Kind of depends on the governance, which i am mot familiar with. Would be a shame for him to lose control when he has done a decent job so far. Transferring governance has turned into a woke mess enough times to be concerning.

[flagged]

These tools all assume you have machines to run the agents on. But for parallel agents I'm pretty convinced you want each agent on its own isolated devbox running your dev environment (not e.g. worktrees on one box) - which isn't trivial to set up and manage.

I'm working this with https://boxes.dev - a workspace for launching and managing claude + codex sessions, each running in its own cloud devbox.

We launched with a desktop app but have gotten a lot of pull for terminal-driven workflows, so have a TUI as well. It's interesting - when the Codex desktop app launched I was so sure that GUI is the future for coding agents (it's just a chat!). But turns out there's no killing the CLI

I've been working on collie[1] a plugin for herdr that runs a web app on the herdr host which can then be served via tailscale for example. This has made accessing my agents on the go a lot easier, paired with session persistence from herdr it's been quite nice to work with.

[1] https://github.com/AltanS/collie

For now I didn't find anything that really works better than using termux on android or echo on ipad, ssh into my desktop with tailscale where I have a tmux session I can attach to. "Input" seems to be the biggest struggle for me versus a normal desktop+keyboard setup. On ipad I use a nuphy bluetooth keyboard which works pretty well and on android I use FUTO Keyboard with voice input.

Overall I think I don't even want to start coding too much on the go with a setup like this but it's nice to have options :)

I don't understand herdr, can you explain it to me? I've tried it, but I don't understand what it gives me over tmux other than the agent sidebar. But people seem excited about it so I feel like I am missing something.

While not terminal-only, I've been using Orca (https://github.com/stablyai/orca) for the past week and it's been a great experience

Since you seem curious about these types of tools, might I interest you in ouijit https://ouijit.com/ https://github.com/ouijit/ouijit

Its intended as a terminal agent orchestrator aimed at power users who want a little more design than what I like to call 'hackerman' UXes.

I’m tracking them all here[0]

The first app in this space was VibeTunnel[1]. It basically lets you start terminals on your Mac and the web app allows you to control these terminals on your iPhone using Tailscale. In the end, it’s just terminals on the web with organization around it.

I forked VibeTunnel and turned it into a Next app and added workspaces with memory and turned it into a software factory which is making open-source alternatives for me while I sleep. I call it VibeGhost and will be launching it soon[2].

0. https://opensource.builders/alternatives/conductor

1. https://github.com/amantus-ai/vibetunnel

2. https://www.vibeghost.org

Pardon the plug, but I thought I'd point out that I think SciCommander [1], or the ideas behind it, could be very relevant for this stack.

It does capture a provenance graph for any ad-hoc shell commands executed in the shell mode, or if prepended by `sci run`, by tracing all new files created from commands, and creating an accompanying json file for every output with metadata, which can later be assembled into a graph specific to any output file, using the `sci tohtml` or `sci toshell` commands (producing an HTML report with an SVG graph, or a reproducing shell script, respecively).

I'm quite bullish of the possibilities with this approach.

And, this was in fact also born out of the thinking to "rip apart" workflow tools, and build them up again using small, well-defined tools that do one thing well.

[1] https://github.com/samuell/scicommander

I was thinking cmux

have you tried Orca at all? I'm wondering between switching to herdr from it

thank you herdr.dev ! it is awesome.

[dead]

The actual header on the page is

> We are building the multiplexer for all work.

imo this would have been a better way to do the HN submission title (with or without [terminal])

> Superlogical: We are building the [terminal] multiplexer for all work

One of the worst I've seen was title editorialized from original "Someone: something [x.com]" to just "Something [x.com]".

Without clicking you’d never know why it’s the title.

At least there's no history of security issues with that.

They can at least deliver baseline value, because it's all text-based. Do agree that it'll become more complex, when/if they decide to create custom protocols.

