File: images(22).jpg (18.6 KB)
> Be Zed
> Literally just a fancy text editor that's also free
> Raise $30 mil
Explain this shit
> Be Cursor editor
> Be written in literal sloppiest of the slop electron shit
> Have valuation at $29 billion
Explain this shit.
Is this some scheme? I'm asking for a friend who is writing his own text editors, maybe he can also get a little bit of some of them billions.
Showing all 68 replies.
>>
>>108997719
Yeah it makes no sense. I guess they did som begging and got some "angel investors" early, and now they are sitting on their asses with their money?? I think they did some begging, and then some boomer gave some money and did not really understand that it was not investment?, but the project was promoted well?, or some EU tax-money nonsense? >:D
>>
>>108997775
This implies that those angels / boomers are stupid. If they just liked the idea of a text editor, they could've made it a non-profit contribution, like with LadyBird. And look at it, LadyBird raised like a million and a browser is a million times more complicated.
>>
>>
>>
>>
>>
>>
>>
>>
>>108997775
Angels are by definition small checks. There are billionaire angels that will give you a small loan of $1m, but those are super rare and require extensive connections and on top of that a lot of luck.
Normally you get 10-20-50k maybe. The point of angels is you want one that's strategic: either a field expert or someone hyperconnected in the investor world to get intros to VCs.
The problem is that VCs don't let you sit on money. If you do that, you will be sued for breach of fiduciary duty.
Also, you can't just find 20 angels and call it a day, because your cap table will be too complicated for investors in later rounds. $30m raise is a very strong B round or a C round, it's not something you're supposed to have until you have extensive proof that people are using and paying for your thing repeatedly, and you are on a strong path toward profitability. But in 2026 it's all muh AI hype without any rhyme or reason.
t. techbio, can't even raise preseed despite already having clients and written agreements for $1m/deliverable.
>>
>>
>>
>>
>>
>>108998174
Basically investors are lemmings. There are a lot that claim to be "idea stage" and "bet on the founder team only" and "people doing something different", but they're all exactly the same: you need to follow a template they've seen before (exact template, can't be a one-off company that has pulled it off before, must be a model that's been followed many times in the past), you need to sell them on a bullshit "we'll literally own the entire world" vision, and most of all, and most of all, everything must be derisked before they take over the whole company for the privilege of giving you a dime. The worst is the "idea stage" investors, who are for some reason least likely of all to invest early.
>>
>>108997840
>>108998141
Okay, what is your explenation then
>>
>>108997719
I thin Zeds buisness model might be a gateway to people buying a shite llm, so its adoption would be mutibly beneficial for multiple parties. Hope it has nothing to do with data storage though, and stealing user information...
>>
>>108999236
>>108999221
For zed I'm pretty sure they're raising under the premise that they'll be the next cursor, i.e. eying a multi-billion acquisition or collaboration agreement by a big guy trying to compete with cursor x spacex.
>>
>>
>>
>>
>>108999340
I worry about it precisely because I'm in the club, obviously. I wouldn't have to worry about it otherwise. I'm offering to make the old boys rich and have all the supposed evidence they claim to want to see that they'll profit a lot from it. Why would they say no to getting richer?
>>
>>
>>
>>
File: always ask.png (53.5 KB)
>>108997719
>>
>>
File: 1780258146328181.jpg (251.4 KB)
>>108997719
cursor had ai in it when the way to use ai was still in an ide and not the ai company's own app that controls ur entire computer
>>
>>
>>108997719
Zed offers an AI gateway, and Cursor does the same thing.
Cursor collects the prompting data and has all data Openai and Anthropic output.
Both things can be used to train AI models to perform very good.
Cursor proved it by post-training the best Chinese open-weight model with their data to perform very close to Opus for a fraction of the cost. This data makes them incredibly valuable, because all value in the AI space is derived from the data you have to actually train the models.
Inference is printing money.
This data will lead to SpaceXAI to be competetive with OpenAI and Anthropic, or at least a lot of people will bet on that.
Cursor's AI Gateway also generates a lot of revenue. Most of it goes to other companies, but they clip at least some out for themselves.
Zed tries to do the same thing.
>>
>>108997719
Cursor is AI first, or rather, AI only and tied to a subscription.
Of course investors fucking LOVE that. The value of a company isn't tied to whether their products provide any real world value, but rather whether investors think the value of the company will increase in the future and nothing else.
Everyone wants to be the guy who invested $100 into some literally who and got $1,000,000 when that company made it and became a big tech company.
If you still think you need to make good products to make a lot of money... hehe lol
>>
>>108999566
>data will lead to SpaceXAI to be competetive with OpenAI and Anthropic
Never thought about it that way. Good analysis, although cursor still has more than just data. They are the innovators and everyone always steals from them. Cursor itself mostly just steals ideas from it's own users who share what and how they do (and obviously share code as well), but also knows how to invest in RnD and get results.
Just throwing money in general direction of progress is not enough. They know how to actually implement and ship stuff, gather feedback, improve etc.
