Hacker Timesnew | past | comments | ask | show | jobs | submit | BinaryPie's commentslogin

Is Kagi worth paying for? It's been on my radar for a while.


I’ve enjoyed it. It’s the first time I’ve not been left wanting by a search alternative. If it were to go away tomorrow, I’m not sure what I’d do for search.


Im happy with it. I have filters which will try to find search results from before 2022, which has greatly improved the quality of results for me


The domain-block feature makes it worth it for me. Finally no longer having medium or userbenchmark pollute my search results.


This. Or quora or pinterest or twitter/x or etc etc

I can outright block domains or just adjust their weight. Great for my personal prefs but also huge with the family account and helping keep the BS out of sight for the kids without going full restrictive.


You can self host searxng and block domains - increase priority on some etc. I do it myself.

https://github.com/searxng/searxng


Just be aware that a small percentage of your money would be going to the Russian government: https://ounapuu.ee/posts/2025/07/17/kagi/


I try not to buy from US companies these days, but Kagi is really so good that I make an exception here, despite the US government getting some of my money.


The EU is still buying billions of dollars of fossil fuels and other resources from Russia.[0]

[0] https://www.aljazeera.com/news/2025/10/3/how-much-of-europes...


Nefarious actions by others shouldn't justify your own. We can do better.


It's a matter of scale. Objectively, yandex is a great resource, and Kagi's results would be degraded without it. Pennies per user go to them. The sum of the entire money that has ever transferred from Kagi to yandex is what? 30 seconds of EU oil and gas purchases?

In short: pick smart battles.


So Yandex is one of their data sources? Is that allowed under new sanctions?


also duckduckgo use(d) yandex. not many alternatives in this space


I use duckduckgo and live in a neighboring country, so I know Russian well (thanks, imperialism) and have to search things in it from time to time. It's still good at those queries, so this is just an excuse.


they used yandex up until 2022, and now they use bing and other


My wife and I got the duo package because we do a lot of writing and need citations and sources. Compared to google and DDG it is less noisy and returns fewer spammy pages. We’re giving it a year to see if it is worth it.


Absolutely, yes. It has completely replaced Google for search for me. Good AI search as well if you're into that (but they don't force you to use it!).


I think generally yes. I tried it out for free for a while, found it was substantially better than Google and DuckDuckGo, and paid for a subscription.

Recently it has not had such a strong quality margin, which I suspect is due to the AI slop that all of the search engines are fighting against (due to errors both ways in their detection). I'm hoping this is temporary.

To be clear, I don't use any of their features except search (and domain filtering).


I think so. I switched to it because I have YouTube tv - essentially Google as cable tv provider - and noticed how commercials became too correlated with recent Google searches for comfort. The only time I end up switching back to Google is for looking up local businesses reviews.


If you expect it to be as good as peak google results, no.

If you need something that’s very noticeably better than its competition, then yes.

If you are okay with all of the terrible that comes with using LLM services as a search engine replacement, then probably no.

If you despise the amount of second guessing and source checking required to use LLMs as search tools, then yes.


It literally gives you google results (+ additional search providers, usually not in top results)... without the added spam. It's therefore superior to "peak google results".

What are you talking about LLM services? default search behavior does not use any LLMs (except any Google might use to reorder their top 10 results internally).


I wasn’t particularly impressed when I have the free trial a whirl - it’s not as bad as DDG, but it’s not anywhere near as good as Google.


Have had the family plan for over a year. Worth it for me. Just so easy to customize.


I generally like what Zed is trying to become. However, all of these features and blog posts are frustraing when they struggle to keep basic editor features stable. Edit a file outside of the editor? It's not going to show up in the project pane or the git diff. Need to work inside a container because it's 2025 and we don't need to clutter our local machine with 100s of dependencies and env managers... well now all the AI stuff is broken. ACP sounds cool until you realize every single CLI in existence works better.

My wish is that Zed gets the core working correctly 100% of the time before moving on to expanding feature sets. For now I'm back in NeoVIM because it always works the first time....

https://github.com/zed-industries/zed/issues/38109

Hopefully soon I can give it another shot at full time usage.


> Need to work inside a container because it's 2025 and we don't need to clutter our local machine with 100s of dependencies and env managers...

“Can I tell you about our lord and saviour Nix?”

Kidding, but seriously though I’ve found having to work in a container to be a bit clumsy, even with good tooling around it. As you said it’s 2025, and there are other ways to have reproducible toolchains that don’t pollute the rest of your system environment Nix or otherwise.


> “Can I tell you about our lord and saviour Nix?”