I think you may have meant to comment on another post :)

I recommend 'ssh -o PubkeyAuthentication=no superlogical.jobs` if you have a public key and want to stay anonymous (beyond revealing your IP).

Why? Because if you use SSH to connect to connect to GitHub, your SSH public key is published at https://github.com/<username>.keys so an SSH server to which you connect could do a reverse lookup.

This is the reason why my ~/.ssh/config has those 2 lines at the end:

    Host *
    PubkeyAuthentication no

NSA and other companies historically had you jumping hoops and writing mini-exploits to be able to see the email you're supposed to send your resume to.

The ssh thing is really fun, but I wish they had another way to view the openings. I'm on a connection with massive latency at the moment and it's pretty horrible UX in this case

In case others were interested, they are not a remote-first company and require you to live in LA, NY or London.

You obviously should connect from Ghostty.

But try also `TERM=dumb ssh superlogical.jobs` for an alternate, monochrome, experience.

Undocumented keys h,j,k,l also work to navigate the UI.

Any other easter eggs?

I must admit that I paused before hitting enter on that command and thought to myself, "Maybe I should download Ghostty and open it with that as a virtue signal." :-)

It’s probably a good filter too for applicants

Better be using ghostty ;)

They're using Go, Swift, and Zig: https://x.com/mitchellh/status/2082623830510710865?s=20

They hire for Go.

Considering it's going to consume libghostty, I'm betting it'll be Zig.

There is PR somewhere in the mosh repo to support block cursor. Just build it manually. Mosh is abandoned, unfortunately.

Right, but I'm pretty sure this is not the first time somebody thought about solving it.

https://xkcd.com/927/

Yes. I find it quite funny that 4 people with the flashiest of titles (head of experience?) came up with this amazing plan:

1. it will be incredible

2. it will be composable

3. it will be safe to use in production

I would totally invest millions into these guys, they have a bright future ahead.

I still don't understand why tmux authors wasted their time on tmux, tho, when we already had perfectly servicable GNU screen.

Why reinvent the wheel?

/s

Another example of the galaxy brain people who like employees to go to an office but then are happy to have offices all over the world, some of which have 8 hours of timezone difference.

Are they the champions of collaboration or the champions of company culture?

Funny how just 3-4 years ago on this forum people were openly prognosticating about how the era of fully remote work was here to stay and the centre of gravity of investment no longer had to be SV and YCombinator was running fully remote sessions, etc. etc.

And here we now have people saying stuff like this again.

Also, long live the "lifestyle business" if it means people just making businesses that bring in non-exponential but justifiable and sustainable income ... like, y'know, a business is supposed to. Good grief.

Or you want to hire the kind of seasoned veterans who want to live in those cities. Often later-career, families, looking for exposure to more than just a thriving tech scene.

Like notable lifestyle businesses Datadog, Squarespace, Etsy, Tinder, and Snap.

Its interesting they recognize they've themselves worked remotely for most of their careers but not for our new company. Maybe its a new trend

I don’t think that’s a parsimonious explanation in a post-COVID world. It seems equally if not more likely that talent is more distributed, and people find fewer reasons to relocate into the Bay Area (in terms of career progression, etc.)

(This is not, in my mind, either a good or a bad thing.)

What do you mean by lifestyle business?

At the risk of sounding like a booster, I have noodled over whether this is a place that AI can legitimately provide some improvements. It's harder than just waving an AI at the problem, which is why we don't really have this yet. But one of the major issues with the intermediary language you're talking about is the rigidity of conventionally-applied programming. AI is a bit more flexible and gooey and more able to function as duct tape and bailing wire... but that does come with several tradeoffs of its own that need to be mitigated to some degree.

As a specific example of this, I've been using emacs since about 1997, but it's only this year that I've started down this sort of "integrate everything" road that it promises, because I've never wanted to learn emacs' specific Lisp variant that is used nowhere else. But now with AI I've done more emacs extensions that are providing real value to me than I have in the last... ever. About my previous limit was reifying a keyboard macro into a command, which still isn't really programming emacs.