There is a bigger mistery than just figuring out why this or that company costs this or that mauch or raises certain money.
Can someone explain how Anthropic exists? To even begin to try and compete with OpenAI they needed compute, data and time. Who gave them compute and data and trusted that over X time they can deliver an product that would beat OpenAI?
Assuming they walked out of OpenAI office with naked asses, fired by Altman or something like that.
How? Explain this shit.
>>
>>
>>108999896
> Just throwing money in general direction of progress is not enough. They know how to actually implement and ship stuff, gather feedback, improve etc.
lots of people know that, I'm sure there are people who know it even better than them. Nut shipping stuff doesn't automatically equal profit. So the question remains, who the hell thinks this is going to pay off?
>>
>>
>>109001478
It doesn't. Zed is maybe a nicer editor, but I will not install it out of principle, because it's probably spyware. Nvim is just fine.
30mil they raised cannot possibly be justified given that Zed is just a thin layer on top of AI which even then people can just turn off. One argument in their favor is that devs spend 8 hours a day in their editor, but it's still hard r imagine how can this make actual money when there are so many free alternatives.
>>
I sometimes feel like I'm wasting time writing my editor, because literally who cares. I wish I could somehow crowdfund, hire maybe 2 more guys and destroy the shit out of Zed, VSCode and Cursor and also make it open source. But it's not realistic.
>>
>>
>>109001969
to add to that: I realized over the years I hate everything that has to do with business. I love the craft and I love making excellent software, but it's never ever profitable. It's almost like a law of nature that in order to be able to make any money, you have to be producing slop and I hate it.
>>
>>109001984
A few reasons:
1. It needs traction. I have 0 traction and I don't know how to get traction.
2. I don't think that even with traction people would send money. Maybe I could raise a hundred bucks if someone's generous enough.
Like I said I love the craft, I enjoy making goo software and I don't enjoy rushing with AI slop. I'm not good at business or marketing.
>>
>>109002003
You can go like zig. Make a 501(c)(3) (or (c)(6) if you prefer) and find people who use your software and have them donate. They can write it off and get tax benefits which is why they're incentivized to do it especially when they benefit from your software.
>>
>>109002031
yeah, it'll take me maybe another year to get to a point where this editor is usable (because I have a day job). Then okay, suppose I form non-profit or whatever... And post on HN. Then what? It's not traction, even if somehow my post gets to the top it'll only be a temporary bump. Maybe I need a person in charge of marketing, because I'm exceptionally bored and unhappy with having to think about that.
>>
>>109002027
2. You can use that as proof of traction which will in turn allow you to unlock more serious funding from bigger fishes. You don't have to fully raise through this. When bigger companies or bigger projects raise this way (e.g. double fine or ankama), that's how they do it: the fan money is just to prove people actually want the product, the real funders will be executive producers or VCs.
1. Why don't you have traction? Does nobody use your editor? Why not? Is it even available anywhere? If it's not available and visible then obviously you will never get users. If you do get users (even just one user) why is the user not using it after trying it out the first time? You must find a way to collect this feedback. The rest will fall into place easier than you think. You have to shill here or on other forums, make announcement posts on reddit and orange reddit, etc. If you don't know how to shill, keep it simple: be honest and say you're looking for feedback.
>>
>>
>>
>>
>>109002063
Also, if anyone can maybe name what killer feature they think an editor should have that no editor currently has, maybe that'll help me too. I think there might be somewhat of an utapped potential somewhere, but I'm fairly retarded and tend to focus on internal architecture and ironing things out.
>>
>>109002063
From what you've said so far the main issue I see is you don't understand how to get the feedback loop you need to get a product out. You say it'll take you a year to make your editor usable, that is not normal. You should have a usable, though spartan, editor today. You should be able to add features to it to reach the original goal at the end, but in the meantime you should be iterating in small steps that are fully functional in-between. Example: you don't need utf-8 support immediately, if that's a concern for you. Thinking of implementing an undo tree? Start with simple linear undo instead. So on and so forth. It takes an hour tops to have a working syntax highlighting editor with something like qt. That should have been version 0. You should have made a post showing this and asking what to implement. Not to actually get info but to see what people had to say. You should be looking at how people use editors and what you can offer that's different. Are you particularly lean? Particularly full-featured? You mentioned zed. Why do people use zed and not your editor? Not today, but when zed was first public and shit like that. When implementing a new feature, make a demo available and have people use it and tell you what they think. Even if the only user you have is a discord friend, that's going to change how you do things a ton. Since it's an editor, start developing using your editor as soon as possible because you'll know naturally which part is most painful. When you think it's kinda comfortable to use even though it is missing things, you should be having at least 3 digit users, if not you spent too long on dev and not enough on user outreach.
>>
>>109002213
This is too open ended, you have to ask better questions like go in an emacs thread and ask what is the thing that is most annoying to them with emacs, same with a vim thread or whatever, etc.
