Full Explanation
AI models don't read words — they read tokens, small chunks of characters. When you misspell a word, most of the token pieces remain the same. The pattern survives, and because meaning in language is highly redundant, the continuation still points to the same idea.
Typos don't disappear — they're still there. But they usually don't matter much because models were trained on messy internet text full of broken grammar, misspellings, and mixed languages. The model isn't fragile.
There is a limit. If a typo changes the structure completely — in code, formulas, IDs, or passwords — meaning collapses because the pattern is gone.
AI doesn't understand spelling. It understands patterns. And as long as the pattern survives, typos usually don't matter.