The ability to indicate "intent" in a way that computers can actually process it as "intent" might change the calculus enough to make this more possible than it used to be in the past.

But those tradeoffs are real. Non-determinism is a problem. Difficulty in knowing what the AI is doing unless you're staring at its output, and even the ability to conceivably monitor an AI's output is only a transient thing we have at the moment, they'll speed up past our ability to monitor them at all soon enough. But it's at least a new area that can be explored and there may be solutions to some problems in there we couldn't reach before.

Looking really cynically, it does seem like it might transition into some kind of cloud-development situation. Your "one access to all development" is one ssh away from our mega 24/7 online compute farm, since we need to keep the sessions persistent. But you can access anything you want!

Hopefully that's not the endgame?

There is value in abstracting the underlying generic systems into more context-relevant primitives for people who just want to get shit done in that context.

Why not meet people where they are at?

If this product isn't valuable, then it will fail. I suspect it will not fail.

Scrollbar is fine for me? Just styled

This might me tmux alternative he had talked about before? Maybe a part of the superlogical ecosystem, but just like libghostty is?

Their intention is to let you live inside, well, Superlogical which can run on your Mac or Windows desktop, and, eventually, on your phone, too.

I still regularly need to edit text files and use nvim for that. Sometimes I even use it to edit a prompt. I like to prompt my system by writing .md files to help get through the planning phase, and then I like to edit the generated todo list.

The agents have to live somewhere. They address your question in the post.

[flagged]

I think IDEs took off plenty well, but several got frozen in time.

Visual Studio as the IDE for building windows client apps, especially WinForms, was and still is a genuinely incredible and productive experience. The ease of which you could click-and-drag a UI, context switch to logic in the code behind the forms, wire up a web service backend, share domain models and have navigation and search tools understand all of it was great. And the TFS integration was quite good too. You really only had to leave VS to do database work or administer some IIS pools.

Unfortunately, it couldn't keep up with the rest of the dev world. "Server Explorer" morphed to include your Azure VMs but froze for 30s each time you opened it. Docker tools came in super late, most Azure services didn't really have a VS tooling experience. The "right click publish" workflow meant for small systems never figured out how to graduate to real-world needs across many teams. Developers had to learn how to use many other tools for modern stuff, and so VS became far less of a center of gravity, often to their annoyance.

I think they definitely have the IDE vision in mind and are trying to move off of terminals, based on two parts of the original article

> so you can close the application, reconnect from another device, and pick up exactly where you left off.

I think this is the remote IDE development dream that no one could quite get right, partly because there wasn't enough reason to run things on a 24/7 device. I think with coding agents that can run in the background this changes.

> Sessions can be accessed through the web and native macOS/iOS applications, and sharing a live session with other people is built in from the start.

I'd really like to point out that "Sharing a live session" is horrible UX for a terminal multiplexer because the terminal has to be the same size (# of rows/cols) on all viewers' screens, and some of these viewers may be on a phone and others on a laptop. Believe it, I tried myself and there's no good way around the problem.

What is interesting is that they added this sentence anyway. This tells me that their long-term endgame is not the terminal. They're most likely trying to build a GUI-based agent IDE and then eventually expand to other work (per their headline of being the "multiplexer for all work") as it all gets more and more agent-based anyways. Well, they're not alone in having that vision.

There are plenty industry-specific niche IDEs that you probably never heard of that make up a significant part of many companies product offering.

Jetbrains is valued around $7bil and their business is almost exclusively in IDEs.

Wolfram Research is valued probably around $100mil and their business also mostly consists of an IDE product.

---

With AI, I think the future of IDEs will be very split. A lot less work will happen in IDEs (in terms of LOC), but at the same time, I think AI will allow for a more sophisticated build-out of IDEs, so e.g. debug experiences will become a lot richer.

