We've raised $17M to build what comes after Git

The interface is still bad. Teaching people to use git is still obnoxious because it's arcane. It's like 1e AD&D. It does everything it might need to, but it feels like every aspect of it is bespoke.

It's also relatively difficult to make certain corrections. Did you ever accidentally commit something that contains a secret that can't be in the repository? Well, you might want to throw that entire repository away and restore it from a backup before the offending commit because it's so difficult to fix and guarantee that it's not hiding in there somewhere and while also not breaking something else.

It's also taken over 10 years to address the SHA-1 limitation, and it's still not complete. It's a little astonishing that it was written so focused on SHA-1 never being a problem that it's taken this long to keep the same basic design and just allow a different hashing algorithm.

> The old model assumed one person, one branch, one terminal, one linear flow.

That sounds exactly like the pre-git model that git solved..

think about all of the discussion we have around the code that gets lost. we certainly have the ability to keep and link all that stuff now. we don't really need to have arguments about squashing or not, we can just keep the fine grained commits if you really want to dig into them and maybe ask that people write a comprehensive summary of the changes in a patch set -in addition-.

but I guess none of that has anythig to do with AI

From git(1):

    LOW-LEVEL COMMANDS (PLUMBING)
       Although Git includes its own porcelain layer, its low-level commands 
       are sufficient to support development of alternative porcelains. 
       Developers of such porcelains might start by reading about git-update-index(1) and git-read-tree(1).

       The interface (input, output, set of options and the semantics) to these 
       low-level commands are meant to be a lot more stable than Porcelain level 
       commands, because these commands are primarily for scripted use.
       The interface to Porcelain commands on the other hand are subject to change 
       in order to improve the end user experience.

       The following description divides the low-level commands into commands that 
       manipulate objects (in the repository, index, and working tree), commands 
       that interrogate and compare objects, and commands that move objects and 
       references between repositories.

I'm not famous though, I'm just a good engineer who is patient, inquisitive, and determined enough to spend the last five years of my life on nothing but this.

My question is: say the investor believes that some new platform will win out over Github. How do I make the case that it will be mine over a famous person's?

Once open source spreads into an area, it tends to kill (commodify) commercial software in that space.

For example, with databases, MySQL and Postgres "won". Yes, there are commercial databases like SQL Server and Oracle but they largely exist through regulatory capture and inertia. It's highly unlike anyone will ever make a commercial general purpose database again. There are always niche cases.

Same with operating systems. Yes we have MacOS and Windows but what are the odds we get another commercial mass OS? I'd say almost zero.

It's the same for source control. Git "won". There are a handful of others (eg Mercurial). But gone are the days of, say, Visual Source Safe.

But when people talk about "what comes after Git" they really mean (IMHO) "what comes after Github", which is a completely different conversation. Because Github absolutely can be superseded by something better. Will it though? I don't know. It has an incredible amount of inertia.

As for AI and anything related to source control, I'd have a hard time betting against Anthropic. But remember the exit could be an HN post of "We're joining Anthropic!". Side note: I really hate this "we're joining X" framing. No, you took the bag. That's fine. But let's be honest.

For people with a proven track record, AI is a gold rush of acquisition more than creating a sustainable business, let alone an IPO. I think that's what this bet is.

I'm surprised to read that, because that's how I've always used Git (and GitHub).

That's what I've understood to be good practice with Git, and it was liberating compared with what came before. One of the nicest things about Git is you can throw things in locally without worrying about how it looks, and make it presentable later.

I will admit, I didn't know jj but I wanted snapshots so I used it, so then when AI made some changes and kept on going and I wanted to go back to a particular change and I used ai to do that. It was actually really frustrating. To the point that I think I accidentally lost one of the good files within the project and I had to settle on good-enough which I had to try to get for hours to that particular point.

My point feels like I should either learn jj properly to use it or to at this point, just ask AI agents to git commit. Another point but I was using ghostty and I had accidentally clicked on the title bar and somehow moved the folder to desktop, I wasn't thinking the most accurately and I just decided to delete it thinking that it must have copied it rather than moved it. (Also dear ghostty why do you make it so easy to move folders, it isn't the best of features and can lead to some honest errors)

My face when I realized that I have deleted the project:

Anyhow decided to restore it with ~/Trash but afterwards realized that the .git/.jj history is removed because it deletes hidden folders (from my understanding) so I definitely lost that good snapshot. I do have the binary of the app which worked good but not the source code of it which is a bit frustrating

These were all just an idea of prototyping/checking how far I can move things with AI. Yeah so my experience for that project has been that I could've even learnt a new language (Odin) and the raylib project to fix that one specific bug in lower time than AI which simply is unable to fix the bug without blowing the whole project in foot.

