They built webgl components and a K8s cluster with GPUs just to generate random artwork for use on their website and release graphics? Please tell me there was a more pressing use case for that kind of infrastructure and code slinging.
Are they just bored over there?
Edit: i'm not trying to rag on JetBrains and I generally LOVE their products. But this work seems way over-engineered and not a great use of resources, unless its the graphic designer building all that stuff :)
They had to suspend the R&D they did in Russia due to Ukraine, but I'm surprised they were able to keep it over there for so long given the other perils of having employees over there (that law which forced most American tech companies to close their R&D centers a decade or so ago).
> that law which forced most American tech companies to close their R&D centers a decade or so ago
Given that Intel and Nvidia had huge R&D centers in Russia until the middle of 2022, I highly doubt such law ever existed. Unrelated, but I also doubt US (as a country) won more than it lost from forcing them to withdraw from Russia, because Huawei literally hired whole divisions with all the expertise and inside knowledge.
> Russia can arrest local employees if they don’t like what their parent company is doing.
That is a real risk for maybe few very high-level expat managers, but certainly not a risk for average programmer Ivan. Besides, as if Google cared about countries arresting local employees - they have multiple offices in China, from all countries!
Make of it what you want, but the piece of anecdata I have from a person working at Russian Intel office in 2014 is that after Crimean referendum US government unofficially recommended big US companies to cease operations in Russia. Some of them agreed, like Google, because it was easy for them due to a small footprint, some of them, as Intel, deemed it unrealistic back then.
> That is a real risk for maybe few very high-level expat managers, but certainly not a risk for average programmer Ivan. Besides, as if Google cared about countries arresting local employees - they have multiple offices in China, from all countries!
Ah, I'm an expert on being expat working in China, and it really isn't much of a risk. China is dodgy, no real rule of law, just rule by law, but they aren't going to go after techies, there is no specific rule that they can go after employees, and the most you are going to get is tit-for-tat that honestly could affect any foreigner in China (e.g. tourists). But Russia made a law saying they could specifically go after employees, and that spooked Google out to moving everyone from St Petersberg to Zurich.
In the same way as Yandex is Dutch and Kaspersky actually used to be British for some time. Where the holding company is registered is not that important these days
That feels very inaccurate, the C/C++ extension for VSCode has 50 million downloads for example. Just anecdotally, most of the developers I know use VSCode for non-webdev things as well. I think you're underestimating the popularity of VSCode in general.
Guess it depends on the business? I work at a Java shop and everyone uses IntelliJ. A few people use VSCode for UI stuff but most also just use IntelliJ.
I use vscode for Golang backend dev work. I even have a subscription to Idea, but it just never took to me. I put in the work to customize Idea to basically be like Goland, and I still just kind of prefer vscode. Also use vscode for more generic things like CI scripts, Ansible, etc.
Jetbrains' approach to IDE's tends to be language-specific. So, taking IDEA (for Java) and making it work for Golang is probably not as slick as it should be.
GoLand[1] is part of the Jetbrains suite, and should be used for Golang development.
I edited it into my original comment, but you can install a bunch of the same extensions that are used in Goland in Idea. I was under the impression that the difference is primarily in the default extensions.
Usually it is extensions - but also keymapping, default utilities, some layout, etc. Basically they structure each specific IDE for what people working in that language would want/need the most.
So, you're kind of going against the grain by trying to make IDEA a general do-all IDE.
This is different than say, Eclipse, where everything and the kitchen sink are under one IDE. This took me a bit of getting used to, but now I kind of like that separation without fear of breaking/worsening one of my "flows" for another lang.
Yeah well the C/C++ extension panicked and crashed when I tried to use it for debugging on macOS so that just about sums up what the extension quality was like.
I sincerely hope it has improved in the past few years since then.
The remote capabilities of jetbrains software is just not comparable at all. I tried again very recently and it was just super backwards. Not as bad as doing straight ssh but not even close to being as useful and ubiquitous as the full editor experience you get on vscode. I even see more and more products offering one click remote access with vscode.
I would pick vscode over any jetbrains tool for absolutely any language. I do all of my Java, scala, and python in vscode. Use whatever tools you like (that’s why there are so many) but vscode is genuinely good for just about everything.
I can’t bring myself to fight with jetbrains insane configurations (what the hell is a facet?!?)… and some very questionable syntax highlighting.
Could you elaborate? I'm in the opposite camp and any time I try to use VSCode with Python, I really miss the Semantic Highlighting there, which helps with orientation in the code quite a lot.