The "just use Nix" people are just like the "write it in Rust" brigade.

And the things both groups promote share exactly the same problem ...

A stupidly enormous learning curve.

That's why your average person prefers to just the job done in a container vs Nix, or in Go vs Rust.

I'm sure Nix is awesome, just like I'm sure Rust is awesome. But honestly, I've got enough going on in my brain at $work and $home without having to wrangle some obscure config syntax (Nix) or obscure low-level language complexities (e.g. Rust's infamous borrow-checker).

Every time I go back and look at Rust or Nix my brain is just like .... no, thanks.


Rust forces you to consider things upfront, or to put it another way, a C developer does what the borrow checker does manually. Sometimes, they do it after asan/ubsan/tsan has reported the issue. Sometimes it's after a segfault. Usually it's after the user hits the bug already.

Nix similarly forced issues to be reasoned with upfront.


Yes and most people are happier with a garbage collector than using C, C++ or Rust. There are domains or projects where you have to use one of those languages, but Rust is being used and promoted far outside those niches.


So you rather stay comfortable doing what you were always doing, which is fine but programming is everything but that...


Complexity for the sake of complexity is not a goal everyone should be striving for.


> Complexity for the sake of complexity is not a goal everyone should be striving for.

THIS !

Go's well maintained stdlib just lets me get on with it.

With Rust meanwhile, I have to decide which of thousands of third-party Rust crates is sufficiently en-vogue and well maintained.

Most of my coding work is backend stuff that calls a lot of HTTP APIs, parses a lot of JSON and does a lot of crypto. All that is ready-to-go "out of the box" with Go stdlib and on top of that is faster to get to production in Go than in Rust.

I don't need or want Rust's complexity. As I said I have enough going on at $work...


I always feel like choosing between rust and go is like picking between node (ts/js) and python when it comes to stdlib out-of-the-box-ness!


I think LLMs have solved that for Nix. I recently moved my homelab from an Ansible/Debian setup to NixOS. As long as I understand at a high level what Nix is and what problems it solves, I can let the LLM deal with the configuration language.


I didn’t say “just use Nix”, I jokingly made a reference to street evangelists and the similarities they have in their approach with people who are excited about a particular technology. In fact I said “Nix or otherwise” implying there are many options when it comes to reproducible per-project environments.


joking or not, people are very tired of the evangelists (fanbois?) from these ecosystems, and this culture around them is one reason others stay away. The broader trend is developers are burnt out on tooling churn


(That’s the joke)


That's a fine opinion to have. You just have to fight a separate shoggoth of a mutable-root system, which mainstream Linux vendors like Google and Valve have avoided.

It's similar to the Rust situation in that many Rust programmers are simply tired C++ programmers fed-up with using Valgrind and Cmake. Nix is the retirement-home for people who clean up package pollution and just want stability at any cost.


In the case of AI tools it is for security reasons, not just reproducible toolchains.


We need more selinux and sandbox-exec in daily dev lives. There's no reason to bring a whole new system along just to restrict some access.


Docker is just a shim on kernel isolation APIs. It’s not any different, but better packaged.

But irrelevant in this case. I dev on macOS. I’m not aware of any other options.


sandbox-exec. It's not great, but it's usable. https://igorstechnoclub.com/sandbox-exec/

> It’s not any different

It's very different. With docker on Mac you're running a VM which runs a wrapped up complete system that runs your app.

With selinux/sandbox-exec you run just your app and can skip the extra packaging needed for docker and mounts. (And get the extra performance)


Does selinux/sandbox-exec work on a Mac? Is this an apples to apples comparison?


sandbox-exec is a Mac exclusive thing.


Wow TIL


Any references to how you use selinux? I've used bwrap and firejail, but I don't feel confident that I'm not leaving holes open?


Honestly - no. To use selinux you need to commit to actually learning how it works and experimenting a bit. I don't think there's an easier way than reading both redhat (https://docs.redhat.com/en/documentation/red_hat_enterprise_...) and NSA (https://www.nsa.gov/portals/75/documents/resources/everyone/...) docs.

If you're happy with firejail, make sure you use whitelists only and you'll be 90% there with what's possible to achieve.


With the AI stuff, it feels like they invested a bit prematurely. When the Agentic editing demo came out (6 months, 10 months ago? It’s a blur), it felt right. Accepting and reviewing edits, live tracking ..etc., felt like pair programming. The ACP addition felt like a natural evolution .