I think the takeaway is to have good backups man. I mean I was being reckless in this project because I had nothing to lose and was just experimenting but there have been cases where people have lost databases in prod. So even backups should be essential if you find any source code which is good to be honest.

I am sure you guys must have lost some source code accidentally which you have worked upon, would love to hear some horror stories to hopefully know that I haven't been the only one who has done some mistake and to also learn something new from these stories. (I am atleast happy in the sense that I learnt the lesson from just an tinkering thing and not something truly prod)

Vibecoding moto.

I've not used this app, but I wonder how tooling like this truly competes against an open source community armed with AI. Like where is the moat here, really? I built a personal tool that does some of this with a basic Claude subscription over the course of a few weeks.

Feels like vibe-coders are the real target market for something like this, but if it takes off, would not be that hard to clone as a FOSS app.

It's not difficult to "escape" - using `git checkout` will tear everything down properly - that's the only task of the `post-checkout` - to determine that you want to go back to using vanilla git commit tooling and remove our shims.

We also don't have a prepare-commit-msg hook - our commit tooling will inject an extra Change-Id header (of the same format and interchangeable with Jujutsu) but that affects nothing that vanilla git cares about.

In any case, Git has become tremendously entrenched over the past couple decades. Anything that hopes to replace it would have to be significantly better to break from the inertia Git has. I’m honestly skeptical as to whether this is even possible in the near future. We’re not at all in the same historical moment as when SVN was beaten out.

1 - https://openai.com/index/openai-to-acquire-astral/

They'll start injecting ads in your commit messages, forcing you to subscribe to a premium plan.

To build better tool than git, probably a few months by tiny team of good developers. Just thinking of problem and making what is needed... So either free time or few hundred thousand at max.

On other hand to replace GitHub. Endless millions will be spend... For some sort of probable gains? It might even make money in long run... But goal is probably to flip it.

but if not just your own work flow, have a dir dedicated to storing prompt history and then each file is titled with the commit id.

As for the flag just agree to some convention and toss it in the commit message

Keep an eye on our blog to see how we're doing this, and how we're doing it in a way that hopefully the entire community joins us in a way where we're not all reinventing the same wheels.

Only useful if it can be reliably verified, which is challenging at best.

The point of git is that it has strong authentication built into the fabric of the thing.

Yes, it could have syntax like

    git notes add -m "Claude prompt: foo fee faa foo" <commit-hash>

OH WAIT YOU CAN DO THAT ALREADY SINCE 2009

Seriously, the 90% complaints about git not being able to do something is just either RTFM or "well, it can, but could use some better porcelain to present to user"

Are you interested in giving https://tangled.org a try? I'd love to hear your thoughts!

Yeah, that is also my take. I'm biased of course since I'm someone working on replacing git through grassroots iteration, but I've been around this block a few times though and I never saw blasting money at a problem produce real innovation.

Wealthy people don't have time to do all due diligence and vetting specially when random startups become unicorn.

git ≠ GitHub

Does AI make reading or writing stacked PRs any nicer? No, it does not.

1) because they can

2) it's their money, not company money, and again, why would you risk your own money when someone else wants to risk their money?

If it's to enable multi-agent scenarios, don't worktrees (at least in the local sense) allow for this?

4. GitButler is a terrible name for this

5. No one will use the "but" command over "git"

6. The founder needs to learn to enunciate the name of his new product better

And also, your central premise is exactly right. The solution to agents and humans working faster will not be better manual oversight of what they're doing. It's like missing the most important principle of agentic development. Supervise, don't gatekeep.

GitButler came about many years ago because I have been using Git for almost the full 20 years of it being around and I thought there could be a better way to do the things it's trying to solve for us. I want version control to do more for us, easier, faster and smarter. Git is still pretty dumb. Plus, now, everything in the dev workflow is changing - it's an interesting problem to think about what a _great_ toolset for how we'll all soon be developing software will be.

As a _single_ example - agents can't use interactive editors, yet _so many_ of the powerful parts of Git absolutely _require_ it. Agents can't interactively rebase, meaning they can't very effectively squash, amend, reword, reorder, absorb. They can't very easily interactively add. They are middling at best when it comes to stacking branches. Git is designed to send patches over email and agents are not concerned with that.

