacks some affordances that you might typically expect in favor of keyboard-driven instant navigations (we have the very ambitious goal of an FCP of 100ms). In case you're curious, here's how we we built it: https://gitdot.io/designs
We recognize that we're making some bold claims here and are also well aware that we have much to learn. Building software is still hard, and that's a fact we seem to relearn everyday.
But we wanted to share what we built so far nonetheless.
Cheers, thank y'all for reading, and till the next —paul & mikkel.
Discussion (290 Comments)Read Original on HackerNews
Second, when I browsed from an actual desktop, and clicked on links for files it was all slow as hell - specifically the part when you click on a file an expect it to just load, you instead get: 1) some layout switch which looks like page reload 2) then it says "loading..." for several seconds.
After looking at the source code, it appears to be React or similar frontend framework... Ugh. I don't know why people choose to use that stuff, just have a regular SSR which would work a hundred times faster and is more pleasant. And if you really want an SPA, don't use React, Vue or Svelte (and similar), it's horrible and always slow.
Finally, since this appears to be a YC company, it shouldn't matter what's it written in. In fact, I don't even know why Rust would be a good thing here when Go or even Rails/Django would work just fine - but again, it just reinforces the meme that if it's written in Rust, you'll surely hear about it.
Overall, the minimalism idea is welcomed, but it supposedly should appeal to people like myself and it doesn't for all the reasons I mentioned above.
FWIW I've noticed that agents are pretty good at writing "High-Level Rust" for most basic applications, which gives you pretty great performance (orders of magnitude faster than RoR), great deployment, probably great security if steered by a senior, and pretty great maintenance again if originally steered by a senior. I feel like this is a not-so-secret secret.
I personally won't be using dynamic languages for anything but toy scripts now. (well, except JS, which is hard to avoid with the massive size of WASM bundles)
P.S. I assume Go is still great as well, but IMO Go no longer has an identity. What are they going for anyways? Garbage collected rust? "How to invent a perfect niche and then throw it all away in 21 days". /rant
It's probably polarizing and I honestly don't know why I felt the need to rant about it. /shrug
I am indeed curious to see how an agent-coded project turns out in the long term.
Coherencey from LLMs decreases as the source increases in volume, waiting to see how this all turns out.
I wonder if feature velocity closely tracks LLM context management innovations.
May be this encourages more modular systems if those improvements can't come fast enough.
I am not a rust evangelist, I am not a rust programmer or programmer at all; however while evaluating tools to use, anything written in Rust and Go at-least gets me to look at the project in more detail, since they most likely are able to ship statically linked binaries, which has been one of the key criteria for my personal evaluation of tools to select and use.
So, you might not consider it as a valid signal, however it might be for other users. Even if it has a negative connotation for you. Which in itself, again might be a good filter in case you don’t want to use it.
How can you still not see any advantage? Or was the point of your comment to say that you think the only real motivation is self or Rust promotion, suggesting some dishonesty amongst the people you're responding to?
But the "Written in rust" stuff comes with so much falsehood baggage about automatic best-in-class performance, automatic perfect safety, automatic rigor, that it is detrimental to just say "written in rust" and not provide a more nuanced statement about why that's good (or just drop it as a headline item!). You might not agree because you can single out one automatic benefit of rust b/c of Cargo (not rust), but even that is suspect b/c Rust can ship with dynamic libraries if configured to do so.
In this case, the credibility and benefits are clearer if they said "Static binary". This is all just context. If you see a benefit to that headline, that's ok.
You might have better devops experience as an end-user of the said software if its statically linked and its a good trend in software sure but it is not unique to go or rust.
Admitting you don't know how to code and still trying to argue rust is better is just wild sorry.
C and C++? You've got to be joking. If the project provides static binaries, sure, but I don't want to have to worry about finding a necronomicon and summoning the correct kind of imp required to properly use whatever insane build system the project is using.
Literally last week I was porting something to Linux and had to rewrite the libwayland build scripts because they only expose a shared object.
There's also an expectation that you're going to install them via your system's package manager, not build them, so a lot of them use insane build systems (autotools, meson).
Huh? It does. Only libc is dynamically linked, by default, which --iirc-- all programs will commonly need anyway. All the rest is statically linked.
In fact, it takes some hoop-jumping to build dynamically linked binaries with cargo.
