14 ms·

Palantir has an outdated software stack (Java/Swing). Their genius lied in creating "mythical image" and cleverly overselling it to the less proficient governme

• jee_accnt 9y ago
  You make it sound like they are not providing any real value, but Palantir has lead to multiple arrests, and some claim, even helped in finding Osama. The complain seems to be that they are overpriced, not useless. Also, there is nothing wrong with a Java/Swing backend. Its not cool, but any company handling that amount of sensitive data will not use a 'cool' stack.
• nulldev 9y ago
  > I can assure you it probably has the worst overconfident brogrammer culture Everyone I know at Palantir has at one point or another mentioned to me (boasted about?) how difficult it is to get a job at Palantir. Sample size < 5, but I'm starting to think this isn't a coincidence.
• joelthelion 9y ago
  > Palantir has an outdated software stack (Java/Swing). How is that supposed to be even vaguely relevant? Palantir sells intelligence, not hip-compliant software.
• phreeza 9y ago
  > Their genius lied in creating "mythical image" and cleverly overselling it to the less proficient government agencies (apparently NSA was smart enough to see through the ruse). A trick now being repeated by Cambridge Analytica. The press eat up the story of "Big Data stole the election", and they do their part acting a bit dodgy, and presto they have that reputation and are probably getting buried in cash. Nobody wants to admit working with them so if they fail to fulfill their promise, nobody finds out. If they do end up on the winning side, the collaboration gets "leaked"... https://www.nytimes.com/2017/03/06/us/politics/cambridge-analytica.html https://www.nytimes.com/2017/03/06/us/politics/cambridge-ana...
• tp3z4u 9y ago
  I have friends that work there and say the same thing. I turned down a job there a long time ago, I did ontologies academically so I already knew they didn't work for what they were trying to do.
  • egwor 9y ago
    What's 'ontologies academically'?
    • criddell 9y ago
      No the OP, but it probably means he did a lot of work in semantic web stuff. https://www.w3.org/OWL/ https://www.w3.org/OWL/ https://www.w3.org/RDF/ https://www.w3.org/RDF/
    • TeMPOraL 9y ago
      I'd shorten that to "what's 'ontologies'"? I did a class on Semantic Web, and spend a fair amount of time reading articles (and a book) about that, and I still don't have any clue what exactly do they mean by "ontology"... It's like the meaning of this word somehow escapes my mind.
    • bane 9y ago
      My guess would be "knowledge models".
  • throwaway825 9y ago
    What were they trying to do?
• nitrogen 9y ago
  Swing isn't actually bad if you have a good look-and-feel. Easier to design a Swing UI than a webapp with similar functionality.
  • Zigurd 9y ago
    The problem with Swing is that everything in Swing looks like a dog's breakfast and nobody under 40 knows how to use it.
    • kodfodrasz 9y ago
      > nobody under 40 knows how to use it. Wow, you must be a grumpy old man, hating those youngsters. They can't even tie their shoelaces alone... Actually I have seen some people under 40, who could use a microwave oven, use a smart phone, and some could even use desktop applications. I believe many of them can find their way around stuff outside the web browser. Regarding developing swing apps as "use swing": I have managed to do it after reading the docs, and I'm under 40. It was actually more pleasant than using HTML+JS.
      • Zigurd 9y ago
        Sort of the opposite, really. I'm one of those old guys who remembers attempts such as Swingweb and GWT to compete with Web UI stacks.
    • nitrogen 9y ago
      The former is fixed by a solid look-and-feel, but the latter is definitely a problem (though I'm certainly under 40, I've never met another Swing dev in person and make most of my money elsewhere).
    • angersock 9y ago
      > nobody under 40 knows how to use it. Reading the docs is a good place to learn. That's how I did it when I was under 40.
