Points 1 and 2 should be fairly obvious: you give up one action (i.e. one turn) to cast True Strike to increase your chances at a spell or ability you don't really want to waste.
The trouble is, in virtually all examples I've ever seen, it
still doesn't make sense, assuming you're fighting something, because of the action economy and the relatively high value of cantrips. And the more powerful your cantrips are, the worse True Strike becomes.
And yes your gut feeling is a gut feeling everyone has. I had the same.
But we're wrong. When you do the math, it almost inevitably works out that actually, on average, that's a bad idea. This is for a few reasons:
1) As noted, the value of cantrips (or other uses of your action, like using a weapon, or grappling or whatever), is quite high. Damaging cantrips scale, too, which often gets overlooked. True Strike does not scale, but it does take your entire Action.
2) Most high-value non-spell damage abilities are declared AFTER you hit in 5E. Not all, but a very large proportion. Smite particularly. You going to do a huge smite? Doesn't matter if you miss, because you don't trigger the smite until you hit.
3) Most spells which involve a roll-to-hit are not drastically more damaging than cantrips, or have other features that make them less-compatible with True Strike (like, they involve multiple attacks, and True Strike only applies to the first).
4) True Strike is a Concentration spell. This is absolutely killer for True Strike. Many amazing spells, from Bless onwards, are Concentration. Any spellcaster doing anything hard, is probably using their Concentration on something already. As a bonus you could easily lose your True Strike entirely if you get damaged, meaning you just blew an action on exactly nothing.
So people, quite naturally "feel like" True Strike should be useful. That feeling is why it's in the game. Because people, including apparently game designers, don't do that math. The moment that they actually do the math, it becomes clear that it's so corner-case in its usage, that's it's quite likely to never see a legitimate use.
Your 1 & 2 seem obvious, that's not actually how it's likely to work out, mathematically, because as I said, most of those "attack abilities" are post-hit (so no benefit, you just lose the damage from the round you cast True Strike instead of attacking) and most of the spells which roll to hit don't gain as much damage, on average, from rolling with Advantage, as you lose from not trying an attack with a cantrip in the previous round. The easier the to-hit roll, the less you gain from True Strike, too.
I think there's probably a point where, when you are attacking a sufficiently high AC (and I think it would have to be unusually high for that level of play), and using a high-damage roll-to-hit spell (one of the nastier Cause Wounds, maybe), where it might make sense, but it's going to be rare. We can do the math on various situations if you like, but I'm going to want participation if I'm going to do math for people.
The partner of True Strike, Blade Ward, is actually useful. It is preemptive healing without a spell slot. In games where a PC is allowed to swap out cantrips, I have taken it at low levels as a bard and then waded into battle when low on spell slots to draw fire from enemies so that their attacks do not take down the fighter, etc...
It falls out of use at higher levels, but it definitely stretched the healing in the party at the lowest levels.
This is only true if your damage is so bad that you can't contribute meaningfully to taking down the enemies, which with most Bards, shouldn't be the case at low levels (though I can concede it may be with some). Still, it's definitely got more potential uses than True Strike, and yes the "I'd rather they tried to hit me and not do much than tried to hit the guy on 5HP when I have no heals left - or just one to save him" is a real thing, potentially.