Personally i do not even write Go myself yet when i see some webapp made in Go i'm pretty much always sure that it'll be a self-contained static binary. Yes, there are probably exceptions (and i haven't surveyed all webapps made with Go) but it has been common enough to feel as the standard approach to me - and if i ever decide to make a binary webapp, i'll probably reach for Go to do it.
But for sure a systems language is going to be far faster on paper and Rails is far from perfect and does have some performance foot guns you need to avoid. And yeah, architecture is everything.
GitHub is so data heavy and there's so little reactivity that it should really be server side rendered.
They claim it is better just because of the language, ignoring the features gap, the size of the team developing and supporting the software and not having solved any issue with the software they compete against.
And to be clear, this in not in favour of GH, it is against the mentality that the programming language makes better products and programmers.
Nowadays when everyone and their dog are either vibe-coding with Rust or constantly shouting about it's superiority, I've lost any interest in the language. I'm learning Zig.
None of this makes Rust worse as a language
Zig, on the other hand, is a breath of fresh air. There's so much to love about it, including language design that values clarity and simplicity, and the cross-compiler which is how I got started on it, by using Zig as a better build system for C projects. Their move away from GitHub and vibe-coded contributions is also commendable, it's more reason I trust their sensible judgement.
I mean, it clearly is better (in certain contexts), see e.g. https://blog.google/security/rust-in-android-move-fast-fix-t...
> We adopted Rust for its security and are seeing a 1000x reduction in memory safety vulnerability density compared to Android’s C and C++ code. But the biggest surprise was Rust's impact on software delivery. With Rust changes having a 4x lower rollback rate and spending 25% less time in code review, the safer path is now also the faster one.
If you really think that switching language is the main driver to get safe programs, the you are on the list of people replaceable by LLMs.
Othewise you have to understand that architettural chioces, concurrency, (weak) cryptographic function and user stupidity have a significant impact, no matter what language you use. Memory management is just a part of the problem.
only for very few companies, memory safety issues are a thing. most of those dont use rust because it lacks tooling and certification paths -_-.
rust is like AI and cloud. heavily marketed, not bad, but not as good as whats on the box -_-. it makes life usually harder for a lot of aspects (ofc cleverly avoid in publications about it -_-)
Fast, Safe, Lightweight, Statically linked (plop a precompiled binary in ~/.local/bin and run it), few/shallow dependencies, senior developers, "Done" software.
Now, certainly no guarantees, enough counter-examples, I know. And attributes that one can get with anything from PHP via Javascript to Lisp as well. Some attributes have stronger correlation than others too.
But, in general, "rust" has a (much) higher chance of meeting these attributes. I care about those attributes above anything else.
> few/shallow dependencies
is just flat-out wrong. There's (sadly) a big culture of sprawling, deep dependencies in the Rust community. There are of course many counterexamples too, but it's definitely not correct that Rust signals the virtue you refer to.
(Incidentally, the misguided idea among some that as long as a dependency isn't through FFI, it doesn't count as a dependency, is one of the things I dislike the most about Rust culture. That, and pervasive bundling.)
I have not heard that idea a single time. There's definitely the idea that FFI dependencies "count more" / add more baggage (because there's a bigger risk that the build fails, it's harder to investigate memory safety, ...). Absolutely not that a non-FFI dependency "does not count".
Recently encountered someone going 'full crab' with rust & vk trying to chase perf, they were not happy when some thoughtfully written modern JS beat the brakes off it even with the overhead of the browser (~2000fps vs 750)
This kind of thoughtless "its rust!!!" shit causes people who don't know better to waste time chasing their tails. Claude seems to encourage it :\
In reality, the programming language tells you all kinds of subtle things: probabilities about the way the software will feel to use, how stable it’s likely to be, how fast, what the author is likely to focus on.
I found one of the best jobs of my life in 2015 by asking “who’s doing interesting things in Atlanta in Go?” Not because I was uncompromisingly settled on Go, but because in 2015, using Go (often) connoted a certain approach, a certain type of engineering, a certain constellation of values.
So please stop pretending the whole gestalt of programming languages and their communities don’t deeply affect the resulting software.
(I say this with no unkindness intended, mostly to all of hackernews)
So called "The vibe". And the vibe emanating from Rust is sometimes unbearable. Like claiming the main feature of the project is that it's written in Rust.
