> I'll make a claim -- does anyone disagree (?) -- the incentives are not pointed in the right direction. Unless developers are paying directly for the functionality they want, that's misalignment. So enshittification seems probable, maybe inevitable, unless something structural changes.
Exactly.
AI autocomplete I just don't care about. OK, but they do offer an AI subscription service. I'm using LLMs almost not at all currently, but at least right now this does seem like a viable path to getting a lot of money. I worry, though, that in the long run, the margins will just get squeezed by competition here, especially when a lot of deep-pocketed competitors can charge nearly nothing and get away with it. Zed also very nicely offers the ability to drop in an API key for their AI agent features... which is awesome and the right thing to do, but not good for profitability.
Obviously, the issue with simply charging money like Jetbrains is that Jetbrains already has built a world where they have a super complete IDE package and even then they only really charge a relatively small amount per month for the privilege. It works because for software the development costs for a given piece of software is relatively "fixed" (not entirely, but close enough) whereas the revenue you get grows as you acquire users. But now that the market is saturated, it's going to be hard for someone to jump in and try to be like Jetbrains in any way, and if you're charging money it's even harder to compete with most of the other code editors that are ~free.
Worse, it's hard to charge money and remain completely open source. Linux users will mostly not even realize you're doing it because the packages will just keep working for free since the software is open source. And Zed really does benefit from being open source, maybe more than many other open source offerings; it really seems like the community did a lot of work on making Zed work well on Linux. (By the way, if any of you who worked on the Wayland support for GPUI happens to read this, thank you very much for the excellent work.) So it's hard to say "just go closed source!"
I hope Zed Industries got a good deal on whatever funding they received and are being somewhat conservative with their funding too. Even though it's hard to compete in this market, they've built something that is very hard to build and basically not paralleled as far as I'm concerned. There's clearly something here and finding a less shitty path to monetizing it would be very, very good for us all.
(And I am willing to pay, though I know they need to find something that will make a lot of people want to pay, not just people like me.)
Maybe a year ago (or couple of months ago, I can’t recall) but I found out about Zed and since I was learning Rust, I wanted to contribute. But then I saw that they have a CLA. So while it won’t be easy to go closed source and not upset their users, it definitely is possible. Any contributions, no matter how important, are already covered by the CLA.
Ah yeah, they have a CLA that forces you to give them a carte blanche copyright license to contributions and they also restrict dependencies to be ones with permissive licensing too.
Pretty obvious what's going on here. I hate seeing it happen, though.
FWIW, we plan to sell "enterprise" builds of Zed with SSO, centralized plugin repositories, business-hosted infra, etc. etc. The sorts of features developers don't care about but businesses do. To do that, we have to be able to re-license the codebase and apply these changes. We have a lot of other ideas for business sustainability but the core, offline code editing experience will always be freely available and open source though :)
I don't really expect you to respond but if this is true why not use a CLA that guarantees the code stays open source? The FSFE FLA exists and is prominently used (though optional) by KDE.
Hey - I'm the founder of WorkOS. Happy to chat about the playbook we see with OSS projects spinning-off a commercial offering. It's pretty common and we work with a lot of these businesses, enabling them to continue investment in the ecosystem too. mg@workos.com
Exactly.
AI autocomplete I just don't care about. OK, but they do offer an AI subscription service. I'm using LLMs almost not at all currently, but at least right now this does seem like a viable path to getting a lot of money. I worry, though, that in the long run, the margins will just get squeezed by competition here, especially when a lot of deep-pocketed competitors can charge nearly nothing and get away with it. Zed also very nicely offers the ability to drop in an API key for their AI agent features... which is awesome and the right thing to do, but not good for profitability.
Obviously, the issue with simply charging money like Jetbrains is that Jetbrains already has built a world where they have a super complete IDE package and even then they only really charge a relatively small amount per month for the privilege. It works because for software the development costs for a given piece of software is relatively "fixed" (not entirely, but close enough) whereas the revenue you get grows as you acquire users. But now that the market is saturated, it's going to be hard for someone to jump in and try to be like Jetbrains in any way, and if you're charging money it's even harder to compete with most of the other code editors that are ~free.
Worse, it's hard to charge money and remain completely open source. Linux users will mostly not even realize you're doing it because the packages will just keep working for free since the software is open source. And Zed really does benefit from being open source, maybe more than many other open source offerings; it really seems like the community did a lot of work on making Zed work well on Linux. (By the way, if any of you who worked on the Wayland support for GPUI happens to read this, thank you very much for the excellent work.) So it's hard to say "just go closed source!"
I hope Zed Industries got a good deal on whatever funding they received and are being somewhat conservative with their funding too. Even though it's hard to compete in this market, they've built something that is very hard to build and basically not paralleled as far as I'm concerned. There's clearly something here and finding a less shitty path to monetizing it would be very, very good for us all.
(And I am willing to pay, though I know they need to find something that will make a lot of people want to pay, not just people like me.)