I would love to debate all of the lessons learned about the history of Git, but I was around for all of that. I know why Git was started, I know what it was meant to do, I understand how it's evolved. I still think there are a lot of interesting things that we could have in our change control tooling and Git is not the perfect solution.

Nearly everyone in this thread suffers from the same basic local maxima blindness that you do. Git is great, GitHub made it more valuable. But maybe the answer to the papercuts we've constantly been dealing with for decades isn't faster horses. (To, you know, mix metaphors)

Also if they really wanted to “replace git” I think that would be much more difficult due to network effects. Everybody is already using git.

What I'd would expect of the next vcs is to go beyond vcs of the files, but of the environment so works on my machine™ and configuring your git-hooks and CI becomes a thing of the past.

Do we need an LSP-like abstraction for environments and build systems instead of yet another definitive build system? IDK, my solution so far is sticking to nix, x86_64, and ignoring Windows and Mac, which is obviously not good enough for like 90%+ of devs.

It’s 17m for a tool which hopes to serve companies and charge money and make more than 17m in profit as a result.

If you look at the set of dev tooling, teams will frequently pay many hundreds per dev on things like CI, Git tools, code review, etc.

And to be fair, GitHub is really quite bad for a lot of workflows. I haven’t used gitbutler, but my team pays ~$30 a month per dev for tools which literally just provide a nicer interface for stacking PRs, because it saves us WAY more than that in time.

This isn’t even an egregious example of VC, it’s just an enterprise dev tooling bet.

Use value != sales value; hype sells.

Ps. not too sure how far $17M gets you toward mini nuclear power plants, but I catch your drift.

What’s the problem?

Do you think less money should be going into VC?

Just some numbers ~1.5M housing units are built in the US with an approx cost of $300k - $400k. That is $450B to $600B going into housing units construction every year.

On the other hand VC has maybe $1T AUM in the US. Maybe 10%-20% of that is deployed every year? So $100b to $200B.

What is wrong with that ratio? Could there be better solutions to make more housing cheaper? (lower regulations, efficient permitting, etc)

Money moving from VC to housing seems without a first principled approach on what problem your solving and how is silly.

There's gobs of amazing technology being built by people who just love to build, have great ideas, and huge talent (now exponentially compounded by LLM assistance, even) -- and 99% of it is ignored by people with $$ and none of them will be paid to work on these things -- let alone get funded to build a business around them -- and the reason isn't the inadequacy of the technology or "lack of a workable business plan": it's lack of social connections or pedigree.

And what this tells me is two things

1. there's a fundamentally sickness to the VC culture coming out of Silicon Valley and it's gotten worse not better with the new restraints in the post-ZIRP era. It's an echo chamber and a social circle, not a means for creating new profitable companies or good infrastructure, and it serves mainly just to feed a pipeline of acquisitions into much bigger fish rather than building tomorrow's new businesses or ideas. This is very different from 80s, 90s tech culture that I grew up in.

2. there's clearly a desperate need for more actual incubators or labs for actual technology, paying people to build "good stuff" independent of the vagaries of what VCs and their ivy league friends are able to pitch.

Frankly: The $$ out there in heavy circulation has been mostly corrosive, not helpful.

Yes, we have higher taxes, yes, we pay more in social security... but in the end we have far less "Working Poor" and I know very, very, very, very few people who have more than 1 job.

But I guess that's just socialist bullshit.

What I am trying to convey is: The US lives in its own bubble, just like the rest of us does.

The difference is that the US hears the US propaganda and the rest of us heard the US propaganda for decades as well, through Hollywood and media.

It reminds me how the Bohemian Club’s slogan, “Weaving Spiders Come Not Here” is a bit farcical given that it is impossible for the club members not to engage in commerce.

You never sort by color, ever! You sort by form, and then throw every color of that specific form in one bin. If you throw every red brick in the same bin, you'll never find a specific formed red brick because to many red bricks. But if you first search by form and then by color, you are much faster.

Because solving problems isn’t the goal, the goal is money (and sometimes a little fame) with the least possible effort, and software can be changed on a whim and is very cheap to manufacture and distribute and “fix in flight”, it’s the perfect vehicle for those who are impatient and don’t really care about understanding and studying a need.