> 7. How does gitdot make money?
> We don't.
> We are fortunate enough to have raised a small pre-seed round from investors we are happy to call friends, and also to be at a point in our lives where we are financially independent and in good health.
The founders of Artifact said the same thing. They had no plan to make money and once they used their own capital they decided not to continue running it anymore (no-one else wanted to invest in Artifact).
So it is only a matter of time until they eventually need to make money, raise money or shut it down.
In comparison CodeBerg [1] and SourceHut [2] both offer Git hosting but don't merely describe themselves as "GitHub but X".
[0] https://gitdot.io/faq
[1] https://codeberg.org/
[2] https://sourcehut.org/
/s, I love zig, but I'm consistently surprised of how popular any post containing zig in the title makes it to the front page almost daily, doesn't happen with any other language.
On mobile I literally refreshed the page because I thought my wi-fi broke... Maybe it looks a bit better on desktop
If I could make one suggestion, I really like the old MacOS "inspector" pattern. Basically a consistent way to get meta-information about any "thing" the user chooses to inspect. Your right sidebar is going towards that, but it would need some work to make it more consistent between views.
GitHub's UI has these weird meta-states/restrictions that are so badly explained in the UI they feel like bugs. Each line gets a [...] menu in github which lets you see the blame/spawn a issue linking to it/get a permalink/etc. It's a totally different UI in the diff view, and then totally different again if you're looking at a comment referencing a line in a diff AND different if it's referencing a permalink to a line in a file, even if it's the same code that would be in that diff!
I want the UI to have obvious "nouns". If the UI is showing me a line of code, even if it's in a diff view, let me "inspect" it and get the exact same meta-info + tools I get for lines of code anywhere. It's "a line", not a weird meta state of "a line, but you're in the comment of a PR linking to this line".
Same concept applies to comments/commits/authors/etc. If the UI shows me a username, I should be able to pull up a "who is that again" inspector. Going into github's commit view, clicking on a name... and being sent to a filtered list of that person's commits makes zero sense to me because this is the ONLY place where that happens. That behaviour should be a "recent commits" button inside some "user inspector".
i'm not aware of the old macOS inspector pattern, but this sounds super interesting and i agree with the critique of inconsistency in github's behavior.
this reminds me a tad of superhuman's right panel too which auto-populates upon writing a time (or typing a name i believe?), which is a feature i do find personally useful as well.
i haven't thought seriously about hovers on nouns quite yet, but this is giving me much to munch on.
thank you sincerely this is dope.
The value of Github is not just hosting projects and issue tracking but the notion that it's probably the largest professional network of developers that can interact with each other, create pull requests and issues on each other's projects, etc. There are many GH alternatives but very few with the same global network of essentially almost every developer out there.
Instead of everybody creating new accounts on each other's not quite Github clones with only a handful of users, why not improve on the state of the art here and allow people to use their user account from a server of their choice and create a proper alternative to Github?
I really hope it works out, would bring so many benefits.
[0] https://codeberg.org/ForgeFed/forgefed
The UI is a bit odd at first, and there are obvious things that need to be fixed (like the invisible input fields), but by the time I had created an account, I'd gotten used to it, and I appreciate how incredibly fast it is.
I'm a bit worried about this:
> How does gitdot make money? > We don't. > We are fortunate enough to have raised a small pre-seed > round from investors (...)
I'd rather know now how you're making money than find out in three years when I'm invested in the platform.
i will also say on accessibility, i recognize the site is a bit too small font in general — and will fix it soon.
I have spent a considerable amount of time learning git, not because I wanted to, but because someone else in my team didn't and inevitably t-boned our repository with a proverbial freight train running commands they didn't understand. It is absolutely unacceptable for a program designed EXACTLY for the purpose of maintaining a history and backup of the evolution of a program to be so unwieldy and occasionally dangerous to use. We can absolutely do better than Git.
Personally I dont care if it is written in Rust or anything else. What I do care is stability, security and features.
I would have enhanced the dashboard a little bit. I understand most of the work was done in the backend. But dashboard is what folks see first. and might lead someone incorrectly judging the work just by looking at that.
Downvote me into oblivion or whatever: It's not relevant to me in the slightest whether you built it in Rust or not, or even Python, Go, Java, Node, PHP etc. There is lots of great and terrible, fast and slow software written in each. OP might only have 3 months of programming experience for all I know.
