Comment on If AI was going to advance exponentially I'd of expected it to take off by now.

• agamemnonymous@sh.itjust.works ⁨3⁩ ⁨days⁩ ago

  They grow proportionately to ax^n . Correspondingly, for values of x < 1, they look very similar to a simple library slope. For values of x > 1, they grow very rapidly. Both portions are part of the function, it doesn’t suddenly “become” exponential at the rapid increase, it’s exponential the whole time.

  source

  • Clinicallydepressedpoochie@lemmy.world ⁨3⁩ ⁨days⁩ ago

    Well there it is

    it doesn’t suddenly “become” exponential at the rapid increase, it’s exponential the whole time.

    source

    • agamemnonymous@sh.itjust.works ⁨3⁩ ⁨days⁩ ago

      What metric are you using? Data can’t really be fit to a curve without data to plot.

      The entire contention is you misunderstanding how exponential functions work., i.e. “if it’s exponential, shouldn’t we be rapidly accelerating by now?” Betrays a fundamental misunderstanding.

      People don’t expect AI to be exponential because of existing data. It’s because once AI starts significantly improving itself, the advancement of AI, x, starts to apply to itself x^2 .

      We won’t know if it is, in fact, exponential until after the “knee” of the curve. But a slow advancement now does not preclude rapid acceleration in the near future. You’ve repeatedly demonstrated throughout the thread that you don’t understand this.

      source

      • Clinicallydepressedpoochie@lemmy.world ⁨3⁩ ⁨days⁩ ago

        Without the “knee” of the curve there is no exponential growth.

        source

        • -> View More Comments