People complaining about investors throwing stupid sums of money at stupid or trivial things unrelated (or only marginally related) to AI? ...sounds to me like the first glimpse of hope I have come across in this industry for half a decade.

[0] https://pickupbricks.com

[1] https://www.youtube.com/watch?v=X25MIpQqLIU

The world doesn’t need this. It would just be more plastic and electronic trash.

You and your kids have hands. Pick them up. It’s what we do in my house.

If you don’t have hands, use your feet.

Perhaps you should have. Based on the link it seems like it's more an extension to than replacement for Git.

The page is mostly sort of fluffy AI hype, but the concrete bits are things like integrating issue tracking and PR logic in one tool/repo, like e.g. fossil does.

Also git proper could use some love too. The UI is still a mess. And the large file support and the submodule/subtree/subrepo situations are quite dismal.

> $17M, one could probably build a vacuum robot prototype that’ll also clean up all of the kids toys and sort LEGO bricks by colour and size.

Doing this robustly is probably quite far from robotics SOTA.

I do wonder, though, if it would have been designed differently if the whole “code forge” sort of application (or whatever GitHub and the like are called) was envisioned at the time. Pull requests aren’t even a concept in git proper, right?

It seems like a kind of important type of tool. Even though git is awesome, we don’t need a monoculture.

That said, if you ever decide solve the tidying the toys problem, start a kickstarter, I pledge to pledge support! :D

And what's the next step? I can't even imagine how rich (and how large the their houses) the parents need to be for them to comfortably buy such dedicated tool. Perhaps 100x~1000x richer than me?

And, while this is just pulled out from my rear side, I feel even getting this passed safety regulation would cost your $17M. It's a fully automated machine working next to toddlers!

On the contrary Github is a proven product.

I think if you have a healthy busy growing well, you shouldnt raise unless you have ambition and urge to go faster.

Irony of thr market is, just like tinder 20% of the companies attract all the attention rest of them try to gran the attention. Those who need capital get the capital, those who need the capital die trying.

Enough friday pessimisim.

I had a few interactions with VCs (both professional and personal), where I didn't care because I wasn't benefitting from them. One of them was "an expert in CRISPR and blockchain" (WTF?) and... well I didn't need much time to see that he did not understand what a "hash" was. He was mostly an expert at repeating stories he had been told about how he would make a ton of money with the latest bullshit he didn't understand.

The truth is, it's like trading. You diversify the investments and hope that the economy goes up (respectively that one of the startups you invested in gets profitable). The only thing a VC has to do is verify that they don't invest in a fraud, but even that is hard given that they never understand the technology enough to say it's worth it (they often invest in shiny bullshit).

While I personally doubt that for $17M one could build such a vacuum robot prototype (for a vacuum cleaner company, investing this amount of money - if it worked - would be a rounding error), I will rather analyze the point that you raised:

It is a very common situation that the workflows of companies is deeply ingrained into some tool

- that they can't get rid of (be it Microsoft Excel (in insurance and finance), be it Git (in software development), ...)

- that is actually a bad fit for the workflow step (Git and Excel often are)

So, this is typical for the kind of problem that companies in sectors in which billions of $/€ are moved do have.

I am actually paid to develop some specialized software for some specialized industrial sector that solves a very specific problem.

So, in my experience the reason why nobody [is] solving actual problems (in the sense of your definition) anymore is simple:

- nobody is willing to pay big money for a solution,

- those entities who are willing to pay big money often fall for sycophantic scammers/consultants.

Because that’s too risky for investors.

Thought so until saw this. Man, he is the co-founder of Github and already seed-funded. How can someone refuse him? 17M is a small amount considering the valuation VS Code Agent wrappers are getting

Just write down how you'll spend the money to make that, what it'll eventually cost to produce, what the market size will be, and what the price will be, and if it's enough return you can easily convince someone to give you $17m to do it all.

We've strayed really far from where technical innovation began

Well, cofounding Github helps

Consider that many of the tech posts here are of the form, "i did X but with Z" as the poster hopes they will be recognized as some master of execution.

This seems ridiculous to you, compared to a very obvious win with a Lego sorting vacuum.

Lego isn’t niche, and the explanation isn’t a weird technical thing that only experts would get and understand how important or valuable it is.

Yet it’s not being done.

Is there nobody who has realised this gap but you? Has nobody managed to convince people with money that it’s worthwhile? Have you tried but failed?

Or is it not many many thousands of people who are wrong but you?

