Digital vs. Analog (epistemic)

In CF, digital vs. analog is a distinction about evaluation:

• Digital. Discrete states: pass/fail, refuted/non-refuted, alive/dead.
• Analog. Continuous: 73.4% confidence, "pretty good".

CF favours digital evaluation where possible.

Why digital wins

• Robust against noise. A small error doesn't change pass to fail.
• Composable. Pass × pass = pass; multiplying analog values is unreliable.
• Easy to communicate. Two states are unambiguous.
• Error correction is fast. No need to update a continuous score.

Why analog loses

• Sensitive to noise. Small errors accumulate.
• Hard to combine. Cross-dimensional addition fails.
• Easy to game. "I'll just nudge the score up."
• Hard to communicate. Two people may score differently.

Why not always digital

• Some things are continuous. Temperature, weight, speed.
• Sometimes you need precision. Sub-breakpoint distinctions matter.
• Convert to digital where you can. Use breakpoints.

CF's recommendation

1. Identify breakpoints. Where does continuous become discrete?
2. Apply pass/fail at breakpoints. Convert to digital.
3. Stay analog only inside categories. For fine-tuning.
4. Re-digitise when communicating. Pass/fail, not "kind of".

Computer science analogue

CF borrows the digital/analog distinction from computing:

• Digital computers use discrete bits; they resist noise.
• Analog computers are faster but fragile; small errors propagate.

The same logic applies to evaluating ideas.

"Digital vs. Analog Thinking" is a CF essay title. — criticalfallibilism.com