Why is "made with Rust" any more relevant than "made with Go" to an end user of a software product? It just doesn't matter.
Pre LLMs it did matter, not rust specifically but the language itself. It meant the difference between asking the developer for a feature/fix and getting it done myself if needed.
Now not so much with LLMs, but it does signal something about the development, deployment and supply chain risks involved.
That said, from a UX perspective, personally i prefer Forgejo - which isn't that much slower either (judging from Codeberg, it isn't instant but it is fast enough to feel "fast").
a lot of this we've really come to know as we dug into both it, gittea, gitlab, and all of their internals.
i think the short answer as to a differentiator is design. our goal's to just build the best product possible, one that we'd love to use and one that we hope developers do too.
some of the stuff we've been thinking about include: stacked diffs as PR primitive, a Nix-based CI (that's reproducible and locally testible), a super simple and intuitive bug tracker, and just making the site super duper fast and pleasant to use.
that is to say, there is a _lot_ of surface area that a software forge covers and i think there's a lot of room to make things better.
hope that's clear enough, apologies for any ambiguities, we do NOT have all the answers quite yet haha
i think the honest answer here is autonomy. the freedom to choose our own tech stack, our own product priorities, and our own design.
i'll also admit that i don't know forgejo's own priorities as well as i do our own, and that is negligence that we're due to correct. maybe they are perfectly aligned and it does make sense for us to join forces.
but just as with any FOSS project, they have the freedom to choose what they work on as do we, and i intend on respecting that.
I think the question is why not try to make the other FOSS forges better instead of reinventing the wheel.
Nothing that claimed to be a better GitHub was even GitHub enough to begin with.
GitFar an internal a drop-in replacement for our devs to just change the hostname and keep using exactly the same APIs with exactly the same request and response shapes with extensions delivered through request and response headers!
Kinda ironic isn't it?
However, I think the value proposition should be much higher than simply "GitHub in Rust". GitHub is still a great, complex product, despite its recent problems. There should be some sort of x10 (or at least x5) improvement that will make people _consider_ migrating.
glad to hear the positive feedback :)
- no more silo-ing, moat BS, by design - they are working on vouching which is becoming more important in the age of AI PRs
> The problems I kept running into:
> - I'm lazy.
That should tell you all you need to know about this super hype new fantastic GitHub written in Rust ! ! ! !
we can see one big initial push, not too suspect. Then a few superficial updates, and then nothing for months.
Unfinished, broken, and ostentiably abandond due to being an unmaintaible nightmare as vibeslop tends to be.
the overall reason why we didn't ship a mobile site is that code is inherently very hard to read on mobile
and i think to do that properly, you have to design explicitly for it. (and that is not, in fact, something that i do want to vibe code)
Aspecially as now adding responsive mobile takes 15 min (or less).
i debated this for a while too, some of my thinking for how it is is that i wanted the focus of a repository page to be _the repository_. so as much as we can, trim things that might detract.
it was also done with the intention that it's actually pretty rare for a user to find or explore repositories on github (more likely you find them here on hn or on twitter), so had the restraint of really trying _not_ to make gitdot anything like social media.
but thank you nonetheless for the feedback, i'll revisit it proper and see if i can make this more intuitive.
If anything since the keyboard shortcut already exists, you could always put it in little cursor menu at the bottom of the screen. But yeah up to you, I look forward to your progress.
But, I wanted to say thanks for posting this and being really open in the comments. It’s hard to get so much feedback so quickly. It’s a firehouse of criticism that’s hard to deal with.
You’re handling it well.
thank you for the empathy.
IMO a team like yours can either:
* Use LLMs, in which case you aren't "anti-AI".
* Not use LLMs currently, but the non-use is not due to following a principle, in which case you aren't "anti-AI".
* Not use LLMs and promise never to do so.
I'm happy you are trying something new. But you hurt yourself by engaging in something very old: disingenuity.
(edits for presentation and grammar)
hope our sentiment is clearer in the comments that i've made, i think i've made the mistake of phrasing here with the "anti", i'll revisit it.
edit: dang has removed it, apologies for the confusion i caused, that was my mistake.