    • RUTHLESS_RUFUS 9y ago
• winter_blue 9y ago
  What's outdated about Java/Swing? Java 8 is a fairly decent language, and Swing isn't bad. In fact the entire suite of JetBrains IDEs (WebStorm, PyCharm, IntelliJ, Android Studio, ReSharper, etc) have their UI implemented in Swing. I suppose they could have used Kotlin, Scala, Haskell, or some other more language with a more advanced type system. But it would have been truly horrible had they written their back-end in an awful (dynamically typed) language like JavaScript or something.
  • svendisnigh 9y ago
    The issue with Java/Swing is that it limits development velocity and makes job less appealing for current employees. While HTML5/JS is not without its own problem. Yet using it over a native framework like Swing, enables the development process to leverage external knowledge (Open Source libraries, etc.). This problem is further worsened when your competitors in enterprise Business Intelligence space can leverage those tools, while you are stuck with maintaining the old stack for government customers.
    • jackmott 9y ago
      nonsense
      • svendisnigh 9y ago
        I welcome you to start your next hot startup in 2017 using Swing/Java instead of React/React Native/Node/Ruby/Python/Go/Rust/C++ etc. The merits of the stack are determined by both intrinsic factors and the market. Ignoring the market (remember Palantir is hiring in SV / Palo Alto) you risk alienating potential hires and current employees. [1] Another great example is EHR/EMR market with Mumps etc. [1] https://www.quora.com/Why-does-Palantir-use-Java https://www.quora.com/Why-does-Palantir-use-Java
        • slap_shot 9y ago
          I'm a recent founder a big data platform and a lot of our code is Java. But better yet, we have embraced Cloud Native, so between containers and microservices it doesn't really matter what language we use. We use Java, Python, Go, Ruby, C++ and our front end is React. Individually, every language we use is both powerful and flawed. But together it builds a pretty amazing platform.
• foolfoolz 9y ago
  I interviewed there a couple years ago. the guy forgot to call me for the phone screen. I emailed the recruiter and they said sorry they can reschedule for 7pm that night when he got home. on the call (which started at 7:15) he talked 90% of the time. he would ask me two or three questions at a time, and when I would get a chance to answer one he would jump in and answer the rest of them for me. I was rejected for lack of technical skills
  • rdtsc 9y ago
    Aren't you glad though. That was very quick way to learn about the company. And you have an awesome story to tell. Same kind of thing happened when I interviewing for AWS. Forgot to call for phone screen. Manager didn't show up on day of interview. Forgot about me during lunch. Had to parrot back the leadership principles while doing even sillier whiteboard puzzles. Then "we'll call you next day" as I had a few offers in hand was more like 2 weeks later.
  • cylinder 9y ago
    It's amazing how many mediocre people are being propped up by investors all over the place.
    • mrlyc 9y ago
      I've worked for a number of startups and observed that some people are great at collecting money but don't know how to spend it wisely after getting it, just as some politicians are great at getting elected but don't know how to govern.
      • TeMPOraL 9y ago
        Convincing investors is an orthogonal skill to running a tech company or building a useful product - much like winning an election is a completely separate skill from actually being competent at managing a country. This unfortunate reality is why I try to avoid dealing with systems that developed professionals who are solely specialized at gaming those systems - like MBAs or career politicians.
• ProAm 9y ago
  Technology stacks do not make or break businesses..... It's foolish to think so.
  • imcoconut 9y ago
    Unless the business is a tech company selling enterprise software. If, as other commenters have mentioned, their software is subsumed by cheap or free open alternatives, that could be detrimental.
  • sytelus 9y ago
    No, they often do. Netscape couldn't catch with IE because of their convoluted stack. Friendster couldn't catch with Facebook for same reason from what I hear. If you are building web app in COBOL or Fortran, you are certainly at huge disadvantage with someone else doing stuff with modern more richer platforms.
    • lokedhs 9y ago
      > “Netscape couldn't catch with IE because of their convoluted stack” Wait what? There are several reasons why IE managed to crush Netscape, but the complexity of their software stack isn't even on that list. I was working in the industry when all of this went down, and I can assure you that Netscape's software stack had absolutely nothing to do with it. The actual story is mostly about Microsoft leveraging their monopoly to push competitors out. This is part of their standard modus operandi since the company was formed. If you think I'm wrong, please give some evidence to it.
