Looks really good! Maybe add in the title that it's MacOS only for now.
There's options for local and cloud-based AI - which is nice. I'd be cool to see what's the compute cost of requests!
I’ve been using warp as my main terminal since getting a private beta key from a coworker a bit more than a year ago.
I don’t really use most of the advanced features it has to offer but I’m quite happy with it. It’s fast, it has a few ergonomic features that I haven’t seen in most macOS terminals, so no complaints here.
If everything stays the way it currently is, I don’t think I’ll move from warp to something else in the near future.
I might look into the AI features though, it seems nice not having to wade through man pages for basic CLI stuff that I don’t really do often but is sometimes necessary
I check out Warp every 6 months or so, because I’d love to see more innovation with the terminal, and the screenshots look great. But the story’s the same every time: I download the app, fire it up, and am greeted by a mandatory ‘sign up’ screen and privacy policy, at which point I close and immediately delete the app.
I will never be okay with a terminal that requires me to have a proprietary login to operate on my own local file system with local tooling.
Terminal includes everything than IDE or web browser can, I would say.
Only in-memory access is restricted and even that might not be the case, if you run root user on terminal.
But how else are you supposed to extract revenue and information from your users? People can't just use software without the vendor knowing who they are and what they're doing! Preposterous. Especially for a terminal where we must implement everything because every last thing a user does in a terminal must be supported by the terminal because this has never been possible before. Our product is the first. No libraries exist which do these things. AI. AI, AI, AI, AI, AI.
Also, when we say that the terminal is AI powered, it serves no real purpose for the user, but it allows us to capture every single keystroke so we can send it to our AI, and collect useful information about what people use terminals for. Information that is available in no other way and has never before been understood by anyone.
Also, how is this company going to make money when they inevitably need to if they don't force you to log in? I swear, no one has the silicon valley hustle mentality anymore...
It appears that you hold the belief that extracting revenue and information from users is an essential aspect of software usage. However, I beg to differ, and I shall explain why.
Firstly, the notion that people cannot use software without the vendor incessantly monitoring their activities is far from accurate. It is indeed preposterous to assume that every user action must be surveilled and recorded. In fact, respecting users' privacy and allowing them to have control over their data should be fundamental principles of any software. Rather than exploiting users for financial gain, we should strive to create software that empowers individuals and respects their rights.
As for your claim that your terminal is the first of its kind, and no libraries exist to accomplish the tasks you mention, it is important to note that innovation builds upon existing technologies. Many robust and efficient libraries exist that enable a wide range of functionalities. Instead of reinventing the wheel, it is more prudent to leverage and contribute to the existing open source community, which encourages collaboration and knowledge sharing.
Regarding the assertion that your terminal is "AI-powered" solely for the purpose of capturing keystrokes and collecting user information, I must express my deep concerns. Transparency and informed consent are crucial in any software endeavor. Users should have full awareness of the data being collected, how it is used, and the potential implications for their privacy. It is unacceptable to exploit the AI buzzword to justify invasive surveillance practices without providing any tangible benefits to the user.
Lastly, the assumption that a company cannot be profitable without coercing users to log in reflects a narrow perspective on business models. While some companies may adopt such tactics, there are numerous successful alternatives that prioritize user trust, value creation, and fair pricing. A sustainable and ethical approach to business involves offering products and services that genuinely benefit users and respecting their privacy choices.
In conclusion, it is essential to shift our mindset away from viewing users as mere sources of revenue and data to be exploited. Let us instead focus on creating software that respects privacy, empowers individuals, and fosters a culture of openness, collaboration, and user-centricity. Only then can we truly advance the field of technology while upholding the values and principles that benefit society as a whole.
That you have to log in to an account use? What logging and data mining are they performing? not advised for anyone that has access to secure or sensitive data.
Hey there, engineer from Warp here. Warp never collects input or output from the console. There is only high-level tracking for monitoring app crashes and some feature usage data. Our philosophy is total transparency around telemetry. We’ve created a Network Log tool you can use to watch all events in real-time [1] and we have a full list of telemetry events we record in Warp’s docs [2]. You can also opt out of telemetry in Warp’s settings.