It's one of the extra things you can add on to the base project structure e.g. a Spring facet for Spring framework projects, or an Android facet for Android apps. https://www.jetbrains.com/help/idea/facet-page.html
I get all of my syntax highlighting, autocomplete, integrated terminal, test panels etc in vscode. Vim is great too with the right configuration and tweaking.
I have had the same experience with PyCharm on Win. I miss PyCharm's approach to exploring nested data data structures visually but I don't miss having to wait 10 minutes every launch because it is "updating indexes."
My embedded team is falling one by one to vscode. Main problem is education/documentation. None of us know what or how to do all the great things or even the simple things in CLion. It also guzzles memory. Also name recog: everyone wants vscode on their resume. 3rd, setting up embedded targets is not clear.
:(
I have been using Jetbrains products for a number of years and like them overall but the development over time has been really disappointing. The improvements tend to be very minor each year and usually come with new problems. Many of the changes are just UI reshuffles for no good reason which is annoying, and they haven't been very good about supporting new technologies.
It feels like a company that started off great but then began coasting on their lead and has largely lost focus. The products are still very good for the most part, but most updates end up being a waste of time.
The latest release fixed an open issue I had for 4 years. It was to make the hover for Typescript types look more pretty. They were motivated to finally fix this cause someone else did it for vscode.
I'm just glad they fixed it and their products are the best in the industry.
> they haven't been very good about supporting new technologies
Depends on your language. For me it has been great. Been using CLion for several years now with the Rust plug-in, and recently they began work on a dedicated IDE for Rust :D
Yeah I primarily use DataGrip and Rider. I use DataSpell with the R plugin, and Pycharm is a godsend when I have to program something in python. I don’t do it very often and really appreciate pycharm’s ease of use.
Rider is a great, lightweight alternative to visual studio 2022. And honestly DataGrip is perfection and shouldn’t be touched.
Also use Rider and I don't think there is stagnant development. Their new predictive debugger is a great example of feature improvement.
Only downside is the cost of CPU/Memory, as I do IntelliJ for web development and Rider for C#. Can be quite heavy. But the products themselves make up for it in base features over something like VS Code.
In terms of new features, for me that's ok. However, the amount of bugs in their bugtracker goes up and up. Also, the CPU/memory usage of their tools is a shame.
They're exercising their own tools and library code to produce this. Tools like Datalore (Jupyter competitor) would be next-to-useless if no one at the team uses them to develop some kind of product.
Maybe they shouldn't try to have 12,345 different products then? The core product line isn't really in a great spot right now. I'm seriously considering switching away from IDEA for the first time in a decade due to how bad the performance has gotten.
> The core product line isn't really in a great spot right now
Pycharm -> Best in class
Rider -> As good as Visual Studio depending on your use case can be better.
Goland -> Best in class
CLion -> Best in class
DataGrip -> Best in class
Doing this is just an alternate form of advertising advertising. Post go viral, they get eyes on their products.
Additionally, it's good that the engineers there use their tools for real-life projects that aren't all just building an IDE. For instance python is in this weird state now that for web dev pycharm is great, but for remote capabilities they've been lacking a good work flow. So them exploring and feeling the pains themselves is good.
If I were a broad tools vendor, I'd very much want an internal team to ship something real using the new hotness to make sure we find ways to support it. How else can they understand the needs of a new market?
Except human beings are not optimization machines and the best way to get great people to bring all of themselves to work is to allow them to be all of themselves at work.
The best and brightest have diverse interests and if they get bored they can just move on. The cost to the company to let them mess around with generative AI for a couple of weeks is far cheaper than having to replace them if they jump ship elsewhere.
Plus, this builds internal expertise with a new area of technology that may or may not end up relevant to their business in the future.
Jetbrains is a bootstrapped superstar company, man. They don't have to answer to people like you and me. For the rest:
1. LinkedIn -> public through MSFT
2. Stack Overflow -> public through PROSY
3. Meta -> public
4. Epic -> public through TCEHY
If you want to solve this problem, start a blindingly successful company with wild ARR and then spend it how you want it to. But if you take it public, you best believe that you now have masters that look like you and me. Notice how employee ratings of CEOs correlates strongly to stock performance? That's what shareholders do too. No one knows what's right. They only know the outcome they like, not how to get there.
Personally, I think this is fantastic. They're doing a good job moving the frontier of some of this marketing stuff. That means a pretty good internal org structure that allows the eng guys to contribute to the branding stuff.
