- Generating revenue from customers is about more than just creating a great product. You also have to reach lots of customers and convince them of your value. Many naive idealists think only product matters (or should matter), and neglect distribution. But most people eventually come to understand that both are necessary, and that this is practically a law of physics, not something to moralize about. (FWIW JetBrains is quite good generating revenue, and I'm fairly certain their revenue dwarfs that of Cursor and Windsurf.)
- Whoever is paying you is your customer, no matter what alternative word we use for it. If you're an employee, your customer is your "employer." If you're being acquired, your customer is your "acquirer."
- In most cases, acquirers are playing the role of investor. Investors value returns. If you want to provide value for an acquirer, then, you need to convince them of the future value of your business should it be acquired. That's usually best done through growth trajectories.
- It's perfectly valid to continue generating revenue year after year without being acquired for eye-watering sums. It's a waste of your emotional energy to become jealous or indignant when others get acquired or succeed with less work. Good for them, just keep doing you. That also goes for the rest of us in the peanut gallery. We don't need to attack recent successes to defend the honor of our favorite incumbents.
Almost everyone here is providing business value in service of these rules of the universe. Those who aren't in cost centers probably need to reflect on this reality more.
Seeing young companies which are pre-revenue, which are competing for an unproven yet crowded sub market (AI coding assistant) out value an established incumbent in the larger space does not compute.
Add to this the fact that there is very little moat for AI coding assistants. Assuming the market as a whole proves itself, there is a very good chance that the winners will be the established incumbent IDEs who can add AI assistance as a feature in their established products.
All of that is to say, current AI valuations in this space look a lot like a bubble.
Which suggests that your framework is lacking.
Here's where:
1. You're neglecting to look at the differences between the fast-rising stars and the comparable incumbents, and instead you're assuming that the incumbents automatically represent a ceiling. In this particular case, JetBrains obviously isn't the most ambitious company on the planet, and isn't focused on hyper growth. There are plenty of avenues for AI IDEs to grow and expand their revenue that have yet to be explored.
2. You're overestimating the importance of concrete moats. Google had no concrete moat either. Just because people can switch easily doesn't mean they necessarily will.
3. These companies aren't pre-revenue. I believe JetBrains is making something like $400-$500 million dollars a year, after 25 years. Cursor is at half of that in just 2 years. Windsurf is also doing big numbers.
4. Related to #3, you're underestimating growth trajectories.
5. You're leaving out the context. Companies that can afford to make $3B acquisitions (a) have tremendous war chests, and (b) have extremely ambitious goals. They're not looking to build the next JetBrains, they're looking to join the pantheon of $1T companies. Achieving massive 10x or 100x or 1000x growth as an investor/owner requires making asymmetrical bets -- bets where if you lose you're still okay, but if you win, you win big.
ie it's a mistake to estimate the potential upside by looking at the size of the current developer market.
As you move you're coding tools towards a less technical customer base - there are two synergistic effects - your potential customer base is much much larger, and they are simultaneously less technically competent on average ( and so less likely to build their own tools if you charge too much ).
Whether those assumptions are true - time will tell - but I definitely see more people thinking software development is now accessible to them.
That is none-sens.
If that were true, all profits would go to the customers. No, you are hired because the company and its owners can extract the profits from your work. Laying you off could mean nothing to customers, but even more profits to them.
Why? What does that mean?
> No, you are hired because the company and its owners can extract the profits from your work. Laying you off could mean nothing to customers, but even more profits to them
Nothing you said here changes the fact that your employer is your customer, paying for your services. If you don't see it that way, it's to your own detriment.
- Russ Hanneman
> And who's worth the most? Companies that lose money.
The next couple of lines:
"Pintrest, Snapchat, no revenue. Amazon has lost money for every fucking quarter for the last 20 fucking years and that Jeff Bezos is the king"
They also have a pre-built customer base to sell them to.
I do think they have a perception issue with devs whose perspective on their products was crystallized back in the 2010s when they were using some old company laptop with 8GB of RAM when they could feel too heavy. With a modern laptop I just don’t care at all if my IDE takes up a few gigs of RAM.
JetBrains also ranks well on things like low-latency input, which surprises a lot of people. They do seem to care about developer experience.
Granted, they're not the worst offenders. When I read that Jira has been updated, I need to work up the courage to look at it because I expect it to just be worse on every level.
My workflow is somewhat Byzantine—mostly just use shells and basic tools like find and grep do most of what IDEs do (sure somewhat worse).
