1. 18
  1. 10

    I can’t name a single thing from Google X or Amazon 126, Apple or Microsoft Research

    Things from Microsoft Research I remember off the top of my mind: F*, lots of work on Haskell and GHC from Simon Peyton Jones.

    And, well, .Net! Whatever you think of MS’s approach to developing and marketing it, it’s a much more modern VM than JVM, and its languages brought many functional programming concepts to a wider audience than ever.

    1. 3

      Isn’t typescript a MS research thing? Singlehandedly fixed many of the problems with JavaScript, IMHO.

      1. 1

        Nope https://www.zdnet.com/article/who-built-microsoft-typescript-and-why/

        Funny thing is that article mentions the original Typescript team worked with an Office365 team for unspecified reasons. I think that’s referring to the people in Office who built a typed JS compiler that relied on typing in comments (this was well before Typescript’s release). I heard a rumor that Typescript was started from that codebase.

      2. 3

        Z3 SMT solver also comes to mind.

        1. 2

          Isn’t F# also from Microsoft’s (UK) research division?

          1. 2

            Also the Xbox Kinect, which was very popular when it first came out.

            1. 1

              And also started a robotics revolution: https://www.wired.com/2011/06/mf-kinect/

            2. 2
              • Midori
              • Lean
              • mimalloc
              • The entire field of functional reactive programming

              … what a wild hot take

            3. 5

              I love Bell Labs as much as the next person, but this is basically what today would be considered HR stuff and isn’t on-topic.

              1. 10

                Idk, I found it kind of compelling to see all these faces. For every Doug McIlroy there are N number of people doing support or other roles that are generally historically invisible but made the organization what it was. Plus it’s a good reminder how huge Bell Labs was and how Unix was just a fraction of their projects.

                  1. 3

                    Ultimately, good managers and HR are what keep researchers at your organisation in the first place. HR can singlehandedly make and break an organisation by how they handle events that come up and how they decide to manage the place. They are perhaps more important to the successful replication of an environment of Bell Labs than the researchers themselves – we can never have those exact brains and minds again, we won’t even have the same melting pot of ideas as we once had, however we can work on facilitating new spaces that work as well as the old ones with our new brains and melting pots. Facilitation is extremely important and an immensely complex and difficult task, and extremely noticeable when it is done badly

                    1. 2

                      That’s true but I don’t see that much unique about their organization chart though, this seems pretty similar to other R&D orgs from that period. It seems flatter than some charts I’ve seen, but flat organizations isn’t some mysterious advantage. I wish there were more here on how business was actually conducted.

                  2. 3

                    Bell Labs is cool and all but the article credits them with the Big Bang theory. Bell Labs is actually credited with discovering the afterglow. Georges Lemaître, a Catholic priest, is the first to theorize an expanding universe from a single-point, and credited such.

                    1. 2

                      I thought this was interesting to see. Research organizations are inherently pretty tricky because it’s not really clear if someone’s work is valuable or not, often until years after it is published. In academia people win grants and publish, but those might not happen as much in an industry research organization. So it’s cool to see the human faces who must have been incredibly motivated and also capable of distinguishing wheat from chaff.

                      Microsoft Research is probably the most similar organization today. Just one group there developed Paxos and TLA+, which are both fundamental to widely-used tech today. While there are larger industry tech research organizations in the US, they are mostly more secretive and don’t publish much of their work.