With the continuous improvement in CLI tools and people’s experience with them, it feels like doing a live review or edit-by-edit approvals all feel like a drag. I personally have come to avoid using the IDE/Editor. I just kick up Claude code - plan mode, auto-accept edits. Once the session is done, switch to the editor and make necessary adjustments. I suspect people with Max subscription and “dangerously-skip-permissions” …etc won’t even care if an editor has AI integration or not.


The only editor integration I think is semi useful is wiring it into Diagnostics/Problems data that the editor has from extensions. Speeds up the agents flow quite a bit when it leans on that to check its work vs always executing (say) “eslint” directly.

But that can be done easily enough with an MCP extension for your editor/IDE of choice


Claude hooks are great for this.


I used to only use JetBrains for AI stuff, now I just open everything in Zed because of its Claude Code integration. Especially with the linters and other nice to haves. I am insanely close to cancelling JetBrains.


Theres a plugin for Claude Code and you can use Claude models from Copilot in IntelliJ. Are those options worse than Zed, especially considering other comments about Zed missing basic editor stuff?


Claude Code in IntelliJ just opens a terminal tab. Claude Code in Zed feels integrated into the editor as if it were a natively supported model.


The copilot plugin with Claude models (Sonnet models) is very integrated in the IDE, and so is Jetbrains AI (even more integrated than Copilot) which I believe can also use those models. So I wonder what's the actual difference.


JetBrains products has Claude Code integration. Is it worse?

I still use Claude Code in cli, as a WebStorm user.


The plugin just opens a terminal whereas in Zed you get a more natively feeling experience. In JetBrains they do support it as one of their agents but now you have to buy into JetBrains AI credit system instead of just using your Claude subscription.

Theres two ways to use Claude Code in JetBrains and instead of helping them make their plugin better they opted to try and make money off of it. Which Zed could have done the same but instead lowered their monthly to let you use your own Claude Code subscription.


Thanks. The JetBrains integration sounds lousy!


I am not getting tons of issues with Zed going out of sync all the time. I wonder if the issue is it silently having issues watching the filesystem due to open fd limits.

I've noticed that not only does it sync but it even will recognize if I rename the folder a workspace is in.

But then, I've run into a couple of strange issues that tell me there is more polish needed:

- Sometimes upon using LSP refactor, it seems like if a bunch of files get renamed, the open buffers will get screwed up somehow. Like, I'll hit save and it will write to the old filename! It's not actually a huge problem, as I can close the buffers, delete whatever excess files I accidentally create, and re-open them without error, but it is confusing as hell.

- I have indeed had issues with the file view not always updating when files are added externally, however it is not constantly. I usually just reload workspace when this happens. It is a minor frustration, but I had many minor recurring frustrations with both VS Code and Neovim before too, so I don't consider it a deal breaker.


yup, my pet peeve is there is no way to disable line wrap. the setting that exist doesn't work and there's no way to actually disable it instead of just increasing the max characters (with set hard limit in the source code).

have a big docs or log,data file where you don't care for the rest of the line ? well too bad better have a spare editor.

this feel to me like it should should be a number #1 priority. "an editor need to nail the editing part".

https://github.com/zed-industries/zed/discussions/26344

on the positive side I do like that you can in-place edit the result of global search.


Huh? That doesn’t track with my experience. I spent a long time trying to find out how to enable line wrap, and Zed’s settings are more extensive than most editors.


I know this is silly but the biggest thing that’s driving me away from it is how god awfully blurry it looks on my 1440p screen :/


> I know this is silly but the biggest thing that’s driving me away from it is how god awfully blurry it looks on my 1440p screen :/

Not silly at all. User experience is important. If you're going to spend as much time in a tool as most programmers do in their editors, it should look and feel nice.


I don't have this problem, but issues like this are always a huge barrier, especially when you want users from a polished code editor like VSCode. Personally, I use it because VSCodium does not support the default Python LSP on my Linux box. I like it, but there are definitely areas that seem rough around the edges. My biggest issue is the size of the font and icons. I use a 4K screen, and while the font is readable, the icons are so tiny that I can barely see them. Since it is a new system, I don't just know where they are like I would if it were VSCode.


Yeah same, plus I have poor eyesight so I have to zoom then to click anything


This is also a big issue for me and has one of the largets issues for a while. [1] I wish there was font hinting to the like of windows font rendering.

[1] https://github.com/zed-industries/zed/issues/7992


That's not silly! Great UX is table stakes.


It's been that way since day one. Wild that they haven't fixed it. Edit: it was fixed in Oct 2025, version 0.207.4


Try the latest version, it should be better

Personally, I never registered that as a problem, but I can see the delta in screenshots between versions.


