3.3#
Education#
1. Millenia-of-Experiences
\
2. Our-Heritage -> 4. Collective-Unconscious -> 5. Decode-Teachers -> 6. Apply-Thrive
/
3. Artifacts-Rituals
It is helpful to compare the learning process of artificial intellgience to the education process of human intelligence what we mean by reproducible research and explained some of the additional benefits.
Dionysus 1, 2, 3#
Millenia, centuries, decades, ovarian or other lottery
Human, colonial, national heritage, and the 10,000 hour rule
Observed events by period (olympics), year (christmas), month (idi), week (synagogue), etc
Sing O Muse 4#
We must not let daylight in upon the magic. Yet Walter Isaacson seeks to hack it
Apollo 5, 6#
Interpreters of the magic & its application to lifes challenges
Journal & note-take, finds resonance of great heritage with personal life,
represent
1. Inputs-Food \ 2. Compute-Teeth -> 4. Latent.Space-Enzymes -> 5. Decode-Absorption -> 6. Represent-Growth / 3. Encode-Gut
3.3.1 Happy-chances#
Luck: having the input
data
& experiences as well as theinfrastructure
to benefit from RLHF over the 10,000 hours
3.3.2 Myth-making#
AI or Machine-Learning
3.3.3 Power-consolidation#
Common Sense
Nutrition#
Looks like all treatments have a large margin of error (
#1
): We had 2 surgeries on the breast, 10 rounds of chemotherapy, 33 rounds of radiation, exemestane, tamoxifen, letrozole, zoladex Yes, the margins of error are large when the regional nodes are involved at the time of initial diagnosis (#2
) One may think of local disease as having a mean & precise estimate of anatomical involvement (#3
). Beyond local you have regional nodes and distant metastases, increasing the variance in phenotype and prognosis
1. Error: Outlier
\
2. Compute: Bias-Variance -> 4. Latent-Space: Uncommunicable -> 5. Decode: Hack -> 6. Represent: Success
/
3. Mean: Average Infrastructure