Research · Guides · Explainers

RAMGPT.

Independent AI research & technical writing

Research topics & method ↗
Explore RAMGPT

The RAMGPT learning series

AI Foundations.

One concept at a time.
A clear path from the basics to running a model.

No computer science background required. Start at the beginning, or find the stage that connects to what you are reading now. Lessons build on one another.

Start with lesson 1 38 published lessons · 6 stages

Written by William and machaMochaLatte, with editorial review by RAMGPT. This growing path includes every published lesson.

Stage 01 · Lessons 1–5

Build the map

Connect AI, machine learning, parameters, tensors, and the forward pass.

  1. 01↗
  2. 02↗
  3. 03↗
  4. 04↗
  5. 05↗

Stage 02 · Lessons 6–8

Understand learning

Follow how loss, gradient descent, and backpropagation change a network.

  1. 06↗
  2. 07↗
  3. 08↗

Stage 03 · Lessons 9–12

Look inside a Transformer

Connect embeddings, tokens, attention, and the Transformer block.

  1. 09↗
  2. 10↗
  3. 11↗
  4. 12↗

Stage 04 · Lessons 13–14

Train for a purpose

Distinguish pretraining from fine-tuning and understand what each changes.

  1. 13↗
  2. 14↗

Stage 05 · Lessons 15–18

Run the model

Follow inference, sampling, context, KV cache, and quantization.

  1. 15↗
  2. 16↗
  3. 17↗
  4. 18↗

Stage 06 · Lessons 19–38

Further lessons

Continue with the latest published lessons beyond the core learning path.

  1. 19↗
  2. 20↗
  3. 21↗
  4. 22↗
  5. 23↗
  6. 24↗
  7. 25↗
  8. 26↗
  9. 27↗
  10. 28↗
  11. 29↗
  12. 30↗
  13. 31↗
  14. 32↗
  15. 33↗
  16. 34↗
  17. 35↗
  18. 36↗
  19. 37↗
  20. 38↗