... something so ambiguous. He insists the "rule" is based on vowel pairings without accepting that it is a general, and not a specific rule. Furthermore, he ignores the actual genesis of where the soft g/hard g divide actually comes from: in words where the origin is Greek or Latin, the soft g usually has e, i, or y behind it, as he said. However, in words of Germanic origin (a not-insignificant contributor to English) along with a few other words, the hard g is often used even with those vowels. But since this is an acronym, of recent vintage, with no ties to either etymology, to dogmatically insist it follow "rules" (which really aren't) is a sign that video is clickbait and not intended to actually solve anything.
Furthermore, pronunciation is determined not by linguists (with apologies to France) but by how words are commonly pronounced. For instance, the word "algae" in American English actually follows the soft g rule because ae is a digraph for the e,i,y vowel family. In British English, it's widely pronounced "wrong" (with a hard g) according to the rules because they're not properly identifying the digraph (suck it you stuffy Brits). But who cares? It's how THEY say it, and saying it with a soft g sounds wrong to them. In other words, you can rail against the masses all you want but if the common pronunciation is a hard g then it doesn't matter.
Lastly, I believe people are generally simply defaulting to the most similar, familiar word they know as a pronunciation guide, and wouldn't you know it but "gif" looks a lot like someone forgot to write the t in "gift"? Makes sense that it's a hard g then. Long live the exception to the rule.