Oh that’s exciting to hear, hopefully it’s better! I’ll give it a shot


I have the same problem on Linux-wayland/4K screen. I can't get it to look anywhere near "okay".


So they're optimizing for 1080p or 4k?

No settings to adjust?


optimized for HiDPI screens (like those on Apple devices)


ohhh... is this that thing where people say "blurry" to describe anti-aliasing, because they actually want it to be pixelated?


Nope, it is the thing where people say "blurry" to describe "blurry".

https://github.com/zed-industries/zed/issues/7992#issuecomme...

This is how Zed looks like on my computer too. VSCode manages to render things crisply while Zed is a blurry mess.


Thank you for the example. That shows exactly what I was talking about, so this issue is clearly (at least somewhat) a matter of taste. To my eyes, those VSCode samples look harsh, crudely pixelated like something out of the 1990s, while the Zed samples look like normal anti-aliased text.


No other editor or app that I use on macOS looks like this. None. This is not just AA, I like properly AA’d text.


Ahh, thanks for the explanation. I guess I was thrown off by the screenshot attached to that bug report upthread, where someone compared zed's rendering against completely non-anti-aliased text rendered by VS Code.


No, this is a thing where the text is honest to God blurry on non-4K displays. The macOS version of Zed on a low DPI display has the worst font rendering of any application I've ever seen, and I used desktop Linux twenty years ago.

It's like they're rendering a high resolution font at low resolution using the simplest possible algorithm without lining it up with the pixel grid. It's very fuzzy. Characters have this weird sort of additive color intensity where strokes intersect that reminds me of Geometry Wars. It's broken.

They are working on it; the Windows build has decent rendering, and apparently the Linux version substantially improved recently. But they haven't gotten to macOS quite yet. I've been checking in on Zed every few weeks since it went public waiting for a fix.


It might be. I can't speak for anybody else, but it's probably the term I'd use myself if I was trying to explain it, even if it's perhaps not exactly the perfect one. The pixel edges just end up not hard enough, and the pixel corners just end up not sharp enough, and my eyes don't seem to like it. Whatever they're looking for when they're looking at text, pixelly text and printed text seem to supply it, and anti-aliased text doesn't.

(Though there must be some cutoff point past which the pixel size becomes irrelevant. I certainly don't mind reading stuff on the iPad.)


It says they're targeting Spring 2026 for their 1.0 release, so I'll treat as beta and put a calendar entry in for April 2026 to check back in on it.


My experience has been so different. Zed seems to always do the right thing for me when I concurrently edit files with other tools. Not doubting your experience or anything, but you must have a very different environment than me. Zed has been absolutely rock solid for the past year on my computer.


My favorite bug was Claude Code kept editing some template file. Zed kept auto-formatting the file, which was BREAKING the template itself. It took me like 30 minutes of watching Claude implode, and literally using "ed" to edit the line, before I realized, then I started asking Claude how to turn off the Zed formatters to which it was like AH THAT MAKES MORE SENSE, which I thought was hilarious after it tried everything from editorconfig onwards.


The thing about "basic" here is that it's subjective. It could be that these issues only happen on your machine or that the staff (or even most people) don't need what you're asking for. Of course they should try to fix them anyway, but their backlog is enormous.


I would say you're being harsh but their file tree refresh is bad(so is vs code but they have a manual button), their ai panel can't beat Claude code and their panes / window layouts are not any better or flexible than other editors. I still use it but I'm close to going back to vanilla vs code. I'm just hanging on to determine if my frustrations are from me just being used to vs code or otherwise


I agree with you. Zed has a great vision and the team clearly cares, but the foundation still feels a bit shaky. Stability in core workflows matters more than any flashy feature, especially for people who just need their editor to disappear into the background. What gives me hope is that the community feedback is loud and consistent. If the team focuses on tightening the fundamentals, Zed could still become the lightweight but powerful editor everyone hoped for. I’m rooting for them, and like you, I’m ready to give it another full try once the basics feel rock solid.


Oh yeah I’ve noticed both the desync and AI not working via ssh.

Also buggy git support, I selected a few things but it committed everything and made me think I lost it all.

But I love Zed when it works. Literally 5h more M4 battery life vs Cursor.


For me, I was happy using Zed until suddenly an update caused it to crash every time I opened it. It was caused by some sort of issue with graphics drivers on linux I guess.

I just checked the issue [1] and it is fixed now, but it's crazy to me that I never really thought of my text editor needing to use my graphics card for rendering.

[1]: https://github.com/zed-industries/zed/issues/37448


This, 100%.

So many of the basic features (e.g., automatic Python venv, Pyright running, etc.) have random bugs that pop up from time to time, making the basic editor unreliable.