      • rrdharan 9y ago
        I was also working in the industry at the time. Microsoft leveraging monopoly power was certainly a/likely the major factor, but Netscape also had serious execution problems trying to get out an email server and directory management product.. at the time they could have stolen significant share from Novell. A lot of this is I believe covered in the famous "never do a rewrite" JWZ essay and in Ben Horowitz's book.
      • 9y ago
• jchw 9y ago
  I'm pretty curious about the "outdated tech stack" part of this. I have used some of their open source software, such as blueprint (a React UI toolkit) and pkg (some Go libraries.) Both of which seem quite modern. That's TypeScript + React and Go. I do see some Java stuff here. But it would appear at least they have talent in very bleeding-edge technology stacks. Is this all for show?
• jitl 9y ago
  This dig at Palantir’s stack is outright false. A ton of their “products” are modern TypeScript + React UIs over reasonable open-source datastores, and they open-source plenty of stuff from that stack. Check out this React component, which is pretty cool: https://palantir.github.io/react-mosaic/ https://palantir.github.io/react-mosaic/ Plenty more examples here: https://palantir.github.io https://palantir.github.io
  • xxSparkleSxx 9y ago
    Either way, I'd kill myself before I worked at Palanatir. And I've put myself in poverty before by quitting a multi-national pharma that bought my company out and I lived on my 10k savings over 15months while I found a new job. Palanatir is not harming the world because of 4-5 shitty executives. They are harming the world because of 100's of shitty employees that only want money. Fuck all of them.
    • linkregister 9y ago
      > They are harming the world because of 100's of shitty employees that only want money. Can you explain this more? I don't work at Palantir, but I do insist I get paid for working every two weeks. How are they harming the world? I mean, aside from converting oxygen to carbon dioxide.
      • insulanus 9y ago
        They are saying "don't work for them" / boycott them. They are arguing that if they had no employees, they would cease to be.
        • linkregister 9y ago
          I still don't get why Palantir is making the world worse, besides being apparently a bad product.
          • comstock 9y ago
• hueving 9y ago
  Java is one of the most widely used languages in the world. I get the impression that people who think using Java is an issue are usually just hiding their own inadequacies by repeating hype they read on the Internet.
  • criddell 9y ago
    Would you still use it on the client side today?
    • hueving 9y ago
      It depends on what you mean by a client. If your world view doesn't extend beyond a browser, then no. However, I would still use it in client applications on servers, desktops, and Android phones.
    • CydeWeys 9y ago
      If I were developing a new game from scratch today then Java would be a contender. Look at the success of LWJGL for instance -- it's what powered Minecraft up until the Microsoft acquisition. Other than playing around with that, I don't think I've done anything substantial with Java-based UIs running client-side since 2003 :/
    • StevePerkins 9y ago
      Assuming that I can't get away with a web app, and need to install something native on the client side... is cross-platform support required? If so, then your choices are: 1. Write separate native clients for Windows (in C#) and OS X (in Swift or Objective-C), and perhaps Linux (in whatever). It'll be hard to hire people who can work on all three, so you'll need more headcount and organizational/technical complexity. 2. Use a cross-platform framework like Qt, Wx, Gtk, etc. This will probably push you toward C++, which can be difficult to hire for. And you'll get constant complaints that the UI doesn't "look native enough" anyway. 3. Use Electron, and whichever JavaScript SPA framework is trendy this week. Have everyone on Hacker News declare your client to be "Flash for the Desktop", and be forced to re-write it with the next trendy framework two years later so that anyone will want to work for you. 4. Use JavaFX, and the JDK's built-in "javapackager" tool to create native installers for Windows, OS X, and Linux. Each of which includes a bundled JRE, with a size smaller than Electron. You won't be able to find a lot of JavaFX experience on the market... but Java developers are easy to find, and it only take a few weeks for decent ones to get up to speed with the framework. Hell yeah, I'd use Java on the client side. None of the options are great, but Java's probably the least-bad.
