Your next step

Why models become more capable

Data, capacity, compute, synthetic examples and answer-time work.

What you’ll learn

Better data can matter more than more copies
Identify a quality improvement in a training collection.
About 4 minutes · Open activity
More capacity changes what can be learned
Interpret parameter count as capacity rather than an intelligence score.
About 4 minutes · Open activity
Compute is work done by hardware
Recognize the physical resource behind training computation.
About 4 minutes · Open activity
Balance capacity, data and training work
Explain why scale involves tradeoffs under a budget.
About 4 minutes · Open activity
Synthetic examples need a quality filter
Recognize useful and harmful roles for generated training material.
About 4 minutes · Open activity
Learning to solve can differ from learning to sound helpful
Distinguish reasoning-oriented training from surface response preference.
About 4 minutes · Open activity
Extra work at answer time is another lever
Distinguish inference-time search or verification from retraining.
About 4 minutes · Open activity
Different gains have different evidence
Identify whether a claimed improvement concerns the model or the system around it.
About 4 minutes · Open activity
Explain a fictional model release
Read four stated changes, identify their mechanisms, and choose the fresh evidence needed before calling the release better for a particular user.
About 5 minutes · Open activity