Warp requires a login to support its collaboration features (for example sharing commands with a team). Login makes it possible for developers to collaborate securely.
> Login makes it possible for developers to collaborate securely.
Is that the only way for developers to collaborate securely? Aren't there other technologies out there can enable secure collaboration without requiring me to sign up for yet another centralized, opaque, data vacuuming operation?
I'll install and try this as the interactive suggestions look nice (finally, software that can RTFM for me - nice! :) , however, from your description of the purpose of login, couldn't the same effect be gained by locally creating a unique (for warp interaction) ssh key that preserves privacy and yet allows developers to securely communicate by sharing their public keys. A similar process might offer more comfort in deciding to allow the telemetry.
> Warp requires a login to support its collaboration features (for example sharing commands with a team). Login makes it possible for developers to collaborate securely.
So if I don't care about collaborating, can I use it without an account?
You are never going to get broad adoption if you require someone to sign up to use something locally. Full stop. People who will install any software that comes across Hacker news are going to pause the moment you make them sign up to use a local application.
All of the explanations about how it's not intrusive and only collects X, Y, or Z doesn't matter. There are many apps I use that send crash reporting data, and the only time they do that is when they crash and then they ask me if it's ok. I don't have to sign up to do that.
I’ve been using warp since launch and have found it to be a very good experience. The main advantages for me are the well-featured and responsive UI, along with workflows being a very well thought out addition.
For a while I stopped using it daily and instead used the built in terminal for VSCode, but found myself going back to it again and again any time I needed the terminal for a non-project related task. I’ve returned to Warp for my primary terminal a few days ago as it just works better and is more enjoyable to use. The AI addition can seem like a gimmick but it does make it easier to figure out what you need to do (and the one click workflow creation from a result is also quite nice).
I’d recommend warp for anyone who values a really well done UX combined with some very user-friendly features, but long-term shell power users that muck about under the hood would probably not like it as much.
Note: my only feedback for the Warp team is how the prompt shrinks vertically when you click outside it to another part of the command history. Always bugs me that the vertical layout shifts when I click around.
Could anyone who's actually using this spare a moment for an elevator pitch? I am one of many developers who have a latent intuition that our shell programming environment could be made more efficient, but I've yet to see a "killer feature" that is an obvious (even 2x would qualify) productivity boost.
One tool that I'm really excited to implement and use, but haven't found the time (and I'm sure someone else has already started), is a command that wraps any other command and exposes a global hotkey that, when pressed, shovels the last N lines to an LLM and asks it for advice.
Further down the line, I'm imagining being on a team where all developers' error logs are piped into a central feed and that feed is used to eliminate a large swathe of duplicated effort (how often is it the case where a whole team spends an hour or two debugging the same regression, broken dev env, etc?).
> One tool that I'm really excited to implement and use, but haven't found the time (and I'm sure someone else has already started), is a command that wraps any other command and exposes a global hotkey that, when pressed, shovels the last N lines to an LLM and asks it for advice.
Sounds like a 1-line config in a decent WM doing a curl post. You could probably turn it into a SaaS ;-P
> shovels the last N lines to an LLM and asks it for advice.
Theoretically, you could set this up locally with llama.cpp (and any terminal that runs a command with a hotkey) right now. A 7B or 13B finetune will start up pretty quick.
Or in a company environment, you could keep it loaded on a server for faster responses.
Also, you could stuff previous outputs into a vector DB and get context from you (or colleague's) previous outputs.
Definitely an interesting terminal. What bugs me is I can't switch the autocomplete command from <enter> to <tab>. Muscle memory keeps doing the wrong thing and in a terminal, that's potentially catastrophic.
<enter> as an autocomplete command in the terminal seems really dangerous. There are a lot of commands where accidentally running the command that was partially formed instead of completing it could cause a lot of problems.
Last time I tried this, the fully-custom rendering with all of the custom UI, and the horrible text rendering forced me to not use the app. Honestly, using the app shows a lot on how much the Rust GUI scene is not a matured state. This is beyond the level of Electron, everything feels... just so unnatural (in a way that it makes Electron feel lovely).