> IDEs never really took off, > still basically just text editors with a file explorer

JetBrains would disagree, as would I.

Nevertheless I agree this is the next _evolution_ in IDEs.

Somewhat related, I recently heard of fentas/atty from:

https://erock.prose.sh/posix-shell-is-all-you-need

You are missing out a bit, but stay content. :)

Multiplexers have a lot of little papercuts and limitations when you want consistent and native-to-the-OS behavior to flow all the way between the OS and the apps and connections in the multiplexers. For example, gesture scrolling in a tmux/zellij pane is hard because the right events don't flow through. It's an itch to scratch like designing a good terminal lib (like ghostty.)

And then, there's all the unifying and smoothing out of data connections--I imagine they'll be a lot to make sure that, e.g., an ssh session or a Claude code session in a pane can be seamlessly reconnected or never drop when the pane is closed. This takes a lot of hacking to make look seamless in most setups right now.

Like mitchell, I write a lot of software around multiplexers, and I've thought about making something similar, but I don't have quite the same level of personal wealth or angel investor connections. :)

herdr and mosh are a couple of reference points for this, there are others.

I've wondered this myself. I stuck with iTerm2 for a really long time but recently switched to Ghostty. Ghostty definitely seems to renders faster (I say "seems to" because I didn't benchmark or anything).

I don't actually use a multiplexer. I'm not sure I understand the use case when most terminal emulators supports tabs + windows (like iTerm2 does). Or you could just launch multiple windows and use your OS's window manager to "multiplex". Are the terminals connected in some way that I'm missing?

Regardless, I think Ghostty is a reminder that you can win in a space that's crowded. There's probably a lot of widely used software that would get dropped en masse immediately if someone had the guts to make a better alternative.

I don't get it either. What's wrong with Terminal and tmux, screen or what have you?

I don't get the whole GPU rendering thing for ghostty either. Its a terminal emulator. It shows plain text. Why would I need GPU acceleration? I never felt "yeah, this terminal needs faster rendering".

Maybe I'm just an idiot

I imagine something more like herdr, but done at the terminal level, not inside. But that’s cmux…

If you like terminals and tmux, can I suggest you spend a few hours playing with herdr? It clicked for me in a way that a lot of other orchestrators didn't.

Concretely, Emacs/elisp isn't particularly great for handling the newer terminal protocols that allow things like Shift+Enter, Ctrl+Space, etc., along with high colour support or better handling of the alt/meta/⌥ key. Hence why the flurry of new terminal emulators.

The other problem is it does have to be hosted somewhere, and if you run it locally, it goes away when you close your laptop. If you run it remotely, you have the latency to the remote site. I suspect Superlogical is going to find a way to fix this problem.

If you can find a way to stop Emacs from being painfully single threaded, yeah.

The biggest strength to Tmux that a lot of the alternatives lack is you could make a shell script to setup your tmux instance(s) exactly as you would want, so by the time you load into it, its got all the tabs, split terminals, etc where you want them.

This is like the “why not use FTP yourself” response to the YC Dropbox page.

Yes, but good, standardized developer tooling is valuable.

Postman could’ve also been a “just use CURL and dump to JSON” for example.

I actually am bullish on this. Tmux and ssh are somewhat clunky and probably under utilized by most people. I struggle to find commands, navigate sessions, see history, and so on.

Also in an “agent” based world where you might ask an agent to execute things for you, how do you manage that context well beyond just giving it its own Tmux session to drive?

It’s a good idea, and I’m surprised no one’s built it sooner.

Sounds like you're using LLMs correctly IMO. The entire appeal of LLMs in dev-tooling is being able to setup hyper-specific, customized workflows and have "extensions" be done by telling the LLM what to fix. For that, the smaller and more minimal the base is, the better.