That and I copy and and paste from my favorite AI chat and that’s it. Paste a code block, or an entire file for context.
Like taking notes with a pencil and paper—which helps information uptake—I believe it’s actually important to slow down and take a moment to think.
If you have fully memorized all of Linux and GLibC and all your supporting libraries then yeah, I guess you don't need powerful tools.
Since I dont need updates every 5 minutes, I disable the update checker if on Windows, and just don't update my software stack all of the time as it inconveniences me.
> I wish it was a more stable product
https://youtrack.jetbrains.com/issue/PY-72046 https://youtrack.jetbrains.com/issue/PY-79448 https://youtrack.jetbrains.com/issue/PY-78077
I can't get how poetry support problems were introduced. They have made new patch release just before Christmas holidays which introduced major problem (IDE periodically updates package lock file). It took them 3 months to fix it. After a few weeks they reintroduced the bug and said that it is feature and added registry option to revert to old functionality. I still do not understand what is the reason for updating package lock file.
I almost always file a report, many of which remain unresolved. In several cases I've reported bugs that are promptly marked as a duplicate of a years-old bug. For example:
* The Go plugin appears to make the IDE is very slow (>10 seconds) at detecting files changed outside the app. https://youtrack.jetbrains.com/issue/GO-18510/IDE-is-very-sl... (1 month old now)
* The file modification icon gets stuck (saving doesn't clear it), possibly because of some Git state tracking. https://youtrack.jetbrains.com/issue/GO-18094/File-modificat... (3 years)
* The "next/previous error" navigation shows the line error but then immediately hides it again. https://youtrack.jetbrains.com/issue/IJPL-156089/Next-Highli... (11 months old)
All three relate to what I would call core functionality, rather some rare, niche language feature. They all disproportionately affect my daily use. How can a serious company let bugs like these hang around, not for one week, or two weeks, or a month, but three years?
My guess from the amount of churn in YouTrack is they're churning out too many features creating more bugs at a higher rate than they can fix them, and are simply overwhelmed, and their triage process must be deprioritizing a lot of basic stuff.
My most recent bug: After upgrading to 2025.1, the IDE freezes completely 1-2 times a day, requiring a force quit. I reported the bug 10 days ago, collecting all the output they ask for in such cases, and all I've gotten are crickets. https://youtrack.jetbrains.com/issue/IDEA-371932/IDE-freezes...
Java/Kotlin is their main thing, and yet neither Maven nor Gradle builds are stable. If your build fails or there are some unresolved dependencies, you restart IDE in hope it works...
AI coding tool trial failed for me -- IDE told me it's not activated even after I activated it on billing portal. And doc implied it might take some time. WTF. Does it take some batch processing?..
People who were able to get AI coding tools working said it's way behind Cursor (although improving, apparently).
Aside from their initial AI plugin rollout fiasco it has been smooth sailing for me.
And if "If your build fails or there are some unresolved dependencies" you check your dependencies and config.
I'm tired of people complaining and not trying to understand how their systems (or an IDE for that matter) work.
Because JetBrains products DO have issues, but rest assured, the things you are complaining about are on the main path of basic features they take care of the most.
Source: at first reluctant but now happy IntelliJ user, after thinking for a long time that Eclipse/Netbeans would be better. I was wrong.
To be fair with OP, I had a similar experience in the previous company. Sometimes after bumping a dependency or something, even if you asked IntelliJ to "Reload All Gradle Projects" something wouldn't work and I needed to restart the IDE. Not saying it was common, but it did happen. This was a Kotlin codebase.
Now working in a company using Scala, I had a few cases where IntelliJ failed to resolve a dependency correctly so it was saying that the class I just created didn't exist (this was in shared library and two different dependencies pulled it, but we did pin the dependency directly and sbt was building it correctly, it was only IntelliJ that was becoming confused). Cleaning up the build sometimes worked, sometimes didn't, but even when it worked it would go back after a while. Eventually we updated all dependencies in all projects and it now always works, but it was painful for a while.
But I can tell you IntelliJ has way more issues than e.g. Visual Studio I used 20+ years ago.
Possibly an interesting data point is that my company pays for every engineers’ Cursor usage, can’t imagine how much it could cost, but they don’t have any encouraged integration with JetBrains… so while JetBrains products are good, I’m wondering if Cursor simply has a better sales team and hype pushing them to higher valuations
Which one?