What do you find wrong with the text rendering? I'm curious because I'm picky about that and have begrudgingly stuck with the macos terminal because despite its other flaws text looks good. I'm not using warp daily but I have it installed and I don't notice the other text problems terminals tend to have.
Oh yeah I also found that to be the case. I had trouble getting some key combos working how I wanted though and didn't feel like sinking more time into learning the config language right then. Been meaning to go back and get it set up correctly though, definitely seemed like a good fit for me if I did it properly.
Hey there, thanks for feedback. We built our own UI framework in Rust to build Warp, and there were some initial rendering quirks--especially with rendering text on low resolution monitors.
I'd love to hear more about the rendering issues you've faced when using Warp so we can fix them!
I like warp and it’s exciting to see innovation in the space. But I’d love for someone to explain how they’ve raised 23M dollars. I’m assuming that’s amounts to a 100M valuation? How does a VC expect to get multiples of that kind of money back selling a slightly nicer terminal?
I would rather see innovative tools that lessen our dependency on 50+ year old tech. This is still a glorified teletype. It uses AI to autosuggest git commands? Contrast with Magit[1], which (while it has a tiny bit of a learning curve, but also nowhere near 23M in funding) actually makes interacting with git a pleasure.
Er, how exactly does a tool that runs inside Emacs lessen our dependency on 50+ year old tech? Ok, Emacs is not quite 50 years old yet (first release 1976), but it's no spring chicken either...
True, and I'd personally rather move away from Emacs to something more modern. (Helix is great, although I appreciate the irony of it being terminal-only, while Emacs supports several different window systems natively.) Magit is the only real reason I'm sticking with Emacs.
Magit itself is powered by <https://github.com/magit/transient>, which I see more as an interaction paradigm than a library; it could enable more ergonomic interaction with other stateful tools that are typically native to the command line / terminal (such as docker/kubectl, systemctl, mpd/mpc, etc). Rather than using Emacs as a middle layer, Transient could build on top of pluggable native toolkit backends, such as Cocoa, Gtk, Win32, or even web or a terminal.
We continue investing into terminals because the terminal remains the lowest common denominator of interacting with a computer. On the other end of the spectrum we have Electron, which has very clear and obvious downsides. I think there is low-hanging fruit with amazing ROI somewhere in the middle, and Magit/Transient is an example of what it could be.
49 comments
[ 3.3 ms ] story [ 63.5 ms ] threadI don’t really use most of the advanced features it has to offer but I’m quite happy with it. It’s fast, it has a few ergonomic features that I haven’t seen in most macOS terminals, so no complaints here. If everything stays the way it currently is, I don’t think I’ll move from warp to something else in the near future.
I will never be okay with a terminal that requires me to have a proprietary login to operate on my own local file system with local tooling.
Totally agree it's not ok.
Also, when we say that the terminal is AI powered, it serves no real purpose for the user, but it allows us to capture every single keystroke so we can send it to our AI, and collect useful information about what people use terminals for. Information that is available in no other way and has never before been understood by anyone.
Also, how is this company going to make money when they inevitably need to if they don't force you to log in? I swear, no one has the silicon valley hustle mentality anymore...
Firstly, the notion that people cannot use software without the vendor incessantly monitoring their activities is far from accurate. It is indeed preposterous to assume that every user action must be surveilled and recorded. In fact, respecting users' privacy and allowing them to have control over their data should be fundamental principles of any software. Rather than exploiting users for financial gain, we should strive to create software that empowers individuals and respects their rights.
As for your claim that your terminal is the first of its kind, and no libraries exist to accomplish the tasks you mention, it is important to note that innovation builds upon existing technologies. Many robust and efficient libraries exist that enable a wide range of functionalities. Instead of reinventing the wheel, it is more prudent to leverage and contribute to the existing open source community, which encourages collaboration and knowledge sharing.
Regarding the assertion that your terminal is "AI-powered" solely for the purpose of capturing keystrokes and collecting user information, I must express my deep concerns. Transparency and informed consent are crucial in any software endeavor. Users should have full awareness of the data being collected, how it is used, and the potential implications for their privacy. It is unacceptable to exploit the AI buzzword to justify invasive surveillance practices without providing any tangible benefits to the user.
