Absolutely.How often have people complained that problems get solved by the casters throwing magic at it? Why do you think they are having problems? IMO, it's because a lot of DM's fail to realise that casting spells is functionally the same as player authorship.
The significant effect on play is not in-character vs out-of-character; it's whether or not the players have to engage the challenge on the GM's terms (in idealised terms, we could say - engage the GM's framed scene via the action resolution mechanics) or get to rewrite the challenge in their own terms (in idealised terms - bypass action resolution and just reframe the scene).
If you think this is not fun when a player does it by dropping a fate point, it's almost certainly not going to become more fun just because the player drops a wish spell instead!
With all due respect, I think you are proving Hussar's point.A PC casting a spell and "rewriting a challenge" is no different than a fighter swinging his sword and doing the same thing. Sure, one is a bit broader and the other a bit narrower in scope, but for all intents and purposes it's the same thing. The authorship is limited to the rules and I suspect that many DMs do not drill down into it to the extent that you have.
In any game, the authorship it limited to the rules. But if the players all want a game in which the GM frames the challenges, but the rules of the game permit the players to reframe or bypass those challenges, it is not an especially big deal whether those rules are framed in in-character or out-of-character terms. It will still be anti-climactic and suck.
This is why high level modules in D&D classically would have a list of banned spell effects. To stop the anti-climax. The fact that there is an ingame rationale (magic works differently on the Abyss, or on the demi-plane, or due to underground magntetic forces, or whatever) is secondary - it's not that someone imagined an ingame limitation then wrote up rules for it! Rather, Gygax recgonised that teleportation is a scene-reframing ability, realised that it can tend to make modules suck, and so banned it from the D-series and then wrote in an in-game figleaf.
Again, I don't agree.Using a spell, when the spell is part of the rules and is used RAW/RAI, is no different for a wizard than a fighter using a sword to do deal with the threat. It's a character using a mechanical ability that is inherent to the character.
<snip>
A player using a plot point or fate point to declare some item exists in the fictional universe of the game is a player using an authorial ability, and is considerably different from a spell.
Consider OGL Conan, which has a modest fate-point mechanic. I could trivially rewrite that as an ingame mechanic - certain behaviours draw the admiration of Crom, who awards those who perform them with a certain number of "boons" - but it wouldn't fundamentally change the play of the game.
The fact that a given ability is inherent to the character is secondary - it has no more than rather minor consequences for some of the narration of ingame events. The real issue is how it affects the content of the fiction, and the ability of the players to enage with that content. If I have to wear down an enemies hit points via the default combat rules, that's one thing; being able to cast Imprisonment, or Power Word Kill, and simply have that enemy dissappear, isn't radically different from the player of a fighter having some sort of rationed ability to say "OK, I win this fight without needing to roll any dice!"
Likewise, whether the boxes appear because I spend a fate point, because I mention the possibility to the GM and s/he follows my lead, or because I cast a "conjure boxes" spell, the basic dynamic of play is the same: I want there to be boxes in the scene, and there are. Or, to give a different example, the diffrence in play experience between Limited Wish or Wish used to undo a bad result, and a fate point spent on a reroll, is pretty minimal - either way, an undesired outcome is reversed.
This is another example that illustrates why I think the Forge is right to focus on the reality of the play experience, rather than infiction considerations. From the infiction point of view, a spell is something a character can do, and it has no constraints other than those of the ingame logic of magic. In practice, however, this produces spells that break the game in well-known ways - like 3.5's Shivering Touch (? the dragon-killing DEX-draining one), like teleport in many cases, like various summoning spells, like many scrying spells.
The 3E DMG advocated a mixture of illusionist GMing (fudge the outcomes) and in-game first techniques (a good GM can handle broken spells and incorporate them into his/her game). I prefer to be upfront - if I don't want a game in which the players get to reframe any challenging situation at their leisure, just drop teleport from the game. If I want to run mysteries, just drop scrying magic. Approach design from the perspective of the real world - what sort of authority do these character abiliites give to the players vs what sort of authority does the GM need to make the game fun for the participants. Then rewrite the ingame fiction to suit.