Care to share your setup? I'm currently working on pretty much the same thing and it's looking good so far, but having a single user makes it hard to get enough feedback.

The server being their server that you eventually have to a pay a subscription to use? Hmm..

Also, interesting permission problems. Are you going to allow a remote server owned by someone else have shell access to all of your computers? Are they allowed to train an LLM from all the things you type?

Sounds super interesting!

What's your approach?

Can you give me one concrete use case that explains how this is an incredible idea? Thanks

Well neither of those is backed by libghostty. They also aren't native. This multiplexer is a native application on macOS. Hopefully we see a native port for Windows and Linux too.

Zmx allows for similar features natively

Was not familiar with the guy and this is certainly a take:

https://en.wikipedia.org/wiki/Tobias_L%C3%BCtke#Voting_right...

> LLMs have destroyed problem solving, creative thinking in general.

People probably had just as many uncreative ideas before, but lacked the ability to execute on them. Remember the "idea guys" constantly looking for tech cofounders to build their "brilliant" visions? Now, LLMs have given them the ability to execute.

"Any sufficiently VC-funded AI startup will inevitably try to re-invent a log.md file" - Spongebob Squarepants

I've never heard of that Superset. I've used the Apache business analytics one though: https://github.com/apache/superset. Unfortunate name collision...

I have a lot of respect for Mitchellh but taking VC money for a tmux clone is wild.

As long as you are rich all the VCs will throw money at you.. fine.. but if you are rich then you don't have to take the VC money either. Just do it out of passion.

My take is that it will be tmux as a service. You will have many different sessions which will have their data stored in some central system (read: raft/etcd) and an API built on top of that for create/list/attach to sessions.

Ghostty is the client side aspect of this (the binary that actually runs on your laptop), but there will be a large backend system that the binary talks to through an API. With the API you can then also start automating sessions, for example when a metrics value trips an alert start an agent to triage with a session you can then attach to for further interrogation.

What Im not sure of is if they are going to build on top of kubernetes or not, which already provides a lot of the features necessary (etcd store, API server, workload orchestration, tenant isolation, extendable API).

(This is all speculation)

What does tmux alone do way better than the newer tools?

The ease of implementation provided by AI tools makes this "work repetition" more likely. Humans are awful at coordination.

[dead]

Would you pay for it to not have data collections and ads?

One of the cofounders is formerly from poolside (poolside.ai), where they produce open-weight coding models. One of their recent models, Laguna, got quite a bit of attention here on HN.

The one they’re currently building. This has to be an agent orchestration and management play.

It's a multiplexer w/ distributed state. If you know about tmux then it's basically tmux but someone else can connect to the same session w/o logging on to the same computer. I can imagine it being useful for debugging production outages b/c several people can use a single "dashboard" for coordinating their work w/o having to go back & forth on slack/zoom/etc. You can ask any chat bot to explain further if you want to look into it: https://share.google/aimode/XadkUVcvpjlYhuWmK

I came to say something similar. On mobile it looks really nice and matches the highlights elsewhere. Cool subtle visual effect.

So, at least 2 of some dudes.

call me an optimist, but I believe that more and more things are getting more viable to do from the terminal. tools like jj remove the need in Git GUI because mental model and UX gets streamlined, and the AI craze incentivizes practically every company to create their own CLI

The post itself says that the effort is independent of ghostty. Their first work will utilize libghostty, and they'll contribute upstream.

In other words, this probably won't directly involve changes to ghostty.

I'm curious from the other direction; where do you see overlap between k8s and what's described on the Superlogical website?

This sounds like given a coherent structure to “I want the prod code here to change the retry policy so pending retries are held in memory in this edge layer instead of the DB writers so load can be shed more cheaply” and it will usher things in from coding, testing, filing tickets, getting prod approvals, using the deploy tools to respond the containers, smoke testing the deploy, looking at Datadog or whatever to ensure the change is having the desired effect. All in one context, the tools and dev/qa/prod systems naturally being orchestrated within the same environment/ session. K8s is just the next to last 100 meters.