Is the problem harder than you think? I’ve worked with robotics but not for a long time and I think the core manipulation is either not really solved or not until recently. I don’t know about yours but my kids also don’t fully dismantle their Lego creations either so would the robot need to take them apart too? That’s a lot of force. And some are special.

How people want Lego sorted is pretty broad. Kids don’t even need it sorted that much. And the volume can be huge for smaller buckets of things.

Is the market not as big as you think? Is it big enough for the cost, I’d buy one for £100 but £1000? £10,000?

How does it compare for most people against having the kids play on a blanket and then tipping it into a bucket? Or those ones that are a circle of cloth with a drawstring so it’s a play area and storage all in one? I 3d printed some sieves and that’s most of the issue right there done.

People are solving actual problems, but lots of problems are hard, and not all of them are profitable.

As a gut feeling, there is such a large overlap of engineers and large Lego collections and willingness to spend lots of money and time saving some time sorting Lego that the small number of implementations usually split over many years is very telling about the difficulty.

For what it’s worth I want this too.

Why are we trying to replace git? What is the problem with git?

I fought for years trying to convince my colleagues to write good commit messages. Now Claude is writing great commit messages but since I'm no longer looking at code - I never see them. I don't think Claude uses them either.

Branches are now irrelevant since all agents work in worktrees by default. But worktrees are awkward since you run out of disk space fast (since we're in a monorepo).

There is a constant discussion ongoing whether we commit our plans or not. Some argue that the whole conversation leading up to the PR should be included (stupid imo).

The game changed completely. It isn't weird that people are wondering if the tools should as well.

Definitely feels like there's opportunity to build something better

They went over this, in the documentary titled "Idiocracy".

https://www.youtube.com/watch?v=GFRzIOna2oQ

https://www.amazon.co.uk/Toy-Storage-Organizer-Lego-Play/dp/...

For example, instead of building a robot to pick up Lego bricks, say you’re building a platform for personal robotics, and it’ll cook you food, do your laundry, repair your fridge. It doesn’t matter if you have any idea how to do this, just say you need $50M and you’ll hire some robotics and vision guys to figure it out. The bigger and bolder the lie, the better.

Solving actual problems are hard, and even harder to get money for (see research). Most VC’s are in it for the returns only, not actually making a change, there are some exceptions but they are far and few apart.

I just looked into this out of idle curiosity, after watching some guy build a LEGO sorting machine. (They work in a warehouse that sells used bricks for model builders.)

Interestingly, this is on the cusp of viability, but training the ML model would still be cost-prohibitive (for me). With $17M, it's within reach, but there's still the obvious mechanical hurdles: Kids don't disassemble their Lego, the conditions are "less than ideal", and even vibrating belts in a warehouse scenario have a lot of trouble keeping bricks separated for the camera to get a clear image.

Robot hands are nowhere near the point where they can reliably (or even unreliably!) take apart two arbitrary Lego bricks that are joined, let alone anything of even mild complexity. This is hard for most humans, and often requires the use of tools! See: https://www.lego.com/en-us/service/help-topics/article/lego-...

The machine vision part is... getting there! You could pull some clever tricks with modern hardware such as bright LED lights, multi-spectral or even hyper-spectral sensors, etc. The algorithms have improved a lot also. Early attempts could only recognise a few dozen distinct shapes, and the most recent models a few hundred, but they're about 2-3 years old, which means "stone ages".

A trick several Lego recognition model training runs used was to photo realistically render 3D models of bricks in random orientations and every possible color, which is far faster than manually labelling photos of real bricks.

These days you could use the NVIDIA Omniverse libraries to heavily accelerate and automate this.

I mean who tf gives some small team millions to put some Nvidia GPU into space and thinking we will have market disrupting GPU clusters in space in 10 years?!

There are so many low hanging fruits in IT Industry to just being solved.

Even just having something like well build, open smart home products whould have been disruptive years ago (until someone like ikea decides to enter that space).

So, even though Git seems to be ok (people who store large binary files or who run huge monorepos would probably disagree), maybe we can do better.

Altavista was kind of okeish for search, yet Google managed to figure out something that was (at that time) way better.

All you need is a camera pointing at the floor with image detection... when there's legos on the floor it triggers a video playing that explains how the kids need to pick up the legos. /s

Let me just state the obvious. Of all the major problems of society, sorting legos isn't one. If you disagree, try emerging from the cellar.