Last time I checked JetBrains' AI tool and it was laughably bad compared to Copilot. My bar was quite low already as I hadn't even used Cursor by the time.
Edit: What I tried is "JetBrains AI Assistant". I haven't tried Junie yet.
> With a modern laptop I just don’t care at all if my IDE takes up a few gigs of RAM.
If Jetbrains reimagines their settings persistence story, I'd be happy to give that a try since otherwise it's quite great still!
There is huge utility in LLM outside generating complete working code.
I think there will be also wave of private LLMs fine-tuned on corporate data, and it will be also good tools.
Also AI should not be lumped together with literal fraud, that's lazy.
Unless you prefer to think of the AI field as representable by a bunch of Indians actually behind software, as the sibling (insincerely and again lazily) reduces it to.
Do most people even know they're a Russian company? Do businesses decide not to invest for that reason?
Many graduate into criminality, but it's not required.
It seems .NET already has a really mature AOT, I'm really hoping .NETs AOT reaches the point where all .NET code can be AOT'd someday.
I feel like Kotlin could do so much more, but its stuck in standstill. There's even some language features that are still missing such as Inline Classes, Pattern Matching, and even Reflection, all things that Java supports directly.
Kotlin did not replace Java, except on Android. So JB now has a beast they have to feed without an enterprise revenue stream. They do get secondary benefits: they use it internally, etc.
We examined Kotlin in detail for a CLI app and based on conversations with Kotlin developers concluded that it was not sufficient of a Java replacement for us to evaluate further. For those in a similar situation, the greatly increased cadence of Java releases has probably permanently foreclosed Kotlin-qua-enterprise language.
Reflection analysis can (and will) be improved but there are hard constraints - a correctly working expression like 'someAssembly.GetType(Console.ReadLine())' by definition would have to root (and force compilation for) every type in the assembly, which is highly undesirable or even sometimes unfeasible for AOT compilation. And there is a lot of code which does exactly this.
The main challenge are packages and frameworks. ASP.NET Core is largely compatible (via minimal API) and so is AvaloniaUI, EF Core has some compatibility assurances and DapperAOT is tailor-made as the name implies, serialization is also a solved problem although you may need to use a different API.
At the end of the day, NativeAOT is not something "to be fully migrated to" because it has fundamental restrictions (some of which also affect other languages like Rust or Go) and having JIT around is a feature for patterns which specifically exploit it but is also a performance optimization (DynamicPGO, better instruction selection especially around SIMD paths, turning static readonly's into JIT constants and apply subsequent optimizations on top of that, this is what makes C# port of Mimalloc so good as it elides dead code with assertions impossible to remove dynamically in C/C++). NativeAOT has its own optimizations, and it will continue to diverge with JIT (e.g. there's a toggle in .NET compiler to repeat some optimization phases, usually it's too expensive for JIT but for AOT it's a good fit, AFAIK there is work to productize this).
The wide perception that JIT-compiled code has to be slower stems from other sources of performance overhead that are typical to languages which happen to use JIT (many of which have "weaker" compilers too), not from the JIT compilation itself. There are technicalities like certain calls have to be indirect in order to support patching the callee address, or inter-procedural analysis which is trivial to prove under AOT may not be so under JIT where new callers/callees may be constructed dynamically or a reJIT invoked which would invalidate the analysis results. JIT also costs additional memory. But it's not a source of worse performance.
Remote SSH is terrible too, handles network latency spikes by repeating keystrokes. I remember spending an evening trying to fix something in the integrated shell and giving up, but sadly forgot what. I like what they do with Go though. Anyway, back to nvim here, not for me.
The valuations are based on trillionish dollar companies fighting over startups.
Honestly it's a little odd JetBrains doesn't seem to be chasing this fad much at all.
Also, Claude Desktop can be configured to serve Jetbrains MCP Server, which will let Claude Desktop (or any other coding AI/LLM) connect and control Jetbrains IDEs, including changing project configuration, listing / finding files, editing files, looking at VCS diffs.
So I believe Jetbrains is addressing the AI coding assistant market, they're not making as much noise and perhaps they should be ... feature-wise I think Jetbrains IDEs + AI integrations will be as good, in the long term, as other systems. At least I hope so, because I can't let go of PyCharm, Webstorm, IntelliJ IDEA, Goland, et al
Jetbrains isn't a silicon valley startup and isn't raising money, so no VC is going to make a 10x return by hyping them. That sadly usually means that you are shut out from the conversation no matter how good your product is.