Lastly, the assumption that a company cannot be profitable without coercing users to log in reflects a narrow perspective on business models. While some companies may adopt such tactics, there are numerous successful alternatives that prioritize user trust, value creation, and fair pricing. A sustainable and ethical approach to business involves offering products and services that genuinely benefit users and respecting their privacy choices.
In conclusion, it is essential to shift our mindset away from viewing users as mere sources of revenue and data to be exploited. Let us instead focus on creating software that respects privacy, empowers individuals, and fosters a culture of openness, collaboration, and user-centricity. Only then can we truly advance the field of technology while upholding the values and principles that benefit society as a whole.
Yes, the sarcasms is quite obvious.
Warp requires a login to support its collaboration features (for example sharing commands with a team). Login makes it possible for developers to collaborate securely.
[1] https://docs.warp.dev/features/network-log [2] https://docs.warp.dev/getting-started/privacy#exhaustive-tel...
Is that the only way for developers to collaborate securely? Aren't there other technologies out there can enable secure collaboration without requiring me to sign up for yet another centralized, opaque, data vacuuming operation?
So if I don't care about collaborating, can I use it without an account?
All of the explanations about how it's not intrusive and only collects X, Y, or Z doesn't matter. There are many apps I use that send crash reporting data, and the only time they do that is when they crash and then they ask me if it's ok. I don't have to sign up to do that.
For a while I stopped using it daily and instead used the built in terminal for VSCode, but found myself going back to it again and again any time I needed the terminal for a non-project related task. I’ve returned to Warp for my primary terminal a few days ago as it just works better and is more enjoyable to use. The AI addition can seem like a gimmick but it does make it easier to figure out what you need to do (and the one click workflow creation from a result is also quite nice).
I’d recommend warp for anyone who values a really well done UX combined with some very user-friendly features, but long-term shell power users that muck about under the hood would probably not like it as much.
Note: my only feedback for the Warp team is how the prompt shrinks vertically when you click outside it to another part of the command history. Always bugs me that the vertical layout shifts when I click around.
One tool that I'm really excited to implement and use, but haven't found the time (and I'm sure someone else has already started), is a command that wraps any other command and exposes a global hotkey that, when pressed, shovels the last N lines to an LLM and asks it for advice.
Further down the line, I'm imagining being on a team where all developers' error logs are piped into a central feed and that feed is used to eliminate a large swathe of duplicated effort (how often is it the case where a whole team spends an hour or two debugging the same regression, broken dev env, etc?).
Sounds like a 1-line config in a decent WM doing a curl post. You could probably turn it into a SaaS ;-P
Theoretically, you could set this up locally with llama.cpp (and any terminal that runs a command with a hotkey) right now. A 7B or 13B finetune will start up pretty quick.
Or in a company environment, you could keep it loaded on a server for faster responses.
Also, you could stuff previous outputs into a vector DB and get context from you (or colleague's) previous outputs.
I'd love to hear more about the rendering issues you've faced when using Warp so we can fix them!
agreed on this - I have a large display (32" 6K) and I like using this to highlight the aesthetics of type rather than sheer information density.
It grinds on my OCD to see only the top half of a cursor/text in the text input box, as I regularly see with Warp.
I do hope they open source anything useful, if anything, they've done before the company is wound up.
[1]: https://magit.vc
Magit itself is powered by <https://github.com/magit/transient>, which I see more as an interaction paradigm than a library; it could enable more ergonomic interaction with other stateful tools that are typically native to the command line / terminal (such as docker/kubectl, systemctl, mpd/mpc, etc). Rather than using Emacs as a middle layer, Transient could build on top of pluggable native toolkit backends, such as Cocoa, Gtk, Win32, or even web or a terminal.
We continue investing into terminals because the terminal remains the lowest common denominator of interacting with a computer. On the other end of the spectrum we have Electron, which has very clear and obvious downsides. I think there is low-hanging fruit with amazing ROI somewhere in the middle, and Magit/Transient is an example of what it could be.
If you are interested in trying Warp (including our new collaboration features), you can download Warp here: https://www.warp.dev