What does anti-AI mean? Don't really see anything about it in the design doc except "no AI copilot".
i replied in a separate comment: https://news.ycombinator.com/item?id=48452052 but to reiterate: 1) no AI copilot and 2) no training or selling of your data.
but overall, the general ethos is to focus on the problems that AI is introducing as of now and how we can help solve them, rather than just build AI features with abandon assuming that they're good.
some stuff that we do know about: the influx of slop PRs / slop issues on popular repositories, losing agency our own of code as we AI generates more, and privacy/sovereignty of code.
i've talked a bit about stacked diffs which we do see as one concrete stab in that direction, but a lot here is to be admittedly sketched out.
https://git-scm.com/about/trademark
> [...] you may not use any of the Marks as a syllable in a new word or as part of a portmanteau (e.g., "Gitalicious", "Gitpedia") used as a mark for a third-party product or service.
> Please be aware that GitHub and GitLab are exceptions to this Policy because they are subject to explicit licensing arrangements that pre-date, and thus take precedence, over this Policy.
You might've known that if you hadn't vibe-launched this while for some reason marketing it as anti-AI, but here we are in a world where basic research is a dead art.
i will be honest and say that we didn't do our due diligence here (we simply assumed that it would be okay to do so, given the existence of GitHub, GitLab, GitKraken, GitButler, and so forth).
it does look like from digging in: https://public-inbox.org/git/20170202022655.2jwvudhvo4hmueaw...
that portmanteaus are prohibited by the policy that the Git PLC enforces, which as Jeff notes in his email above, does grant incumbent advantages to grandfathered names (e.g., GitHub, GitLab).
we'll reach out to the conservancy, ask for explicit permission, and if not, rebrand.
But you are squinting really hard if you equate "programmers not being good lawyers" with "they obviously vibe-launched the product".
Public domain and open source are two completely unrelated concepts. If open source were public domain, you could not license usage of it (MIT, GPL, etc.). What is the point of confidently asserting something you're completely ignorant about as though it were factual?
Looking forward to seeing even more features, gitdot.
> AI.
> We view AI as an implementation detail — and do not think that using it is necessarily good.
> In fact, we think it makes many products worse by acting as a bandaid for poor design.
> That isn’t to say we are blind to it, but that we will be judicious in our use of it instead.
Not sure I follow. What feature are the developers referring to? I understand that AI will power tools that may or may not fit a particular use case. How is AI a feature and what does it mean to be anti-AI?
https://gitdot.io/bkdevs/gitdot/files
so take this to mean - 1) no AI copilot in the app and 2) no training on your data nor selling of it.
our take on AI is that we should focus on building tools that help address its limitations; one of the things we're particularly keen on is building stacked diffs into reviews as a primitive, so it's easier to review a large AI-generated (or assisted) PR. (e.g., diff 1 for API changes, diff 2 for backend wireup, diff 3 for front-end changes)
i think to do that, we're going to try and hook into the subscriptions that people already have and are paying for: Claude, Codex, rather than package our own, but some of that is a bit hacky to do.
hope that's clearer and thank you for asking
That sounds more like anti-enshittification.
Please don’t do this.
Charge a fair price, in fact find a fair price and double it.
I don’t want a free GitHub clone, I want one that works.
How about 50$
50 private repos, 50GBs of git LFS storage. Add collaborators for free.
Actually respond to customers. At this level you only need 1000 paying customers to make it worth while for 2 developers.
You'd want something that doesn't allow training on private repos, so ideally not a private company that can change owners and managers and hence policies over time.
It's an interesting business case - open source website, revenue generating with some type of "open" business structure.
Its first repos could be of itself.
I had this idea for enabling pull requests to be in "pipelines" with each request crowd-funded (as an option) to thereby entice action. A small split could return to the main site itself for upkeep.
Or, if we go a step further you pay a licensing fee for Gitdot and self host it.
The cookie jar of AI training data is very hard to resist even if it’s a non profit
I think 50$ is the minimum you could charge and still have responsive customer support.
I could be wrong but I think besides being able to self host for free you can also pay for something where they help you set it up and stuff
Something nifty also that you can do is that instead of them training on your data they enable you to train your own models on it and download the weights.
Definitely agreed its hard to trust anything you cant audit 100%. I think a non profit structured the right way would probably be pretty trustworthy not to do anything egregious like training on your data though. Or some way maybe to make the thing transparent so you know what happens with the data and you can basically audit the company somehow. I dont know if you would need to go as far as encrypted repos, that seems a little too paranoid but that would be a really interesting idea for a company although the target audience is probably limited
No loading animation, but my screen jitters while loading in stuff. My internet speed is fine, so it's a performance/bug issue.