Anyway for me the problem is that there are only two relatively universal lightweight editors that can feasibly be used in the terminal: emacs and vim. Both have really shitty plugin systems, but nearly all their usability is in plugins. Vim performs way better without plugins than emacs. Plugins are buggy and often don't work well together. They are also slow and use lots of resources even for simple things. So while I "need" some plugins for pro work, I tend to avoid them in day to day, which makes me slower than I need to be (but I prefer comfort, makes me last longer which kills the speed advantage). I need the editor to be lightweight because sometimes I run it on a remote session. This used to be less of a problem some years ago but everything's on kubernetes now and it's not uncommon to need to debug something by dropping an editor in a container. This is often surprisingly slow, partly due to resource limits. I also do a lot of work on code that lives on remote machines in various ways where that's a problem. And having the same editor workflow everywhere is very neat. Emacs prefers living on your desktop and remote connecting, which can be a bit broken in more complicated auth setups, but otherwise is a superior workflow though. I don't think anything conveniently supports connecting across multiple ssh tunnels into a kubernetes cluster to a specific container.
>>
File: ick_select_words_to_EOL_SOL (1).mp4 (688.6 KB)
>>109002240
Let me then be more specific.
I already have very good utf-grapheme support, syntax highlight, theming etc. I have a very flexible keybinding and modes architecture which allow you to basically make it behave both like vim, helix, emacs or regular non-modal editors. It's written from scratch in D and Raylib to make it flexible enough to literally implement anything visually. I'm currently working on a widget/pane/flexboz system after which it'll be very close to usable. The reason I say another year until it's usable as my main editor is because I have a job which is exhausting and because it's not easy to write stuff that you have to think through instead of ai slopping it. Goal is not just a great editor, but also excellent maintainable code. Attached a small demo.
>>
>>109002310
Ah yeah, I remember one of the threads you make. Like others had pointed out to you, the problem is that your demo is useless. You have to demo something useful. Think about the usecase first. Using the editor yourself day to day will automatically reveal such usecases. One day you'll be working on it and use one of the things your editor can do uniquely well and it will dawn on you that THIS is the thing you need to show people. But you have to keep in the back of your mind that this is the kind of things you have to watch out for.
>>
>>
>>
>>
>>109002340
Same comment as before. The result is >>109002341
There is nothing you're showing that's compelling from a user perspective.
It seems you really like your multicursors. Personally I've always found them to be overhyped and very rarely useful, when you have block edit instead. But maybe there's something special that's very useful with them when combined with other features. Can you use them to make some common refactoring pattern easier for example? More than just :s in vim.
Otherwise just like >>109002351 said I would have said that one thing these "text" editors don't often have is ast-level operations. To me it's things like understanding the relationship between a declaration and definition in C and intelligently fixing them in refactors, or fixing call sites correctly when you change function spec, etc.
>>
>>109002341
I actually use neovim with neovide. It's fine, switched to it from vim a few months ago (been a vim user for 20 years). Those are good editors, but even nvim with neovide can hardly be called modern. Plugins are very often cumbersome and slow, like just yesterday I had tagbar freeze the editor, started profiling, turns out some interaction between it and airline made it super slow and I still haven't figure out way. But in general, I think vim/neovim is both overly complex and inflexible at the same time, while their code is maze.
Please understand though, I've started this as a hobby because I loved the idea of my own editor. I'm not trying to convert you or explain why my editor is better.
>>
>>109002365
> Multicursors
Interesting that you interpreted it as that, but those aren't multiple cursors, I also don't actually understand them. What you're seeing is multiple selection, which is also scriptable. I haven't gotten to it, but because you can compose commands in my editor, you can do a lot of interesting things with it potentially. Think about it like this: imagine you can select, pause, move elsewhere, then continue selecting from there and imagine those actions are scriptable/programmable.
>>
>>109002401
I dunno man, what you're describing sounds awfully like multicursors.
But here's the thing: it doesn't matter at all. What matters is what we see or do with it. Show don't tell applies as usual, and I really don't have a use case in mind for deleting multiple things that don't adhere to an easily :s-able pattern (I have encountered a few in my life but they're too rare to care about).
>>
>>
>>
>>108997719
Neither Zed nor Cursor are just "fancy editors".
They're first and foremost cloud services. Zed has Zed AI and Cursor also has several packages, with the most profitable one target at Enterprises. The text editing apps are the entry points for their services.
Also valuation and the actual wealth they have are completely different and mostly unrelated things. Valuation is the number of shares times the current value of each share that the investors gave it, most likely derived from the money they got from the investors.
Cursor doesn't have $29 billion laying around. Even if they perform an IPO they most likely won't reach it and if they do, it just means that their valuation is even higher than previously believed.
>>
>>109002240
>>109004088(me)
Which is to say thank you for the advice/posting
>>
>>
>>
>>108997719
zed is written by the ex atom devs
cursor was one of the first ai editors and anything ai related gets a high valuation due to being potentially disruptive. I am more wondering whatever happened to phind and perplexity both seem dead.
>>
>>