Andy Grove's Paranoia: Universal Lessons (1996)

(hbr.org)

72 points | by tosh 1802 days ago

8 comments

  • firecall 1801 days ago
    If you've never read Andy Grove's book "High Output Management" then you are missing out. It's one of the only management books I'll ever recommend. Truly great practical advice.
  • gkolli 1801 days ago
    Pairs very well with Grove's dictum of Objectives and Key Statements (OKRs). These and Grove's management philosophy are captured very well in John Doerr's book 'Measure what Matters'
    • godelmachine 1801 days ago
      Second the book - “Measure what matters”
  • sandymcmurray 1801 days ago
    I toured Intel around this time and attended a press Q&A with Andy Grove. It was amazing. He was lively and fun and thoughtful. We received a printed copy of a booklet entitled WIRLII - What's It Really Like Inside Intel. It covered corporate culture, which was very Andy-centric, including the habit of starting meetings exactly on time (and not allowing late people in). Powerpoint was widespread, but our tour guides used low-tech "foils" (aka clear sheets on an overhead projector) for the presentations. Fascinating guy.
  • lifeisstillgood 1801 days ago
    Swardley's landscape mapping is a strong replacement for the overall mapping idea (Porter through Grove) well worth looking up
  • rdlecler1 1801 days ago
    I haven’t read Grove but I find a lot of these management books just okay. Any tough critiques from HN?
  • martin1b 1799 days ago
    So to sum up. Motto for successful management : Jobs - "Stay hungry" and Grove -"Stay paranoid".

    Sounds like Andy Grove was a work horse.

  • darkerside 1801 days ago
    So much of this holds up, and then you get to RISC vs CISC. It's amazing Intel has managed to survive and even thrive after failing so hard to adapt to the changing environment.
    • Taniwha 1801 days ago
      Intel were in a sense lucky, their x86 was just about the most RISCy of the contemporary CISCs - opcodes are (almost) all one memory operand, no indirect memory modes, no autoincrement modes (except for stack) etc etc which means instruction retry on things like page faults are easy because of few side effects
      • ghaff 1801 days ago
        And they were basically able to transition to chips that were architecturally pretty much RISC while maintaining CISC compatibility.

        Furthermore, they did bet too much on frequency with x86 and compile-time optimization with Itanium, in part because a major partner was so focused on single thread performance. But, although AMD had some short-term success with Opteron, Intel was able to retrench and win out. It continues to have challenges in mobile and other areas but is still more successful than not.

      • darkerside 1800 days ago
        Very interesting. It's lower level than my experience, so thanks for sharing.
    • AmVess 1801 days ago
      How exactly have they failed? They are making more money now than they ever have.

      They are behind on tech right now, but they are still the global CPU juggernaut.

      Failure isn't a label I would attach to them.

  • notsosmart 1801 days ago
    Here is a more readable version https://hbr.org/1996/11/inside-intel