Me literally searching for emacs in the webpage.

"With most of them being VC or high-count angel investors, they can combined push whatever product is being built here into 200+ companies independent of whether it is good or not"

Ah, I see you have figured it out the thing that has always been the case.

I think the generous take would be...

- it demonstrates they have funding (pre-empting "no-one will fund something weird like this")

- its transparent and open (pre-empting "which bastards are the ones who really own it?")

> why else would a ~billionaire that could easily fund it himself pull in that many investors that early?

Investors can be good for funding but they can also be good for education. Later, they're also very good for strategy.

> What happened to the good old way of figuring out product market fit first? Is that the only way a product can succeed nowadays

Generally it's very challenging. I have two products I'm building this way and you face a lot of up-hill climb where people with VC backing are getting the door opened for them. Of course, because of those same pressures to deliver and succeed, it's equally possible those same doors will eventually get closed on them.

I think Mitchell also lives in a bit of a bubble now. He went from friendly open source guy to very rich friendly open source guy. That's no knock on him, but I would consider viewing business from the perspective of someone who has been a CEO, CTO, and IC at the same company (his company) as his only job who became insanely wealthy as a result. I have a friend who has largely only worked in big tech, and successful startups, who tells me all the time level and salary don't matter to him. They don't matter to him because he found success early. Again, not a knock but sustained success will shape how you view the world and how you play the game.

> I'm excited to see what Superlogical will build, but the startup scene seems to grow more clique-y by the day.

Insert "It always was" meme

> why else would a ~billionair that could easily fund it himself pull in that many investors that early

That was my question, too. Does Mitchell even need to take outside funding? I get why you would prefer to use others' poker chips instead of your own, especially if they're happy to give them to you, but usually funding comes with strings attached.

> why else would a ~billionair that could easily fund it himself pull in that many investors that early?

Something something "never spend your own money if you can spend someone else's money." That plus the perception of having those specific investors gives the idea/business fast authority/prestige.

It's not the gangster move, but it's a move.

Of course they need VC money! How else are they gonna hire the worlds best vibe coders?

It's really awkward when you consider projects like Bun, which were VC funded, and ultimately bought out by a large company, what would they have even done to pay back the VCs that funded them?

Ghostty and libghostty was product market fit.

Superlogical (allegedly) uses it for the multiplexer.

Mitchell Hashimoto is just the icing on the cake and he has a track record of building defining dev tools used by millions of devs.

So it makes sense why you see 15 angel investors.

More money / investment in the dev tool space is great for everyone!

> Tobias Lutke /../ has most recently advocated for a pay-to-play version of democracy where wealth determines your ability to participate

He’s just looking at america post Citizens United and wishing he could ull the same shenans in Canada. It’s been working very well for the folks and orgs who can afford it

I also recoiled a bit when I say that Lutke was on the investors list.

Does Mitchell even need to take outside funding?

It's really unfortunate that the same traits that make you great at building a business (within capitalism) also often make you have psychopathic tendencies.

[dead]

[flagged]

People are only allowed to take funding from investors whose politics align with yours?

The vision of having a shared interface for all work seems pretty clear to me.

It's just another AI gobbledygook product whose sole purpose for existing is to cash in on the hype bubble, a few years from now this'll just be another in a series of infinite other VC grifts

It’s good to be skeptical, but this guy is the inventor of quite a bit of architecture software supporting modern systems

I think it’ll be fine

It’s a terminal multiplexer. Nothing in this announcement suggests to me that using this multiplexer will require the user to get root or super-root or any additional permissions. It’s what the user already has access to, but more “durable” (which I take a mean terminal session history on by default). And I don’t believe any short-lived secrets like ephemeral keys could have their lifetime extended by a terminal multiplexer. And nothing really supersedes RBAC which any respectable org should be using always.