> start a blindingly successful company with wild ARR and then spend it how you want it to.
So you mean actually innovating in the space that your products occupy and become market leader there? Maybe in a niche that no one occupied before?
IDE's before JetBrains were complete ass-backwards, no wonder they're so successful.
> That's what shareholders do too. No one knows what's right. They only know the outcome they like, not how to get there.
Maybe don't start out trying to build a product that requires an instant 10 million dollars in cash injection? Before the recession there were YC backed startups on here that offered products that hat no market fit outside the silicon valley bubble.
TSMC started with $58 M at $200 M valuation in the late 1980s. I think it’s like 15% annual growth rate since then and the centerpiece of a nation’s geopolitical doctrine. You can build rocket ships in many different ways. You get all sorts of outcomes.
No one really knows outcome till they give it a shot. People used to make fun of the idea of a software guy building spaceships and cars and he’s doing it better than the incumbents. No one knows what will work till it’s done.
TSMC wasn't bootstrapped in the stricter sense though. Those 58$ million came from Phillips, who also provided the technology and skill transfer for their first manufacturing process.
> People used to make fun of the idea of a software guy building spaceships and cars and he’s doing it better than the incumbents.
Reducing the success of SpaceX to Elon Musks persona alone is, realistically speaking, delusional at best. But everyone is entitled to their own opinion.
Exactly. TSMC wasn’t bootstrapped. It’s okay to build products that require a large amount of capital to start. Many of them are very successful.
As for Elon Musk. Many had access to the labour pool that he did. Many tried to do what his companies did. His companies succeeded. Once could be chance, twice is skill. But to each their own.
It was on HN a while ago though I don’t remember the source. It was 250 M pre-pandemic almost all recurring or something like that going off memory. Founders owned the whole thing.
Let's not sugercoating this. When a company invests in image generation AI, its intention is to replace artists.
And before "but who spend money on this? it's just simple motion graphics my nephew can make in After Effect..." yeah we know you can make Dropbox with rsync in an afternoon.
"Photo from a close-up view of a metallic surface filling the entire frame, with a gradient of blue, teal, purple, and pink hues. The texture is smooth and shows a deep wave, but with few lines, emanating light and creating a sense of depth."
82 comments
[ 2.7 ms ] story [ 168 ms ] threadAre they just bored over there?
Edit: i'm not trying to rag on JetBrains and I generally LOVE their products. But this work seems way over-engineered and not a great use of resources, unless its the graphic designer building all that stuff :)
- Хороший понт дороже денег.
An English equivalent would be:
- Making a splash is better than cash.
Given that Intel and Nvidia had huge R&D centers in Russia until the middle of 2022, I highly doubt such law ever existed. Unrelated, but I also doubt US (as a country) won more than it lost from forcing them to withdraw from Russia, because Huawei literally hired whole divisions with all the expertise and inside knowledge.
https://www.cnbc.com/amp/2014/12/12/google-closes-russia-off...
Russia can arrest local employees if they don’t like what their parent company is doing.
That is a real risk for maybe few very high-level expat managers, but certainly not a risk for average programmer Ivan. Besides, as if Google cared about countries arresting local employees - they have multiple offices in China, from all countries!
Make of it what you want, but the piece of anecdata I have from a person working at Russian Intel office in 2014 is that after Crimean referendum US government unofficially recommended big US companies to cease operations in Russia. Some of them agreed, like Google, because it was easy for them due to a small footprint, some of them, as Intel, deemed it unrealistic back then.
Ah, I'm an expert on being expat working in China, and it really isn't much of a risk. China is dodgy, no real rule of law, just rule by law, but they aren't going to go after techies, there is no specific rule that they can go after employees, and the most you are going to get is tit-for-tat that honestly could affect any foreigner in China (e.g. tourists). But Russia made a law saying they could specifically go after employees, and that spooked Google out to moving everyone from St Petersberg to Zurich.
This thing is decoration. It's useless. And that's why it's fine.
VSCode, while used for other things, mainly is a web front-end IDE.
GoLand[1] is part of the Jetbrains suite, and should be used for Golang development.
[1] https://www.jetbrains.com/go/
So, you're kind of going against the grain by trying to make IDEA a general do-all IDE.
This is different than say, Eclipse, where everything and the kitchen sink are under one IDE. This took me a bit of getting used to, but now I kind of like that separation without fear of breaking/worsening one of my "flows" for another lang.
