mutually orthogonal communities : Nonlinear Function
Created: July 10, 2020
Modified: May 22, 2021

mutually orthogonal communities

This page is from my personal notes, and has not been specifically reviewed for public consumption. It might be incomplete, wrong, outdated, or stupid. Caveat lector.
  • This was originally a section of breakup.org, written several years ago.
  • this is more related to jobs and identity, but for cases when I get jealous or feel inadequate about my work, I think it's worth making a list of different communities, all with important priorities, all of which have some level of (implicit or implied) disdain for the others. so when I feel inadequate because I don't live up to some 'local' notion of status (e.g., in grad school, not publishing a lot of papers), I can remember that those notions are only local and real status/self-worth comes from within.
  • academic "applied" AI research: vision/robotics/NLP
    • disdain for PhdAdvisor-type work. (Evan Shelhamer still thinks PhdAdvisor does logic research)
    • disdain for industry (industry "sponsors" are all just trying to get close to the BAIR research glory while working on uninteresting special cases)
  • theoretical academic CS research: learning theory, Mike Jordan stuff, and complexity/algorithms
    • not necessarily disdain for stuff "lower" down the stack, but a lack of connection or understanding. like NameRedacted wants to understand the universe at a deep mathematical level, and doesn't see tech jobs as satisfying.
  • probabilistic modeling and inference research
    • no one in AI gives a shit about MCMC or probprog work, to a first approximation
    • even I think most probprog "research" papers are uninteresting engineering
    • meanwhile probprog people think of everyone else as doing special-purpose hacks
  • CS systems research
    • no one in AI thinks systems research is interesting at all
    • it's all about doing things we already know how to do, not adding fundamental new capabilities
  • AI safety / AGI research
  • tech startups and startup culture
    • tons of people worship YC, and other VCs
    • but e.g., larry page would look down on them as relatively tiny and niche
    • meanwhile academia fautls them for working on trivialities
    • and they fault academia for not building anything real/useful
    • and they fault big companies for being slow/stuck/
  • big company / managerial status hierarchies
    • once you're in the hierarchy, you think it's important. being a higher-level manager is 'more successful' than being an individual contributor.
    • CEOs of bigger companies are more successful than
  • finance
    • everyone in tech has profound disdain for finance. it has no social purpose, no broader knowledge gain, is basically zero-sum and driven by greedy money-grubbing.
    • meanwhile, those in finance know that it's about creating value. It's about wise investment of society's resources, of seeing the big picture. It's about knowing how the world really works. It's about developing market mechanisms and incentives, which are some of the most power social tech that we have.
  • THE REAL WORLD
    • black-swan events happen. in case of nuclear war or global conflict or alien invasion or massive pandemic or whatever crazy thing, none of these status hierarchies will matter.