However that is a load bearing if.
It would be like investing in a horse-and-buggy whip manufacturers around the turn of the century.
They can't invest in AI the amounts that Microsoft can, of course.
In the worst case you’ll still be able to use AI tooling similar or equal to what you have in VS code. Even Windsurf has an JetBrains plugin.
Just vibe code away and never look at the source it’s generating.
That or they just drank the Kool aid
The same was true for Web3, Crypto, Machine Learning...
What actual value does bitcoin have? As an example, gold can be used to make high fidelity cabling.
I'm not being sarcastic at all. I feel I need to reiterate that because you got it wrong.
If you look at previous fads, all the way to be dotcom boom of late 1990s, this very approach seemed to work well for a number of buzzword-compliant pre-revenue "businesses", and their founders / owners. There was a crash after that, but the wisest were able to shield some of the money from it.
But then there’s another side where the fad has real value and the people investing have money because they’re smart and they recognize the value. And you have to be to make value in the fad.
But on some level, it’s also stupid to pursue the fad because it’s so hyper competitive and effect of luck is going to be magnified.
All these people with money didn’t all get it by being stupid or fddy or vaporware so they’re responding to signals of value and their responses are reliable indicators in general.
But on some level, it’s also stupid to pursue the fad because it’s so hyper competitive and effect of luck is going to be magnified.
- but 'luck's meaning is nuanced and depends on how you define it.
The problem with 'luck' is it's a rough model. A hypothesis based on incomplete information or insufficient analysis or non-comprehensive insight. Was it luck? Or was it something else? It's too easy to falsely ascribe significance to mere randomness (both mistaking randomness for coordination, and mistaking coordination for randomness), as I think many people consider luck. Almost as if "luck is a lottery". The wrong view of luck can confuse one's view of agency.
Also, because in the common conception 'luck' deals with the unknown as well as achievement, debates on luck conjure political and religious conflict that can be super contentious and feel intensely personal - essentially intolerable differences for many!
I don't see luck that way - but I agree that luck, in how I view it, is needed!
What is your view of what "luck" is, in detail?
Additionally, network effects and experience input and VC concentration in Silicon Valley are very real and very rational. For VCs, why go anywhere else if the best of the world are flocking towards you already?
This might change in the near future, but I doubt it.
Very false. A lot of the best in the world have no intention of ever living in silicon valley. In my circles people dread even a week long trip there.
Atom was an editor made by GitHub that competed with Sublime.
After creating Atom, GitHub pulled the editor guts out of it and initially called it "Atom Shell". https://github.com/mapbox/atom-shell
This then had a name change of Atom Shell to Electron to decouple Electron from Atom (the editor). https://www.electronjs.org/blog/electron
Microsoft built several key tools on top of Electron (VSCode being the relevant one here) and became very interested in maintaining control of it... and so bought GitHub when it was up for sale.
Eventually, Atom was sunsetted. https://github.blog/news-insights/product-news/sunsetting-at...
> Atom has not had significant feature development for the past several years, though we’ve conducted maintenance and security updates during this period to ensure we’re being good stewards of the project and product. As new cloud-based tools have emerged and evolved over the years, Atom community involvement has declined significantly. As a result, we’ve decided to sunset Atom so we can focus on enhancing the developer experience in the cloud with GitHub Codespaces.
> This is a tough goodbye. It’s worth reflecting that Atom has served as the foundation for the Electron framework, which paved the way for the creation of thousands of apps, including Microsoft Visual Studio Code, Slack, and our very own GitHub Desktop. However, reliability, security, and performance are core to GitHub, and in order to best serve the developer community, we are archiving Atom to prioritize technologies that enable the future of software development.
But overall I hope you’re right, I really want great agentic coding in PHPStorm.
I have PTSD from accidentally opening CLion or PyCharm. Fans starts spinning and there is dozens of seconds wait to close this thing down.
I can't really comment on whether they are good or not since from my perspective I see people who are productive and unproductive using both of these tools, so to me it just looks like mostly a matter of preference. But newer developers don't seem to like these established IDEs and see them as you said: big, slow and laggy.
Some people like slow IDEs, because that gives them time to have a cuppa or browse Reddit when the "index is updating". I have no time for that.
IWhen i open a file /project from a collegue that uses VS Code is filled with errors and warnings because their VS code text editor is not actually understanding the code they are editing.
I do not use those but I believe you, in all N-2 dimensions