• tritium 9y ago
  outdated software stack ... no meaningful ML/AI strategy Yeah, because not being buzzwordy enough is just SO much more important than having meaningful products and services that fulfill a vital role.
  • warent 9y ago
    What exactly is "buzzwordy" about "software stack" or "ML"? This is legitimate jargon that means specific things. We use these terms routinely at Google internally.
    • sgift 9y ago
      In context with 'strategy' it implies you need a "ML strategy" to somehow deliver value to customers. That's buzzwordy.
      • mattacular 9y ago
        They weren't saying every company needs an ML strategy they were saying Palantir does.
      • CydeWeys 9y ago
        In the context of the business area that Palantir is in, though, they absolutely do need machine learning to maximize the value of their product. Imagine trying to build a program to beat AlphaGo, or a search engine better than Google, or a best-in-breed image classifier, without using machine learning. You can't do it. Machine learning is a critically important technique that can solve problems better than anything else can. Palantir's work contains many such problems. Not using the best tool for the job, and not even having a roadmap to do so, means they're ripe for disruption by some other company coming along doing the work better. Machine learning isn't just some buzzword used to attract VC money. It really, truly, does work.
  • jsmthrowaway 9y ago
    ML is extremely relevant to what Palantir does, particularly on the broad-scale analysis stuff they purport to do. Extremely relevant. It is surprising to read that they don't have a strategy -- not weighing in on whether that's accurate or passing judgment, just saying it's a more surprising observation than you're implying.
• throwaway055 9y ago
  I've heard Palantir is maybe the most credentials-/prestige-focused tech company in Silicon Valley. I remember thinking that sounds like a good way to build a below-average team that seems very impressive from the outside. When you look at what they're doing, this strategy actually starts to make a lot of sense. They're selling a product that's somewhat incomprehensible. The value of that product is very difficult to measure, due to both its complexity and the "how do you put a dollar figure on social impact"-factor. They're selling that product to government agencies, which probably can't properly evaluate that kind of purchase and don't care much about cost or ROI (due to government's inherently inefficient incentive structures). Credentials matter for winning contracts, because if you're the person who makes that decision, credentials and prestige are something tangible you can point to and prove you did your job. This doesn't sound like the makings of a great respectable tech giant, though. This sounds more like one of those companies built specifically to exploit inefficiencies in government spending.
  • nnq 9y ago
    > credentials and prestige are something tangible you can point to and prove you did your job Like in "nobody ever got fired for choosing IBM"? Seems like they've just copied a page from an the old book... and this still works, and lots of more competent startups who could do their jobs better will not get those contracts.
    • rapind 9y ago
      There are also probably lots of startups that would create a complete mess of it. Picking / evaluating vendors is hard and super risky for the average manager. While I personally would seek out more nimble vendors, I certainly understand the attraction to established shops.
  • fergie 9y ago
    Yup- when you start thinking of Palantir as an IBM, a Cap Gemini or an Accenture, it starts making a lot more sense.
    • 9y ago
• sytelus 9y ago
  If you want to write cross platform desktop apps, your choices are pretty much either Java, C++ and Python. Rest of the languages/platform either don't have great library support or simply not mature enough to deploy at scale. Now if you want some performance, more mature UX library and eliminate Python, you are limited to C++ and Java. Its much more harder to do UX stuff with C++ and finding great C++ developers has its own challenges. So usually it all comes down to Java. Of course, one can argue that why in the world someone is trying to build desktop client applications in 2017. Usually answer to that question is Internet is not as universal, cheap and fast as some people tend to believe.
  • drxzcl 9y ago
    I have extensive experience writing professional (i.e. Not consumer oriented) desktop applications in both Python and C++, and you're right on the money. Java may not be sexy or purty, but it works and you can actually hire people to do it.
• dajohnson89 9y ago
  Are you calling Java outdated? Or swing? Either way I think that's quite debatable.