I also did not initially understand the UI, but that'll come as I use it more
my fixation here is to make everything load instant, but that is dependent on server latencies, which right now is admittedly slow as we only have one server in the US.
but thank you for giving it a shot nonetheless!
The first unique characteristic is that it was built in Rust? Why does it matter from a user perspective? I was expecting the first point to be something that would convince me to check it out.
Unless the goal is to find people to collaborate on building the software. I got a bit confused.
Looking good regardless :)
but i will say that the point of our post isn't to really sell anyone here or on anything. we kinda know that our product as-is isn't ready to use at real scale (we lack issues, prs, ci, gotta fix a lot of bugs, etc.)
we did just want to sincerely share what we built. and rust is a part of that, we chose it cause we wanted to learn it and we then quickly found out that we really liked it too.
At least they aren't using five digit years...
Me, on an international team, as learned that flawless english sentences as a metric about as silly as dress codes in business. I'll take my california laidback-ness to your stuffy NYC banker rules.
What an irony that those exact safety guarantees that made it attractive are now a detriment because they make the project smell like AI.
Surely it is at least in part. We're talking about the announcement of a new open source project that isn't ready to use yet which would presumably enjoy people helping make it work for them, and a new startup around that open source project that is likely to be hiring in the future (or even maybe right now? Didn't check).
> Why does it matter from a user perspective
It implies that there was likely a small degree of rigour in the construction... not a large one (very little software has that)... and not a guarantee... but likely more than the abysmally small average in the software world.
You can take these arguments to an unbounded level of depth; someone can be tired of my responses to the responses to so me someone happening to care about something different than them. It's all equally silly, and if you're convinced that your level of depth is reasonable but the others aren't, I'd highly recommend trying to think about why you're so confident that your opinion is important but ones that come after aren't, because it's very unclear to anyone who doesn't already agree with you, and that's usually a sign that there's no point in saying it at all.
Why rewrite from scratch? Wouldn't it be advantageous to start from an existing forge (there are many these days) and add a new UI?
I was gonna call mine EOL and I already bought the domain eol.sh...then again, I could just do mine in TypeScript and launch it anyway.
Most people would also agree that building a better Github is not a super easy two-person task.
> We don't.
This cannot last forever. What's the plan when it runs out?
Did you check if I name my project 朽木?
git itself is decentralised - all repos are equal. Mr T designed it that way because ... well that's all that was needed for Linux kernel development back in the day and it still seems to work. The management stuff can be managed quite well via email and some choice socials. Obviously that nonsense cannot possibly scale to the size of your enterprise thingies!
Yet again we have a better Big Brother than Big Brother ... this time with Rust, yum!
I see code reviews is in the roadmap, I can't wait to try it.
it's very convenient to be sent a link (or find it on hacker news) and to be able to click around files, read the README, understand what a repository is about without having to clone it and open locally. plus -- if you only need a barebones git server with no web UI, git provides this by default.
if the question is, will you build a CLI / what will be in it? the answer is yes, we do have a barebones CLI for auth as of now, do envision things like managing issues / PRs from the CLI, but want to make sure that strike the right balance there.
i think TUIs can be deceptively hard to build well, and admittedly, it hasn't been a priority for us quite yet.
Don't try to make me install a random program if I can view it in my sandboxed browser safely.
Also, browsers have greasemonkey to help me personalize websites easily. TUIs don't.
Loading files is very slow but I assume that's because HN is hammering the server.
I am not a believer in negative advertising. So I don't give a poop you are anti-ai. Or "better" than Github (better for who??). Just imply you are a code forge thats made for serious developers who need something engineered to be fast and reliable.
I wish you the best of luck, I can see Linear coming out with git repos after coming out with a diff reader. I have a suspicion there's space for many code forges in the market as you build out more features, especially if you lean into your products hacker-y-ness
Hmmm... no. Why should I? Just let me use git.
That said though, one of my pet peeves around browsing Github Web from the terminal was having to click "skip to content" just to get the body. So you definitely delivered there (after having read your design post). Good luck with the rest of the year.
yeah, that admittedly is not priority, but thanks for taking the time to check it out.