I sincerely hope it has improved in the past few years since then.
[1] https://plugins.jetbrains.com/plugin/21962-dev-containers
[2] https://containers.dev/
I can’t bring myself to fight with jetbrains insane configurations (what the hell is a facet?!?)… and some very questionable syntax highlighting.
Could you elaborate? I'm in the opposite camp and any time I try to use VSCode with Python, I really miss the Semantic Highlighting there, which helps with orientation in the code quite a lot.
But some people do prefer an actual IDE, as it can give a huge benefit.
And then it became difficult with PyCharm because MacBook Air started to struggle.
Reluctantly switched to VS code few years ago and I have not missed anything since then. Performance was great even then. So is it now.
Have used it for almost everything from Go to Typescript and everything in between.
It feels like a company that started off great but then began coasting on their lead and has largely lost focus. The products are still very good for the most part, but most updates end up being a waste of time.
I'm just glad they fixed it and their products are the best in the industry.
https://www.theregister.com/2011/11/04/bofh_2011_episode_17/
Depends on your language. For me it has been great. Been using CLion for several years now with the Rust plug-in, and recently they began work on a dedicated IDE for Rust :D
Rider is a great, lightweight alternative to visual studio 2022. And honestly DataGrip is perfection and shouldn’t be touched.
Only downside is the cost of CPU/Memory, as I do IntelliJ for web development and Rider for C#. Can be quite heavy. But the products themselves make up for it in base features over something like VS Code.
I'm still angry over the MS Office ribbon.
Pycharm -> Best in class Rider -> As good as Visual Studio depending on your use case can be better. Goland -> Best in class CLion -> Best in class DataGrip -> Best in class
Additionally, it's good that the engineers there use their tools for real-life projects that aren't all just building an IDE. For instance python is in this weird state now that for web dev pycharm is great, but for remote capabilities they've been lacking a good work flow. So them exploring and feeling the pains themselves is good.
The best and brightest have diverse interests and if they get bored they can just move on. The cost to the company to let them mess around with generative AI for a couple of weeks is far cheaper than having to replace them if they jump ship elsewhere.
Plus, this builds internal expertise with a new area of technology that may or may not end up relevant to their business in the future.
- LinkedIn lays off 563 engineering roles
- Stack Overflow is laying off another 28%
- Meta to lay off employees in metaverse silicon unit on Wednesday
- Epic lay off 830 people thanks to “unrealistic” metaverse ambitions
Apparently they don't have such deep pockets...
1. LinkedIn -> public through MSFT
2. Stack Overflow -> public through PROSY
3. Meta -> public
4. Epic -> public through TCEHY
If you want to solve this problem, start a blindingly successful company with wild ARR and then spend it how you want it to. But if you take it public, you best believe that you now have masters that look like you and me. Notice how employee ratings of CEOs correlates strongly to stock performance? That's what shareholders do too. No one knows what's right. They only know the outcome they like, not how to get there.
Personally, I think this is fantastic. They're doing a good job moving the frontier of some of this marketing stuff. That means a pretty good internal org structure that allows the eng guys to contribute to the branding stuff.
So you mean actually innovating in the space that your products occupy and become market leader there? Maybe in a niche that no one occupied before?
IDE's before JetBrains were complete ass-backwards, no wonder they're so successful.
> That's what shareholders do too. No one knows what's right. They only know the outcome they like, not how to get there.
Maybe don't start out trying to build a product that requires an instant 10 million dollars in cash injection? Before the recession there were YC backed startups on here that offered products that hat no market fit outside the silicon valley bubble.
No one really knows outcome till they give it a shot. People used to make fun of the idea of a software guy building spaceships and cars and he’s doing it better than the incumbents. No one knows what will work till it’s done.
> People used to make fun of the idea of a software guy building spaceships and cars and he’s doing it better than the incumbents.
Reducing the success of SpaceX to Elon Musks persona alone is, realistically speaking, delusional at best. But everyone is entitled to their own opinion.
As for Elon Musk. Many had access to the labour pool that he did. Many tried to do what his companies did. His companies succeeded. Once could be chance, twice is skill. But to each their own.
What is Jetbrains estimated ARR?
And before "but who spend money on this? it's just simple motion graphics my nephew can make in After Effect..." yeah we know you can make Dropbox with rsync in an afternoon.
edit: a sample gallery is here http://www.bottlenose.net/share/evolvotron/gallery.htm