As someone who spent embarrassingly long trying to log in to a system one afternoon because I was misreading an "l" as an "I", I can say that it does sometimes matter - it doesn't need to be in the same word.

Then again, since you set the challenge - lamb and Iamb

I can see no good reason for us to ever use the same symbol to represent two different letters, that's just silly business.

Edit: for the record, I didn't downvote you. Interestingly enough, I just noticed that it looks as though I can no longer downvote direct replies to my comments. Seems like a reasonable rule to have in place.

It's not just about words. Written media is no longer limited to a few basic structures and voices. Grammar needs to be able to shift away from English in-line while maintaining semantic clarity, and sometimes that means the reader needs to distinguish between graphemes without any of the typical contextual cues that classic written languages provide.

The most common issue is hand-copying a piece of text between devices, such as a cryptocurrency address or a URL. Sure, it would be great if those sorts of non-linguistic bits of information were always tagged appropriately such that they show up in a different font, but that rarely happens in reality. Even here, on Hacker News, we can only talk in plaintext with a few minor markup features.

A general-purpose web font should be able to handle a variety of linguistic and non-linguistic strings and not choke up on edge cases. In that respect, I really like the differentiation between a capital I and a lowercase L in Clear Sans, no matter the inclusion of 'Sans' in its name.

Io and behold — lambic pentameter.

(lo, as in a lota; lambic, as with an Iamb.)

A password with any of these characters might give you trouble if they're not easily distinguishable: I1l