I bet they will sort it out, and is most likely a top priority.
I do like that it reinforces this rule of thumb:
Tell us why we should care outside of the marketing fluff - these aren’t highlights - if anything they are quite off putting.
Your project needs to stand on its own actual merits.
Critique over, congratulations on launching something or building something anyway.
But what makes this different? And why have you chosen that philosophy - outside of marketing fluff
You seem to be experienced devs, so I’m sure you already know, but don’t listen to the contrarians on HN. They’ll suck the life out of you because it’s not 100% the way they like it.
It’s like “baby on board” stickers on cars - why are you telling me that?
Then again Rust is so complex maybe it’s worth self congratulations.
lol. really stretching the word "better"
If you are YC backed then say "YC Winter 2026" or whatever in your title frankly - people will work it out anyway.
They said it was a better GitHub, which is a very high bar (despite the regular complaints about GitHub). They also said it was anti-AI, despite it being vibe-coded.
Also, know your competitive space. Posting on HN is like pitching - not knowing about your competitors is not really going to work - you need an answer for, for example, why someone picks Gitdot over SourceHut.
I was thinking about creating my own git forge given the unreliability of Github and I wouldn't be able to create just at the moment incredibly reliable software like git forge although I could use AI to create a minimalist piece of software, I didn't because I didn't want to create yet another AI slop fighting another AI slop (github/gitlab).
Forejo is incredible but I have always wanted to get more alternatives in this field.
Much thanks for making it. I have signed up and I have high hopes for it too and I will try to either self host this on my servers or gitdot.io as well as Github
I recommend making a small community in matrix (preferred), fluxer.gg, discord etc. as I'd like to join it.
PS: small personal thing that I have made which helps in making communities: https://mirror.forum
I am definitely interested in gitdot.io! This seems incredible
I wish nothing but the best for you folks. Gonna create a local copy of the source code of gitdot.io right now!
Thanks for open-sourcing the efforts too. I really appreciate it :-D The software is so nice!
I genuinely hope that you guys and the project blows up and if you guys might ever hire a junior dev, I hope you all could remember me as the world right now needed such software that you have made!! :-D (Although I am more interested in managing servers/golang but that's because rust is hard to learn as a beginner but that's different topic but I like rust's ideas too and rust is a great/preferred language with golang for this type of service :-D)
It might be a bit difficult to self-host at the moment (as we don't have a good documentation to do so) but you can try and let us know how it goes.
This particular issue is solved in GitHub proper, and derives from the Windows 95 tree view widget [1], which I seem to remember from Windows 3.x but can't find a screenshot.
The hover behavior is just not an intuitive or accessible default. I can't imagine someone being able to use this if they have a hard time clicking without moving the mouse. It also wastes resources fetching file/directory contents while the user is moving the cursor to a predetermined file they presumably wish to open.
[1]: https://learn.microsoft.com/en-us/windows/win32/winauto/syst...
i think i will get rid of the change on hover, it is a distraction and perhaps was just my ambition to show people how fast we load.
edit: fixed
Personally while I appreciate something not being AI slop, writing something in Rust has no meaning to me.
and yeahhhh, i do try to be very non-marketing in all that i say, but something about the title made me a bit ambitious, apologies.
> mobile support to come
Cmon lol. Give opus 20min and it will give you a mobile site throw in a better-looking desktop site for fun.
If you're sell is being an unenshittified alternative to github, just give it time
Personally if I wanted to leave Github, I would just move over to Gitlab.
If I'm not mistaken about that, you should remedy that to ensure email providers don't dump your emails to spam.
The software was never the moat. The skills and experience crystalised within it was, and remains, the product.
Common man, you're not even 5% of a Github replacement. Don't act like one. You've built a Git web UI with accounts, the easy part.
> Building software is still hard
You don't say.
"Don't be snarky."
"Please respond to the strongest plausible interpretation of what someone says, not a weaker one that's easier to criticize. Assume good faith."
Obviously this is a project at an early stage and the title expresses what they're working on, not a claim to 100% feature parity.
https://news.ycombinator.com/newsguidelines.html
Well actually, spacex owned xai is now coding AI in C while earlier it was rust; I guess with the attitude that a language good enough to control rockets is fine for ai.
In 2026 not being mobile first is a bit of a disappointment to be honest