• downandout 9y ago
  >Palantir has an outdated software stack (Java/Swing). This is the first fallacy of software engineering. In general, nobody cares what the tech stack powering a solution to a given problem is. If it solves the problem, then it is considered a success by customers. There are exceptions to this rule, but there aren't many, and they don't really apply to Palantir's model. In their case, as long as the system spits out the data it is supposed to, they could have used Visual Basic to program the backend and customers would still buy it if it worked as advertised.
  • hluska 9y ago
    What about recruiting?
    • user5994461 9y ago
      It's one of the most used and most known languages. There is no issue recruiting. Better yet, it's been widely used for a while, so there is no issue recruiting experienced developers.
      • hluska 9y ago
        I wasn't talking about Java in particular. Instead, I was focusing on this sentence: In general, nobody cares what the tech stack powering a solution to a given problem is. When I choose a tech stack, I'm always concerned about recruiting. Some stacks are easier/cheaper to recruit in than others.
• mattmanser 9y ago
  Java/Swing isn't even a data format, it's the interface tech. Nor is it outdated. Like, if you'd said COBOL it might be a valid point. It's as ignorant of basic technology as saying my car runs on a particular type of rubber so I can't use batteries. Honestly, for a tech site I'm ashamed this is the top comment, just seems to be an excuse to rag on Palantir and decry not using the present FOTM.
  • CydeWeys 9y ago
    I moved from a "sexy" NodeJS stack at my last job to a staid Java 7 stack at my current job (we can't move to Java 8 until App Engine supports it). The code at my current job is much better written, is more thoroughly tested, and is more pleasant to continue development on. The codebase at my previous job was a rat's nest, caused primarily by lower-skilled developers, no time allocated to technical debt or refactoring, a breakneck pace of product release, lack of a good automated testing suite, lack of good continuous integration and presubmits, and the weakly-typed/dynamic nature of JavaScript itself which doesn't scale well to larger server-side codebases. So long as a given language/platform isn't deprecated (and Java definitely isn't), I'd rather work on a solid codebase in it than a weak codebase in a more alluring, recent language.
    • itcmcgrath 9y ago
      FYI, Java 8 is now in beta for App Engine Standard: https://cloud.google.com/appengine/docs/standard/java/quickstart-java8 https://cloud.google.com/appengine/docs/standard/java/quicks...
      • CydeWeys 9y ago
        Oh, I'm well aware. Unfortunately it's been in beta for awhile now, and keeps slipping deadlines. We can't use it until it's out of beta because we run an important production service with external contractual obligations and we need guaranteed support for things working. The "no SLA" policy on the beta makes it a non-starter for us.
    • pyre 9y ago
• throwaway825 9y ago
  > Palantir has an outdated software stack (Java/Swing) What? How is this even correlated to the value provided by the software they sell? So they're SOL because they didn't do a rewrite? > they have no meaningful ML/AI strategy They mostly sell software and services around data integration of disparate data sources and the ability to query that information and make connections easily. Their flagship product is a graph database with access controls built in. ML/AI is so dependent on the particulars of a given data, so I don't know how valid this criticism is. If they're able to provide value without productizing machine learning, sounds good to me. > Finally the political environment along association with Thiel isn't helping them make any new friends. Agreed
• bane 9y ago
  In support of your comment on ML/AI...quora has a post on that https://www.quora.com/How-is-the-state-of-ML-in-Palantir/answers/32457283?srid=iykg https://www.quora.com/How-is-the-state-of-ML-in-Palantir/ans...
• bitmapbrother 9y ago
  >Palantir has an outdated software stack (Java/Swing). Sadly in 2017, their tech stack is falling apart, they have no meaningful ML/AI strategy If I were to guess I would say you failed the interview. You should probably brush up on your Java and Swing.
  • civilitty 9y ago
    That sounds incredibly defensive. Not one substantive response to any of OP's claims, huh?
    • bitmapbrother 9y ago
      When you open with a substanceless proclamation that Java is outdated anything further must be taken with a large grain of salt.