My fear is, if they keep going in this direction (adding bloat without fixing basic functionality), they'll be a perfect fit for Microsoft acquisition, at which point I'll have to switch careers, because there isn't any other editor out there that I like.


Yeah, I have been fighting Zed to get agents to use podman on my host, but Flatpak is sandboxed and makes it almost impossible. The ideal solution would be that Zed could use podman or docker to spin up a container where agents could run free!


Agreed, we should support this!


It’s not just a nice to have. It’s a hard requirement. If I can’t launch agents within a sandboxed container, I can’t use your product - by security policy.


Curious about the failure to detect FS changes made outside of Zed. Are you on Linux?


I have observed this too, mostly for content that is changed in a symlinked directory, but also generally. I'm on Fedora Silverblue running Zed Preview as a Flatpak. It works great in most other ways though, snappy and beautiful, but the sandboxed environment provides some additional challenges.


I agree with you 100%. If basic functionalities are not airtight, there’s no way I am going to deal with growing pains just because they want to get paid on AI fluffs. Contrast this with something like Ghostty.


How does Neovim handle outside changes then? Or is there a plugin to make it work? AFAIK it doesn't reload any bufferes when files change. IntelliJ is the only other one I know that does it transparently.


In vanilla neovim you can use autoread... this does depend on a focus event like entering and leaving the pane or switching buffers. However, for my workflow which is "go to a different terminal pane and do some things then switch back" as soon as I focus the buffer it updates.

Where as with Zed it'll just keep showing the old content and in fact closing and opening file wont even change what it shows in Zed. It's really really annoying. I have to exit zed and open it again. This means if you are working with AI agents you end up having to do this often.


Very strange, this has always worked for me with Zed - both with local and remote ssh sessions.

I use `"autosave": "on_focus_change"` which may be keeping my buffer in sync with the file contents as I switch between terminal and zed.


You can configure Neovim to either auto reload or to ask you whenever a file is changed. (Can't remember which one is the default, probably to ask.)


I'm ~80% sure auto-reload is on by default, and 100% sure that it's on in a fresh LazyVim install.


vim.o.autoread (default on) When a file has been detected to have been changed outside of Vim and it has not been changed inside of Vim, automatically read it again.


Thanks, just found something about that.


Plain vim asks what to do by default. A single 'l' does a reload.


What I like in zed (pretty new to me) is that I dont have to deal with vim/neovim plugings to make the basic stuff work. Zed workds out of the box(for rust).


You may enjoy Helix as well: https://helix-editor.com/


Agreed. I hear way too much about Zed considering the editor doesn't allow a window to be popped onto a second monitor.


On MacOS 10.15, I have 2 monitors plus the built in on a macbook pro, and I have Zed windows on all 3 of them all the time.


You have to open three Zed instances, don't you?


No, this works fine with a single Zed instance.


I stand corrected.


nope. Single zed app instance.


All of our tests including computer science exams were written on paper with a pencil. We can just go back to that.


There used to be tons of these and even turn key libraries that do all of this and much much more. Especially in the early to mid 00s. Did you do any research into them before you built this?


You forgot "hacker" :D


Wait are you saying I have to be content with who I am?

What do you mean I can't expect my employer to validate my passions in life?


> Wait are you saying I have to be content with who I am?

When did he ever tell you to stop bettering yourself?

> What do you mean I can't expect my employer to validate my passions in life?

Because your passions are yours, and yours alone. You could get miraculosuly lucky and get an employer that shares some of your passions and lets you work in them... or you could remove chance from the equation and write code on your own, and open source it.


Sarcasm.... The internet doesn't get it.


Keep on the path for your green card. 2 years is not such a big deal in the grand scheme of things.


And try to enjoy life in the mean time. Life is not about the destination, the trip is also important. You can do things on your free time, but you can also find ways to enjoy your job more.



The future is Orange. /sigh


I choose Freedom.

Being able to work the hours I want to work. I can come into work at 10 instead of 8. I can leave work at 3 if I wanted to. As long as my work gets done.

Not having a boss micromanage me. Not having to worry about office politics (for the most part). Also the freedom to run with great ideas and not have the bureaucracy of some corp jobs.

Also Being able to work on a wide range of technologies and work with people from many different verticals and experience levels is a great experience and keeps my job (programming) fresh.

Finally startups generally only last a couple years. This allows me to change jobs every few years and lets me do something else. This is really beneficial as I'm able to feel refreshed and not trapped in a grind.


Guidelines | FAQ | Lists | API | Security | Legal | Apply to YC | Contact

Search: